Correspondence: Fernando Santos, División de Microbiología, Universidad de Alicante, Apartado 99, 03080 Alicante, Spain. Tel.: 34 965903870; fax: 34 965909569; e-mail: email@example.com
Tuz Lake is an inland thalassohaline water body located in central Anatolia that contributes to 60% of the total salt production in Turkey per year. The microbiota inhabiting this lake has been studied by FISH, denaturing gradient gel electrophoresis of PCR-amplified fragments of 16S rRNA genes, and 16S rRNA gene clone library analysis. Total cell counts per milliliter (1.38 × 107) were in the range of the values normally found for hypersaline environments. The proportion of Bacteria to Archaea in the community detectable by FISH was one to three. 16S rRNA gene clone libraries indicated that the archaeal assemblage was dominated by members of the Square Haloarchaea of the Walsby group, although some other groups were also found. Bacteria were dominated by members of the Bacteroidetes, including Salinibacter ruber-related phylotypes. Because members of Bacteroidetes are widely present in different hypersaline environments, a phylogenetic analysis of 16S rRNA gene sequences from Bacteroidetes retrieved from these environments was carried out in order to ascertain whether they formed a unique cluster. Sequences retrieved from Tuz Lake and a group of sequences from other hypersaline environments clustered together in a branch that could be considered as the ‘halophilic branch’ within the Bacteroidetes phylum.
Tuz Lake is an inland hypersaline water body with salinity above 30% NaCl, located in the central plateau of Turkey, 120 km south of Ankara. It is the second largest lake of Turkey, with a length of 90 km, a width of 35 km, and a total surface of 1665 km2 within a closed basin. The lake started its formation with tensional movements during the Late Cretaceous, followed by compressional episodes that, during the Late Eocene, resulted in the isolation of the basin from an open sea (Dirik & Erol, 2000). Water flows into the Tuz Lake through the Melendiz Stream and drainage channels of the Konya plain. In summer, the lake dries out and a 30-cm layer of salt forms because of the evaporation. In winter, water is not more than 2 m deep. Three hundred thousand tons of salt, which is 60% of the total salt production in Turkey, are obtained per year from the lake (Uygun & Sen, 1978; Çamur & Mutlu, 1995).
In previous studies based on the isolation of heterotrophic microorganisms (Birbir & Sesal, 2003) and identification of cultivable Archaea, Tuz Lake was found to contain a ‘viable, diverse, potentially and industrially important microbial community’ although so far (Birbir et al., 2007; Özcan et al., 2007) only the culturable archaeal assemblage has been partially characterized. Here, we present the study of the prokaryotic community of Tuz Lake using a culture-independent approach that includes FISH, denaturing gradient gel electrophoresis (DGGE) of 16S rRNA gene fragments PCR amplified from DNA extracted from the lake, and 16S rRNA gene library analysis. The aim of this work was to describe the prokaryotic community of a new inland hypersaline environment, uncharacterized previously. Although considerable information has been retrieved from crystallizer ponds from solar salterns (Oren, 2002), inland thalassohaline environments have not been extensively characterized by culture-independent techniques (Oren, 2002; Maturrano et al., 2006).
Many hypersaline environments with salinities approaching saturation (Elevi Bardavid et al., 2008) are dominated by the square archaeon of Walsby [the recently isolated Haloquadratum walsbyi (Burns et al., 2007)], with bacteria such as Salinibacter and Salicola also present as a minor component of the prokaryotic communities (Maturrano et al., 2006; Burns et al., 2007). However, some differences among the microbiota of different hypersaline environments have also been found, such as the presence of many Halobacterium-related phylotypes in 16S rRNA gene libraries retrieved from Andean salterns of many Halobacterium-related phylotypes (Maturrano et al., 2006) or the dominance of Halorubrum spp. in Adriatic salterns (Pašićet al., 2005). The second goal of this study was thus to see whether the Tuz Lake prokaryotic community had the structure found in saltern crystallizer ponds or presented some specific characteristics. In this way, we intended to gain a deeper insight into the differences in the prokaryotic communities inhabiting geographically distant hypersaline environments.
