Analysis of the diversity of aerobic, thermophilic endospore-forming bacteria in two Algerian hot springs using cultural and non-cultural methods

Abstract

Terrestrial hot environments are important resources for isolation of thermophilic microorganisms. Few studies have been made on microbial diversity of Algerian geothermal sites. This paper reports the diversity of thermophilic, aerobic endospore-forming bacteria from water and sediment samples taken from Hammam Ouled Ali and Hammam Debagh, two hot springs with a wide range of temperatures and a very rich mineral composition, located in the region of Guelma, north-east of Algeria using culture-dependent and culture-independent approaches Sequences of the V4 region of the 16S rRNA gene from environmental DNA extracted from sediment samples were analyzed and a set of isolates from water and sediment have been characterized by phenotypic and molecular methods. Phylogenetic surveys using environmental DNA sequences indicated that three families dominated the two hot springs: Planococcaceae, Bacillaceae, and Paenibacillaceae. Phenotypic characterization revealed the morphological, biochemical, and physiological properties of these microorganisms, all of which exhibited a range of common extracellular enzymatic activities. Amplified ribosomal DNA restriction analysis (ARDRA) was used to cluster isolates into different phylotypic groups and phylogenetic analysis of 16S rRNA gene sequences of selected isolates showed that all were closely related to four genera of thermophilic Bacilli: Bacillus, Anoxybacillus, Geobacillus, and Brevibacillus. Our results provide important insights into the microbial ecology of Guelma hot springs. They showed that the phylogenetic diversity of bacterial communities within the two studied hot springs was mostly aerobic, with the presence of taxonomic groups of great biotechnological interest. Bioprospection of thermozymes and other biomolecules within these communities will probably provide a data basis for their industrial exploitation.

Introduction

Thermophilic and hyperthermophilic microorganisms inhabiting hot aqueous environments such as terrestrial hot springs have unique molecular adaptation capacities and can be a significant source of bioactive molecules (Gomes et al. 2016).

The phylum Firmicutes consists of at least 26 families and 223 genera. Some members of the Firmicutes are thermophiles and/or halophiles (Schleifer 2009).

Inside this phylum, thermophilic endospore-forming bacteria constitute an important taxonomic group with widely cited potential biotechnological applications (Raddadi et al. 2015). Many of their bioactive compounds such as antimicrobial agents, compatible solutes, and biopolymers can find their way into industrial, pharmaceutical, and medical use, but their thermozymes, especially hydrolases, are maybe the most exploited and studied resource (Charlesworth and Burns 2016; Kambourova et al. 2016; Dumorne et al. 2017). For example, thermostable amylases, cellulases, and proteases of great industrial interest have been isolated from thermophilic strains of the Firmicutes (Elleuche et al. 2015; Contesini et al. 2017; Zhang et al. 2018). Indeed, there is an increasing focus on the utilization of renewable sources to satisfy the exponentially growing industrial needs. Microorganisms living in extreme habitats are an ideal source for polymer degraders, which allow to perform biotransformation reactions at non-conventional conditions under which many proteins are completely denatured (Antranikian 2007).

Algeria contains numerous and largely unstudied extreme environments. For example, over than 240 geothermal springs have been recorded in Northern Algeria but the data on their biodiversity is very limited. These springs might be considered as naturally well protected, renewable, and quasi-infinite resources (Fekraoui and Kedaid 2005; Amarouche-Yala et al. 2014), but can, on the other hand, be particularly vulnerable to microbial contamination due to human activities and their biodiversity could be significantly affected (Field 1999).

In this work, we investigate the diversity of thermophilic, aerobic endospore-forming bacteria of Hammam Ouled Ali and Hammam Debagh, two hot springs located in the region of Guelma, North-East Algeria, using culture-dependent and culture-independent methods.

Materials and methods

Studied sites, sampling, and isolation procedure

Environmental samples were collected from four different sites of two hot springs: Ouled Ali (36°34′ N; 7°23′ E) and Debagh (36° 27′ N; 7°16′ E) located in the Department of Guelma, North-East Algeria (Fig. 1). These springs are used for their therapeutic properties as baths and spas and are found to be near travertine deposits with a very rich mineral salt composition. Hammam Debagh is the hottest spring in Algeria, with temperatures rising up to 98 °C.