According to our results, the prokaryotic community in Tuz Lake is dominated by Haloquadratum and Salinibacter, thus resembling crystallizer ponds from artificial solar salterns (Elevi Bardavid et al., 2008). As in other hypersaline environments (Burns et al., 2004; Maturrano et al., 2006), sequences related to other extremely halophilic Bacteria and Archaea were also found. While Haloquadratum phylotypes retrieved from the lake were very closely related to those repeatedly found in salterns around the world (Australia, Peruvian Andes and both sides of the Mediterranean), the Salinibacter assemblage was quite site specific.
Materials and methods
Samples were taken from two different locations (W and E, located in the West and the East shores, coordinates N 38°46′25E 33°12′57 and N 38°44′14 E 33°36′01, respectively) of the Tuz Lake in May, July, and October 2005 and named as May W, July W, July E, and October W. Sample July W was chosen for a more complete characterization of the microbial community. The total salt concentration of this sample was 32% as measured using a hand refractometer (Eclipse).
Diamidino-2-phenylindole (DAPI) counts and FISH
Sample fixation was carried out as described previously (Antón et al., 1999) using the protocol optimized for fixation of extremely halophilic microorganisms. Hybridization, DAPI staining, and microscopy were carried out as described previously (Snaidr et al., 1997). At least two filters were analyzed and cells were counted in 30 different microscopic fields. The probes used for in situ hybridization were Arc915 for Archaea, Eub338 for Bacteria, and Non338 for nonspecific hybridization (Amann et al., 1990). The formamide concentration in the hybridization buffer was 35%.
Nucleic acid extraction
Microorganisms were collected by filtration of 50 mL of sample on a 0.22-μm pore size GV filter (Durapore, Millipore). The filter was cut into small pieces with sterile scissors, and placed in RNAse- and DNAse-free 2-mL cryotubes containing 600 μL of extraction buffer (100 mM Tris-HCl, 100 mM EDTA pH 8.0). Six microliters of lysozyme (3 mg mL−1) was added and incubated at 37 °C for 15 min. Then, 9 μL of proteinase K (150 μg mL−1) and 60 μL of 10% sodium dodecyl sulfate (SDS) were added to the tubes and incubated at 37 °C for 30 more minutes. After the addition of 120 μL 5 M NaCl and 90 μL CTAB solution (10% CTAB, 0.7 M NaCl), the tubes were incubated at 65 °C for 10 min, immersed into liquid nitrogen for 2 min, and incubated again for 2 min at 65 °C. The freeze-and-thaw steps were repeated three times. Nine hundred microliters of phenol : chloroform : isoamylalcohol (25 : 24 : 1) (PCI) was added, mixed, and centrifuged at 16 000 g for 5 min at 4 °C. The aqueous phase was removed and one volume of PCI was added, vortexed, and centrifuged again (two to three times) until a clear interphase between the aqueous and the organic phases was observed. Finally, nucleic acids were precipitated with ethanol and resuspended in 50 μL of sterile deionized water. To check the quality of nucleic acids, they were run in 1% agarose (LE, FMC Products, Rockland, ME) gels and visualized under UV light after ethidium bromide staining. Extracts were stored at −70 °C until used.
PCR amplification of 16S rRNA genes
16S rRNA genes were PCR amplified from total DNA using universal primers for Bacteria and Archaea. Reaction mixtures contained 1.5 mM MgCl2, 50 mM KCl, 10 mM Tris-HCl (pH 9), 200 μM of each deoxyribonucleotide triphosphate (dATP, dCTP, dGTP, and dTTP) (Invitrogen Corporation, Carlsbad, CA), 1 U of Taq DNA Polymerase (Invitrogen), 0.2 mM of each oligonucleotide primer, and 50 ng of template DNA in a total volume of 50 μL. The sequences of the forward primers were 21F 5’-TTCCGGTTGATCCTGCCGGA-3′ (DeLong, 1992) for the Archaea and 27f 5′-AGAGTTTGATCATGGCTCAG-3′ (Lane et al., 1985) for the Bacteria. The reverse primer for both domains was 1492r 5′-GGTTACCTTGTTACGACTT-3′ (Lane et al., 1985). The following conditions were used for amplification: a cycle of 94 °C for 3 min, 30 cycles of 94 °C for 15 s, 55 °C for 30 s, and 72 °C for 2 min; plus an extension step of 7 min at 72 °C. Negative controls were included with no addition of template DNA. Five microliters of PCR products were loaded onto 1% agarose gels in 1 × Tris–acetic acid–EDTA (TAE) buffer, stained, and visualized as above.
16S rRNA genes were PCR amplified from all the samples for DGGE analysis with the following primer sets: 341F-GC (5′-GCclamp-CCTACGGGAGGCAGCAG) and 907R (5′-CCGTCAATTCCTTTRAGTTT-3′) for Bacteria; 344F-GC (5′-GCclamp-ACGGGGCGCAGCAGGCGCGA) and 907R for Archaea. (Muyzer et al., 1993). The PCR program for the Bacteria primer set was: 94 °C for 5 min, 65 °C 1 min, 72 °C 3 min, and nine touchdown cycles of: 94 °C for 1 min, 65 °C (with the decreasing 1 °C each cycle) 1 min, 72 °C 3 min, followed by 20 cycles of: 94 °C for 1 min, 55 °C 1 min, and 72 °C 3 min. During the final cycle, the length of the extension step was increased to 10 min. The PCR program for the Archaea primer set was: 94 °C for 5 min (initial denaturation), and 30 cycles of 94 °C for 30 s, 56 °C for 45 s, and 72 °C for 2 min. The length of the extension step was increased to 7 min in the last cycle.
Cloning of PCR products
Ligation of the PCR products with the pCRII-TOPO vector, transformation of Escherichia coli TOP10, and selection of the transformants were carried out using the TOPO TA cloning kit (Invitrogen Corporation) according to the manufacturer's protocol. Two (one for Archaea, one for Bacteria) 16S rRNA gene libraries were generated from sample July W. Each library was generated with the pooled products of at least three independent PCR reactions.
Analysis of libraries and clone selection
Clones were screened for redundancies by amplified rDNA restriction analysis (ARDRA) (Vaneechoutte et al., 1992) with the enzymes HinfI and MboI (New England Biolabs). Enzymatic digestions were performed by incubating 10 μL of the amplified inserts with 5 U of enzyme and the corresponding enzyme buffer. The digestion products were analyzed in 2% agarose gels in 0.5 × Tris–boric acid–EDTA (TBE) buffer, stained, and visualized as above. Clones representing the different restriction patterns were selected for sequencing.
DGGE was performed with the DCode System (Bio-Rad, Hercules, CA). Electrophoresis conditions were 70 V for 16 h, with a linear gradient of denaturing agents from 40% to 70% (100% denaturing agents were 7 M urea and 40% deionized formamide) in a 6% 37.5 : 1 acrylamide : bis-acrylamide 0.75 mm thick gel, with 1 × TAE as running buffer. Gels were photographed with a Typhoon 9410 (Amersham Biosciences) imaging system after ethidium bromide staining. Selected bands were excised, resuspended in 20 μL Milli Q water, and incubated at 4 °C overnight. An aliquot of the supernatant was used for PCR reamplification with the original primer set, and 50 ng of reamplified PCR products were used for the sequencing reaction.