Fig. 1

Map of Algeria depicting sampling locations

Water and sediment samples were used for isolation of thermophilic bacteria from each site. Four to eight sediment samples of 20 g from each site were used for the extraction of environmental DNA. Temperature and pH values of the sampling sites were measured in situ, and pH was rechecked at the laboratory. Samples were collected in sterile containers and were stored at 4 °C until further use.

Environmental DNA extraction and sequencing

Metagenomic DNA of 23 sediment samples was extracted using a previously described protocol (Miller et al. 1999). The quantity and quality of the genomic DNA was measured with a NanoDrop Spectrophotometer (Thermo Fisher Scientific). Extracted metagenomic DNA was purified with the GeneJET PCR Purification Kit (Thermo Fisher Scientific).

16S rRNA gene V4 variable region PCR primers (515/806 with forward primer barcode) were used in a 30-cycle PCR (5 cycles used on PCR products) using the HotStarTaq Plus Master Mix Kit (Qiagen, USA) under the following conditions: 94 °C for 3 min, followed by 28 cycles of 94 °C for 30 s, 53 °C for 40 s, and 72 °C for 1 min, after which a final elongation step at 72 °C for 5 min was performed. After amplification, PCR products were checked in 2% (w/v) agarose gel to determine the success of amplification and the relative intensity of bands. Multiple samples were pooled together in equal proportions based on their molecular weight and DNA concentrations. Pooled samples were purified using calibrated Ampure XP beads, and the pooled and purified PCR product was used to prepare a DNA library using the Illumina TruSeq DNA library preparation protocol. Sequencing was performed at MRDNA (Shallowater, TX, USA) on an Illumina MiSeq following the manufacturer’s guidelines.

Diversity survey of aerobic endospore-forming bacteria in environmental DNA

Sequence data were processed using the MRDNA analysis pipeline (MR DNA, Shallowater, TX, USA). Sequences were joined and depleted of barcodes. Sequences < 150 bp and sequences with ambiguous base calls were removed. Sequences were denoised, operational taxonomic units (OTUs) generated, and chimeras removed using UCHIME (Edgar et al. 2011). OTUs were defined by clustering at 3% divergence (97% similarity) by CD-HIT (Li and Godzik 2006). Final OTUs were taxonomically classified using BLASTn against a curated database derived from Greengenes, RDPII, and NCBI (DeSantis et al. 2006; Cole et al. 2014; Benson et al. 2018).

Availability of data

The sequence data sets were submitted to NCBI Sequence Read Archive (SRA) under the accession number (SRP153378) for public access.

Isolation of bacteria

For the isolation of aerobic thermophilic bacteria, 1 g aliquots of sediment was added to enrichment broth (0.4% peptone, 0.2% yeast extract, 0.2% NaCl w/v) and incubated at 55 °C for 24 h. Of water samples or enrichment broths, 0.1 mL volumes were then sprayed onto five different growth media: M1, Thermus medium agar (Atlas 2005); M2, Medium 256 agar (Atlas 2005); M3, Nutrient agar (Sigma); M4, Plate Counting Agar (Sigma); M5, Actinomycetes Agar (Difco), pH adjusted to 7.2 ± 0.2. Plates were incubated at 55 °C for 72 h.

Further purification of isolates was done on Thermus agar plates. The isolated colonies were stored at 4 °C.

Phenotypic characterization

The selected isolates were characterized using 103 phenotypic tests based on minimal standards for describing new taxa of aerobic, endospore-forming bacteria and criteria described by the Bergey’s Manual of Systematic Bacteriology (Logan et al. 2009; De Vos et al. 2011).

Colonies morphology was studied by inoculating the cultures on M1 plates and incubating at 55 °C for 24 h. Cell morphology was observed by light microscopy after Gram staining. The presence of endospores was investigated using the Schaeffer-Fulton technique (Bartholomew and Mittwer 1950).