Sequencing and sequence analysis
The nucleotide sequences of the cloned products were determined from plasmid preparations (Wizard Plus SV Minipreps DNA Purification System, Promega, Madison, WI). Cloned products and purified PCR products were sequenced using the Big Dye Terminator Cycle Sequencing kit and an ABI PRISM 3100 Genetic Analyser (Applied Biosystems, Foster City, CA). 16S rRNA gene sequences were initially compared with reference sequences at NCBI (http://www.ncbi.nlm.nih.gov) using blast (Altschul et al., 1990). The correlation between primary and secondary structures was analyzed using the arb sequence editor (Ludwig et al., 2004). Nearly complete (for the 16S rRNA gene clones) and partial (for the DGGE bands) sequences obtained in this study were added to an alignment of over 50 000 primary aligned gene structures available at http://db-central.arb-home.de/. Sequences not included in the arb database were drawn from the GenBank database. Phylogenetic analysis was performed using the three algorithms implemented in the arb package: maximum-likelihood using fastdnaml and tree-puzzle programs (Strimmer & Von Haeseler, 1996; Schmidt et al., 2002), neighbor joining, and maximum parsimony. Partial sequences from DGGE analysis were added to the tree using the arb parsimony tool. All the sequences obtained in this study were analyzed simultaneously by mallard 1.02 software (Ashelford et al., 2006) in order to detect chimeras and other artifacts. Furthemore, each individual sequence was screened by pintail 1.1 software according to the authors (Ashelford et al., 2005) to confirm the presence of chimeric sequences.
Isolation of microorganisms
In an attempt to isolate new Salinibacter spp., the following medium was used: to a liter of a solution of salts, named as 25% SW and containing (g L−1): NaBr 0.65, NaHCO3 0.17, KCl 5, CaCl2 0.72, MgSO4·7H2O 49.5, MgCl2·6H2O 34.6, NaCl 195, 1 g yeast extract (YE), and 20 g of agar were added. Several dilutions (from 10−1 to 10−5) of the original water sample were used to inoculate the plates by a ‘top agar’ technique. Briefly, 0.1 mL of inoculum and 3 mL of 0.7% agar in 25% SW were mixed and the mixture was immediately poured onto plates containing SW 25% and 0.1% YE. After incubation of the plates for 60 days at 37 °C, the 269 colonies grown in the plate inoculated with dilution 10−4 were restreaked in the same medium. DNA from each colony was extracted by boiling and used as a template for two independent PCRs with primers specific for Bacteria and Archaea, respectively, as described above. Colonies yielding PCR products only with either Bacteria- or Archaea-specific primers were identified, respectively, as Bacteria or Archaea. The 16S rRNA genes from the eight bacterial colonies were sequenced, while 29 colonies were randomly selected among the 261 archaeal colonies for ARDRA and sequencing of selected clones from each ARDRA pattern as described above.
Nucleotide sequence accession numbers
The 16S rRNA gene sequences determined in this study have been deposited in the GenBank database under accession numbers EF459702–EF459730.
Bacterial and archaeal communities from the water samples were compared by DGGE analysis of PCR amplified 16S rRNA gene fragments, as shown in Fig. 1. Selected bands were cut from the gel, reamplified, and sequenced (Table 1). Bands that could be successfully sequenced are labeled with arrows. All the bacterial bands (labeled with a number followed by ‘B’) analyzed corresponded to members of the Bacteroidetes, most of them (bands 2B, 4B, 7B, and 8B) to phylotypes among the Salinibacter cluster. Band 1B was related to environmental sequence NCh63AT12 retrieved from Lake Tebenquiche in the Salar de Atacama in Northern Chile (C. Demergasso et al., unpublished data), with no match to cultured bacteria. Band 4B was related to a sequence retrieved from the athalassohaline Lake Chaka, in Northwestern China (Jiang et al., 2006), with Salinibacter ruber Pola 18 as the closest cultured relative. Band 8B had the highest similarity to the uncultured Salinibacter phylotype EHB-2 (Antón et al., 2000). Three out of the eight archaeal (number plus ‘A’) bands analyzed (i.e. 4A, 7A and 8A) were related to H. walsbyi DSM 16790, ‘the square archaeon’ recently isolated from Spanish and Australian salterns (Burns et al., 2007). Band 11A was distantly related to a group (ss057) of environmental sequences retrieved from soil samples taken along a salinity gradient (Walsh et al., 2005), with Halorubrum sp. as the closest cultured relative. Band 12A was related to the environmental sequence A54 obtained from a Permo-Triassic rock salt (Radax et al., 2001) and also to ss057sequences; Halorubrum sacharovorum was the closest cultured relative. Band 14A was related to the haloarchaeon CSW2.24.4, isolated from an Australian crystallizer pond, that branched just outside the Square Haloarchaea of the Walsby (SHOW) group. In fact, the cultured archaeon closest to band DGGE14A was H. walsbyi, with 92% similarity. Sequences related to CSW 2.24.4 were also retrieved from a Spanish coastal multi-pond saltern (Legault et al., 2006). Bands 16A and 17A were very closely related to Halorubrum orientale and to the haloarchaeon CSW5.28.5, isolated from the above-mentioned Australian crystallizer pond (Burns et al., 2004). Sequences very closely related to CSW5.28.5 have also been retrieved from crystallizers of an Adriatic solar saltern (Pašic et al., 2005). Because the number of bands in the DGGE for sample July W appeared to be higher than in the rest of the samples, it was chosen for a more detailed analysis.