The physiological and biochemical characteristics of each isolate were determined using standard procedures, following 24 to 48 h incubation at optimal growth temperatures, in triplicate. Temperature, pH, and NaCl requirements were studied by inoculating the cultures on M1 agar/broth and incubating at a range of temperatures (10–80 °C, with an interval of 5 °C), pH levels (5.0–10.0, with an interval of pH 0.5 unit), and NaCl concentrations (0.5, 1.0, 2.0, 3.0, 3.5, 4.0, 5.0, 7.5, 10.0, 12.5, and 15% w/v) to record the minimum, optimum, and maximum temperature-, pH-, and salinity-dependent growth of all isolates. Catalase activity was tested with 10% (v/v) H₂O₂, oxidase production was checked using Oxidase Strips (Sigma), and β-galactosidase production was checked using Ortho-Nitrophenyl-β-Galactoside (ONPG) disks (Fluka). Carbon substrate utilization was evaluated using 1% (w/v) sugars: (D(+)-glucose, D(+)-fructose, D(+)-galactose, D(+)-mannose, D(+)-saccharose, D(+)-maltose, D(+)-lactose, and dextrin), 0.1% (w/v) amino acids (L-tyrosine, L-glycine, L-glutamic acid, L-threonine, and L-lysine), 0.1% (w/v) organic acids (lactate, acetate, citrate, propionate, and succinate), 0.1% (w/v) alcohols (mannitol, glycerol, ethanol, and methanol), and 1% (w/v) polysaccharides (starch, pectin) on agar plates containing 0.05% (w/v) NH₄Cl and 0.005% (w/v) Na₂HPO₄. Acid production from growth on sugars (1%, w/v) was recorded on nutrient broth containing 0.003% of phenol red as a pH indicator. Other standard biochemical tests carried out included the following: Methyl Red (MR) and Voges Proskauer (VP) reactions, formation of indole, urease production, mannitol motility, and triple sugar iron tests (Tindall et al. 2007).

Screening for extracellular hydrolase production

The isolates were tested for extracellular enzymatic activities by growth on solid media. Amylolytic activity was tested on a growth medium containing 1% (w/v) of starch (Gordon et al. 1973). Caseinolytic and gelatinolytic activities were tested on growth media containing 1% (w/v) of casein and 0.4% (w/v) of gelatin, respectively (Frazier 1926; Priest et al. 1988). Cellulolytic activities were investigated using a growth media containing 0.5% (w/v) of carboxymethyl cellulose (CMC) (Bragger et al. 1989). Pectinolytic activities were determined using 1% (w/v) pectin (Soares et al. 1999), and extracellular lipases were screened in growth media containing 1% (v/v) of Tween 20 or Tween 80 (Khyami-Horani 1996). All cultures were incubated at optimal growth temperature for 48 to 72 h.

Phenogram construction

A dendrogram of the phenotypic characteristics of the isolates was constructed using NTSYS-pc ver. 2.02i (Exeter Software, Setauket, NY). A binary 0/1 matrix, based on the absence or presence of a phenotypic feature, was assembled. Simple matching (SM) coefficients were calculated, and a dendrogram was constructed using the unweighted pair group method with arithmetic mean (UPGMA) (Gower 1971).

DNA extraction, 16S rRNA gene amplification

All strains were grown aerobically on M1 medium (pH 7.2) at 55 °C for 24 h. Genomic DNA was extracted using a modified protocol as described previously (Miller et al. 1999). The quantity and quality of the genomic DNA was measured using a NanoDrop Spectrophotometer (Thermo Fisher Scientific).

The 16S rRNA gene was amplified by polymerase chain reaction (PCR) using universal bacterial primers E9F (GAGTTTGATCCTGGCTCA) (Farrelly et al. 1995) and U1510R (GGTTACCTTGTTACGACTT) (Reysenbach and Pace 1995).

A typical PCR contained the following (final concentration): 1× DreamTaq Buffer, 1% (v/v) bovine serum albumin, 1.25 U DreamTaq Polymerase (Thermo Scientific), 1 μM (each) primer, 200 μM of each deoxynucleoside triphosphate, and 10 to 100 ng template DNA in a 50-mL reaction volume. PCR conditions were as follows: 95 °C for 3 min; 30 cycles of 95 °C for 30 s, 52 °C for 30 s, 72 °C for 85 s; and a final incubation at 72 °C for 5 min. PCR products were electrophoresed and visualized on a 1% (w/v) agarose gel.

Amplified ribosomal DNA restriction analysis and sequencing

Restriction digestion amplified ribosomal DNA restriction analysis (ARDRA) profiles on agarose gels, generated by AluI and HaeDIII fast digests (FD) (Thermo Fisher Scientific), were analyzed to avoid sequencing redundant clones.

The reaction mixture containing 3.33 μl amplified 16S rRNA gene product, 0.67 μl Tango buffer (10×) for AluI or 0.67 μl FD buffer (10×) for HaeIII FD, 0.33 μl AluI or 0.33 HaeIII FD (10 U/μl), and 5.67 μl deionized water was incubated at 37 °C for 3 h (AluI) or 5 min (HaeIII FD). AluI was inactivated by heating at 65 °C for 20 min. Restriction digest products were loaded into 2.5% (w/v) agarose gels, electrophoresed at 90 V for 90 min. Restriction fragment size was estimated by inclusion of a 100-bp DNA ladder (Thermo Fisher Scientific). The gel was visualized using a Bio-Rad Imager Gel Doc XR System, and the restriction patterns of each sample were compared. Fragments smaller than 100 bp were not included in this analysis. The NTYSIS-pc package (version 2.02i, Exeter Software, Setauket, NY) was used to score similarity and clustering analysis using the binary data. Dice coefficients were used to calculate the similarity among the isolates and dendrograms were constructed using the UPGMA method (Nei and Li 1979).

Amplicons of the selected isolates from the ARDRA were then purified with the NucleoSpin Gel and PCR Clean-up (Thermo Fisher Scientific). E9F and U1510R primers were used for capillary sequencing at the Central DNA Sequencing Facility, University of Stellenbosch.

Phylogenetic analysis

Identities with 16S rRNA sequences of described taxa were investigated using the nBLAST tool against the EzBioCloud Database of cultured organisms (Yoon et al. 2017). Multiple sequence alignments were performed with MEGA 7 software using the ClustalW algorithm (Thompson et al. 1994). 16S rRNA gene-based phylogenetic trees were constructed, based on neighbor-joining (Saitou and Nei 1986) and maximum composite likelihood models (Tamura et al. 2004) with 1000 bootstrap replications (Felsenstein 1985) using MEGA 7 program package (Kumar et al. 2016). The 16S rRNA gene sequence of Sulfobacillus acidophilus DSM 10332^T was used as outgroup.

Nucleotide accession numbers

Nucleotide sequences have been deposited in the NCBI database under accession numbers MF136820 to MF136840.

Results

Sampling sites

The samples for this study were collected from four different sites at two hot pools: Debagh1 (Db1), Debagh2 (Db2), Debagh3 (Db3), and Ouled Ali (OA). Temperature and pH values for each site are shown in Table 1. Chemical analyses of the spring waters presented by different studies showed some common parameters between the two hot springs: high temperature, slightly acid to neutral pH, high concentrations of silica, major concentrations of sodium chloride with large content of calcium sulfate, gas emissions of carbon dioxide and nitrogen at the thermal vents (Saibi 2009; Bahri et al. 2011; Amarouche-Yala et al. 2015).

Table 1 Localization and physicochemical data of the sampling sites

Aerobic endospore-forming bacteria communities in hot springs sediments

We performed a MiSeq sequencing of the V4 region of 16S rRNA amplicons from metagenomic DNA extracted from sediment samples from the four sites. Filtration of the data generated 2,238,462 good sequence reads from the 23 pooled samples. Db1 samples had the most important reads (887,026), followed by Db2 (511,956), OA (479,579), and Db3 (359,901). About 193,318 good reads were classified under archaea, while 2,002,967 good reads were classified under bacteria. For the bacterial diversity, the result showed 30 phyla (see supplementary). Firmicutes abundance among other bacterial phyla is shown in Fig. 2.