Table 1. DGGE bands and their closest matches in GenBank
Bacteria and Archaea 16S rRNA gene clone libraries from sample July W
A total of 200 clones (100 for Archaea and 100 for Bacteria) were analyzed by ARDRA, which yielded a total of nine different patterns for Bacteria and nine for Archaea. At least one clone per restriction pattern was chosen for complete sequencing. Using the chimera sequence detection softwares described in Ashelford et al. (2005, 2006), two ARDRA groups from the archaeal library (corresponding to 13 clones) and none for the bacterial were identified as possible chimeric sequences and were thus eliminated from the analysis. Table 2 shows the best match with the sequences in databases obtained by blast analysis of the selected clones. Most (82%) of the bacterial clones were related to different uncultured members of the Bacteroidetes. For all these sequences, except 4-01B, S. ruber was the closest cultured relative. Sequence 4-01B did not match with any cultured bacterium, although it was affiliated to the Bacteroidetes group. In addition, 15 clones were related to Haloanaerobacter lacunarum, which was isolated from hypersaline Lake Chokrak on the Kerk Peninsula (Zhilina et al., 1992). Finally, three bacterial clones were very closely related to Acinetobacter sp. TUT1001, a proteobacterium isolated from a household biowaste composting reactor (Hiraishi et al., 2003), as well as to 16S rRNA gene sequences obtained from oil- or phenol-degrading bacteria.
Table 2. Bacteria and Archaea 16S rRNA gene clones analyzed by ARDRA and identification of the almost complete sequence of selected clones
Many (69%) archaeal clones were very closely related (more than 99% similarity) to H. walsbyi DSM 16790. The rest of the sequences had their closest matches to uncultured haloarchaea. ARDRA pattern IIA, represented by clone 1-39 A, was related to a sequence (WN-FW-186) retrieved from alkaline hypersaline lakes of the Wadi An Natrun in Egypt (N.M. Mesbah et al., unpublished data). The closest cultured organism was Halosimplex carlbadense, a halophilic archaeon isolated from unsterilized salt crystals taken from the 250-million-year-old Salado formation in southeastern New Mexico (Vreeland et al., 2002). Sequences related to H. carlbadense have also been recovered in 16S rRNA gene libraries obtained from a crystallizer pond in a solar saltern from Alicante, Spain (Legault et al., 2006). Ten archaeal clones (ARDRA pattern IVA) were related to the uncultured haloarchaeon MSP1, an archaeal phylotype detected in an East African alkaline saltern (Grant et al., 1999), and to Halorhabdus utahensis, an extremely halophilic archaeon isolated from the Great Salt Lake in Utah (Wainøet al., 2000). Finally, eight clones (ARDRA pattern VIIA) were related to the halophilic archaeon CSW1.15.5, isolated from an Australian crystallizer pond (Burns et al., 2004), which belongs to the novel ADL group that had been detected previously only in the Antarctic Deep Lake (Bowman et al., 2000).
FISH analysis of sample July W indicated that the prokaryotic community was dominated by Archaea (58% of the total counts, three-fourths of the community detectable by FISH), although Bacteria were also present in high proportions (21.7% of the total counts). DAPI counts of the Tuz Lake samples were 1.38 × 107 cells mL−1. Square cells and curved rods, the characteristic morphologies of H. walsbyi and S. ruber, respectively (Antón et al., 1999, 2000; Maturrano et al., 2006), were observed in the samples.