Fig. 2

Percentages of the major phyla in each site (the sequence percentage is above 0.5% in at least one site)

Prevalence of Firmicutes within the total bacterial community in each of the four sites was significantly different. It was very low in Db2 and OA (1.66 and 6.13%, respectively) in comparison with Db1 (12.86%) and Db3 (17.92%) pools. In general, Firmicutes were relatively a minor taxonomic group in the four sites. Dominant phyla were different in Db1 (Deinococcus-Thermus), Db2 (Aquificae), Db3 and OA (Chloroflexi).

When studying Firmicutes diversity present in each site at the family taxonomic level, we notice that aerobic endospore-formers are dominant in the four sites (Fig. 3). In fact, the presence of three major families of aerobic thermophilic endospore-forming bacilli is observed: Planococcaceae, Bacillaceae, and the Paenibacillaceae which were the Firmicutes dominant family in Db1 (61.61%) and Db3 (59.08%). Clostridiaceae and Thermoanaerobacteraceae were the most noticeable anaerobic endospore-forming families present in the sediment samples.

Fig. 3

Percentages of the major families within the Firmicutes phylum in each site (the sequence percentage is above 0.5% in at least one site)

The major thermophilic aerobic endospore-forming genera of Firmicutes phylum present within the bacterial communities of the four sites were Paenibacillus, Psychrobacillus, and Sporosarcina with at least 1% of the total sequences (see supplementary). Genera such as Bacillus, Geobacillus, Anoxybacillus, and Brevibacillus were minor (less than 1%) in both hot springs.

Isolation of thermophilic bacteria

One hundred seven aerobic, thermophilic, chemoheterotrophic bacterial isolates were obtained from the cultivation of the different samples collected from the four sites. Literature indicates that the temperature used for the isolation procedure (55 °C) allows to identify a wide range of thermophilic heterotrophic aerobic bacterial taxa from both hot and mild environments (Trujillo 2001; Logan et al. 2009; Battista and Rainey 2015; De Vos 2015; De Vos et al. 2015). The selection and purification of these isolates were based on the morphological variability of their colonies and their microscopic aspect. The Ouled Ali site samples provided 49 isolates whereas 58 isolates came from Debagh three sites. Sediment samples from both hot springs yielded 63 isolates while water samples provided 44 isolates. Thirty-five isolates were collected from growth medium M1, 25 from M2, 33 from M3, 11 from M4, and 3 isolates from M5.

Phenotypic characterization and extracellular hydrolase production

Phenotypic characterization and screening for extracellular enzymatic activities was performed for 20 isolates selected by ARDRA analysis (Table 2). Colonies were smooth or moist, of circular shape, cream or transparent and generally non-pigmented after incubation at 55 °C for 48 h. Cells were typically Gram-positive, rod-shaped, and single or in chains. Based on microscopic observation, ellipsoidal endospores were located in either a terminal or subterminal position.

Table 2 Phenotypic characterization of the strains

Strains were moderately thermophilic and neutrophilic, with growth ranges between 30 and 75 °C and optimum growth between 50 and 60 °C, and at pH values from 5.0 to 9.0 with optimum growth between pH 6.5 and 7.5. All the strains were able to grow at 0% (w/v) NaCl and tolerated salt concentrations between 1 and 10% (w/v). Optimum growth occurred at 3% (w/v) for the most halotolerant strain.

Isolates were aerobic and were able to produce catalase and/or oxidase. The majority of the strains were not able to produce β-galactosidase. Glucose was used as carbon source by all strains while the utilization of other substrates was variable. Fermentation of sugars, RM, and VP reactions were negative in most cases, indicating the presence of aerobic metabolic pathways. Indole and urease tests were negative. Mannitol-mobility and triple sugar iron test results were variable (Table 2).

All of the 21 strains showed positive results for at least one of the seven extracellular hydrolytic activities tested. Amylases and proteases were the most commonly produced extracellular enzymes while lipolytic and cellulolytic activities were the least expressed enzymatic activities (Table 2).

The phenotype-based UPGMA dendrogram of selected strains is shown in supplementary. Strains showed no particular pattern related to the physicochemical properties of their isolation sites. Clusters did not reflect any specific phenotypic differences.

Amplified ribosomal DNA restriction analysis

ARDRA analysis was conducted in order to eliminate redundant clones from the set of isolates. Analysis profiles, using AluI and HaeIII of 45 isolates, were combined and used to construct a dendrogram using UPGMA cluster analysis based upon the similarity index calculated using Dice coefficient (Fig. 4). In total, 23 different bands, ranging in size from 100 to 1000 bp, were observed. AluI and HaeIII are the most frequently used enzymes for ARDRA of members of the genus Bacillus, giving the highest number of differentiating bands (Blanc et al. 1997; McMullan et al. 2004; Kuisiene et al. 2007). We noted that restriction digestion with AluI was more discriminative than HaeIII FD. Isolates could be assigned to 28 distinct clusters, divided into five groups at similarity levels of 70% (group V), 84% (groups III and IV), 87% (group II), and 92% (group I). Results were compared with phenotypic characters to select representatives of each pattern group for further molecular characterization.

Fig. 4

Differences in restriction patterns. Dendrogram of 45 isolates ARDRA distances obtained by digestion with AluI + HaeIII FD. Isolates are regrouped under seven different pattern groups: I, II, III, IV, and V. The dendrogram was constructed using the UPGMA algorithm and Dice coefficients

Molecular characterization and phylogenetic analysis

Twenty-one organisms were selected for phylogenetic analysis: 8 originated from Debagh hot spring and 13 from Ouled Ali. 16S rRNA sequences were compared to the EzBioCloud version 2018-05 using the nBLAST tool, yielding sequence identities ranging from 98.69 to 100% between the selected strains and type strains (see supplementary). Isolates were affiliated with the following species: Anoxybacillus gonensis (Belduz et al. 2003), Anoxybacillus flavithermus (Pikuta et al. 2000; Dai et al. 2011), Anoxybacillus thermarum (Poli et al. 2009), Bacillus paralicheniformis (Dunlap et al. 2015), Bacillus licheniformis (Weigmann 1898), Geobacillus thermoleovorans (Nazina et al. 2001), and Brevibacillus thermoruber (Manachini et al. 1985; Shida et al. 1996).

The phylogenetic analysis is shown in Fig. 5. Strains affiliated to the Bacillaceae (Logan and Vos 2015b) and the Paenibacillaceae families (De Vos et al. 2015) were mostly related to four genera: Bacillus, Anoxybacillus, Geobacillus, and Brevibacillus with a predominance of Bacillus (10 strains) and Anoxybacillus (9 strains).

Fig. 5

Phylogenetic tree based on 16S rRNA gene sequences showing the relationship between the strains of this study (indicated in bold letters) and strains of related genera of the family Bacillaceae. The strains and their corresponding Genbank accession numbers are shown following the organism name and indicated in parentheses. The phylogenetic tree was made using the neighbor-joining method with maximum composite likelihood model implemented in MEGA 7. The tree includes the 16S rRNA gene sequence of Sulfobacillus acidophilus DSM 10332^T as outgroup. Bootstrap consensus trees were inferred from 1000 replicates, only bootstrap values > 60% are indicated. The scale bar represents 0.02 nucleotide changes per position

Discussion

In this work, the diversity of thermophilic endospore-forming bacteria in two Algerian Hot Springs with variable water temperatures has been studied with culture-dependent and culture-independent approaches.

The prevalence of the Firmicutes among the bacterial communities of the two hot springs by MiSeq sequencing of the V4 region of 16S rRNA gene extracted from sediment samples was investigated. The results showed that this phylum was minor in the four sites. Indeed, bacterial communities in sites Db1 (91 °C) and Db2 (95 °C) were dominated by two other aerobic chemoheterotrophic and more thermophilic phyla: Deinococcus-Thermus and Aquificae, respectively (Reysenbach et al. 2001; Battista 2016), while phototrophic thermophilic phyla were the most abundant when temperatures were under 80 °C in Db 3 (75 °C) and OA (54 °C), where Chloroflexi was the most prevalent phylum (Castenholz 2001). We also found that the prevalence of Firmicutes was more important when the temperature of the pools increased: from 6.13% in OA (54 °C) to 12.87% in Db3 (75 °C), to finally reach its highest percentage (17.92%) in Db1 (91 °C). On the other hand, it seems that the limit temperature for the growth of the members of this group was below the temperature of Db2 pool (95 °C) where the prevalence of representative OTUs dropped to 1.66%.

Results of microbial diversity studies for similar environments using the same metagenomic approach were variable. In the Manikaran hot springs (India), Firmicutes (28 to 84%), Aquificae (2 to 64%), and Deinococcus-Thermus (1 to 18%) were the dominant phyla (Bhatia et al. 2015). In Jakrem hot springs (India), Firmicutes (61%), Chloroflexi (21%), and Cyanobacteria (13%) were the most abundant (Panda et al. 2015). Firmicutes represented 7.5% of the microbial diversity in Eryuan hot spring (China) (Menzel et al. 2015) and between 6 and 21% of the total microbial diversity in Ma’in hot spring (Jordan) (Hussein et al. 2017). In Murtazaabad hot springs (Pakistan), Chloroflexi and Firmicutes represented 29 and 10% of the bacterial diversity, respectively (Amin et al. 2017).

The diversity of Firmicutes within the four sites was mainly aerobic. Three families of aerobic endospore-forming Firmicutes were dominant, especially in Debagh sites: Paenibacillaceae (De Vos et al. 2015), Bacillaceae (Logan and Vos 2015a), and Planococcaceae (Shivaji et al. 2014). Paenibacillus (Ash et al. 1993), Psychrobacillus (Krishnamurthi et al. 2010), and Sporosarcina (Yoon et al. 2001) were, respectively, the most abundant genera of each family. In OA, where the temperature was the lowest among all the studied points, the distribution between aerobic and anaerobic families was more equal with more diversity than in Debagh sites. This is confirmed by the fact that more variability was found in the strains isolated from OA than in Db. Higher temperatures tend to limit the diversity of the environments (Gómez and Parro 2012). The limited depth of the studied pools (between 50 cm and 1 m) might be favorable to the presence of aerobic thermophiles since it is known that dissolved oxygen quantity is more important in surface water (Hayashi and Rosenberry 2002).

Ninety-three isolates have been cultured on non-specific media (M1, M2, and M3). Most aerobic endospore-formers are chemoorganotrophs with aerobic metabolism and, despite their very wide diversity, will grow well on routine media such as nutrient agar or trypticase soy agar (Logan and Allan 2008).

From an initial set of 107 isolates, 13 strains from Ouled Ali and 8 strains from Debagh were selected for molecular characterization. Ninety percent of these strains belonged to the Bacillus and Anoxybacillus genera, with no significative distribution differences among the two hot springs of the study, in addition to two strains affiliated to Geobacillus and Brevibacillus isolated from Debagh and Ouled Ali hot springs, respectively. Reports have been made about the presence of representatives of these genera in hot springs located in geographically different sites: for example in Armenia (Panosyan 2017), Bulgaria (Derekova et al. 2008), China (Song et al. 2013), Morocco (Aanniz et al. 2015), Tunisia (Sayeh et al. 2010), Turkey (Adiguzel et al. 2009), and Russia (Foti et al. 2008). Isolates had typical phenotypic characters described for Bacillales members, they were commonly Gram-positive, rod-shaped, chemoorganotrophic, aerobic, and moderate thermophiles (De Vos et al. 2011; Logan and Vos 2015b). Identity between the strains of the study and type strains of databanks was between 98 and 100%. These type strains were isolated from different hot and mild environments. Thermophilic aerobic endospore-forming bacteria are widely distributed in water, soil, and many other environments and it is not uncommon to isolate them from mesophilic environments (Logan and Halket 2011; Müller et al. 2014).

All of the selected strains showed at least one extracellular enzyme activity among the tested activities. Amylases and caseinases production was numerically the most important and further studies of these enzymes might be interesting. For example, the proteasic activity of one of the strains isolated in this study, Brevibacillus sp. OA30, was investigated and an acid protease with remarkable properties was characterized (Gomri et al. 2018). Thermophilic aerobic endospore-forming bacteria are well known to be a good source of thermoenzymes, including xylanases, proteases, amylases, peroxidases, glucose isomerases, lipases, and DNA restriction enzymes (Maurer 2004; Schallmey et al. 2004; Satyanarayana et al. 2013; Panosyan 2017).

A dendrogram was constructed to discriminate the 21 isolates based on their phenotypic characters. It seems that the clusterization of the OTUs was generally not influenced by the strains origin. The physiology of the endospore-forming bacteria is influenced by complex multifactorial conditions of their environments and temperature might not be the decisive parameter in the variation of the phenotypic features of these microorganisms (Mandic-Mulec et al. 2016; Gauvry et al. 2017). A natural environment provides both sufficient nutrients and favorable physical and chemical conditions of temperature, pH, availability of water, redox, etc. (Carlin 2011).

ARDRA might be an interesting tool for the screening and the identification of new strains of heterotrophic thermophilic endospore-forming bacteria. Application of this technique in studies interested in the genetic variability of bacilli species revealed its ability to expose the exact lineage of the species rapidly (Kumar et al. 2014; Tiwari and Thakur 2014; Jain et al. 2017). The diversity rate of these groups can be very low in hot pool environments and ARDRA might help eliminate redundant clones in the prospecting process (Blanc et al. 1997; Brock 2012).

When comparing results from culture-dependent and culture-independent methods, we observed that the genera to which our isolated strains belonged (Anoxybacillus, Bacillus, Geobacillus, and Brevibacillus) had a very low distribution among the four sites. Similar observations were made on other hot springs in Malaysia (Chan et al. 2015), Tunisie (Sayeh et al. 2010), India (Kikani et al. 2015), and Bulgaria (Derekova et al. 2008) where these genera were the most frequently isolated while there abundance in environmental DNA was very low. It is important to mention that in culture-independent studies, the abundance of taxonomic groups of endospore-forming organisms belonging to the Firmicutes risk to be underestimated. The cell lysis of such organisms is often difficult, and so subsequent DNA purification and/or PCR-based applications may also be biased (De Vos 2011; Urbieta et al. 2015).

16S rRNA gene-based methods provide good phylogenetic information to the genus level, but, in themselves, give little information on function. Importantly, traditional approaches that group isolates on the basis of common metabolic properties may be limited in terms of phylogenetic insights, but provide clues to environmental factors favoring and selecting for particular groups and can be strong indicators of potential ecosystem function. A combined approach might allow to establish more clearly the links between diversity, community structure, physiological diversity, and ecosystem function, rather than merely characterizing the presence, absence, and identity of strains present (Mandic-Mulec and Prosser 2011).

Conclusion

This work is one of the few published studies describing the diversity of thermophilic bacteria of hot springs in Algeria. The sequencing of the V4 region of the 16S rRNA gene from environmental DNA extracted from sediment samples showed that Firmicutes were present as a minor phylum in the four sites with aerobic endospore-forming members belonging to Planococcaceae, Bacillaceae, and Paenibacillaceae families as major representatives.

Twenty-one thermophilic endospore-forming bacteria strains were isolated from Debagh and Ouled Ali hot springs and selected for phenotypic and genotypic characterization. Phylogenetic analysis based on the 16S rRNA sequences revealed that these strains were affiliated with genera Anoxybacillus, Bacillus, Brevibacillus, and Geobacillus distributed between the two hot springs. All of the 21 isolates had at least one extracellular hydrolytic activity. Classical culture-based methods remain important for understanding the molecular adaptations of microbial guilds, particularly those isolated from extreme habitats with potential applications. One of the perspectives of this work is the use of Next-Generation Sequencing methods to explore the studied hot springs with a focus on bioprospection of thermozymes and other biomolecules of great biotechnological potential.