Isolation of microorganisms
In an attempt to isolate the new phylotypes of Salinibacter found in the clone library, a water sample was used to inoculate the medium that had been used previously for the isolation of S. ruber (Antón et al., 2000). Out of the 269 colonies obtained, eight were Bacteria and 261 were Archaea. All the Bacteria isolates had the same ARDRA pattern that, as revealed by sequencing of two representatives (7-45 and 7-46 in Fig. 2), corresponded to S. ruber EHB-1. Twenty-nine archaeal colonies were randomly chosen and analyzed by ARDRA, which yielded two different patterns. The most frequent pattern (25 out of the 29 analyzed colonies) was represented by sequence 8-49 (see Fig. 3) and was very closely related (99% identity) to Haloarcula hispanica strain HLR4, isolated from a salt mountain in Taiwan (M. Lai & Y. You, unpublished data). The other ARDRA pattern was represented by sequence 8-38 (Fig. 3), which was found to be related (97%) to archaeon 4R (M. Aponte et al., unpublished data), isolated from salted anchovies in Salerno, Italy. The closest cultured relative was ‘Halorubrum californiensis’ (P.T. Pesenti et al., unpublished data), with 92% nucleotide identity in the 16S rRNA gene.
Comparison of clones, isolates, and DGGE patterns
The phylogenetic affiliations of clones and isolates were studied by inferring trees for Archaea and Bacteria (Figs 2 and 3, respectively), using only almost complete 16S rRNA gene sequences. DGGE band sequences were added to the trees by parsimony.
For Bacteria, clone library analysis gave a larger picture of diversity than DGGE because all the DGGE bands had their counterparts in the library, while many clones were not related to DGGE bands. The culture approach we used did not allow for the isolation of the more abundant members of the community detected by cloning since only some phylotypes of S. ruber related to cultured bacteria could be retrieved. Although, according to the molecular data, Salinibacter spp. appear to be an important component of the bacterial assemblage, the isolates we obtained were not very closely related to the clones and DGGE sequences. Also, a ‘Tuz Lake Salinibacter spp. group’ was found to be clearly different from the previously Salinibacter described phylotypes EHB-1 and EHB-2. Members of this group could not be retrieved by cultivation from the Tuz lake water samples. This Tuz Lake cluster is separated from the EHB-2 group with a puzzle quartet value of 97%, higher than the value that supports the separation between clusters EHB-1 and EHB-2. This group presents the following signature sequence in positions 1015–1030 (E. coli numbering): 5′-TCCGCAAGGACCCC-3′.
Both molecular approaches revealed that haloarchaea belonging to the SHOW (Square Haloarchaea of Walsby) were a very important part of the microbial community. DGGE bands 4, 7, and 8 were identical or almost identical (above 99% in the complete length of the DGGE band, positions 436–801 in the clones) to clones 1-13A, 3-3A, 3-09, and 3-46A. The other DGGE bands, affiliating within the Halorubrum sp. cluster, did not have matches with the analyzed archaeal clones. However, there were some sequences (1-39A and 3-07A, corresponding to 20% of the clones) that represented new archaeal groups and did not have counterparts in the DGGE analysis. The archaeal diversity recovered by plating on 25% total salts 0.1% yeast extract plates was, as expected, quite low, although we were able to culture a completely new archaeon, corresponding to sequence 8-49 (Fig. 3), whose phenotype is presently being characterized.
The composition of the Tuz Lake microbiota resembles in many aspects that of crystallizers from coastal solar salterns both in the abundance and the diversity of prokaryotes. The total counts in Tuz Lake are within the range reported for this kind of environments (Elevi Bardavid et al., 2008). The proportion of Bacteria in Tuz Lake water is one of the highest reported for hypersaline environments (although there are only very few studies providing these kind of data: Antón et al., 2000; Rosselló-Mora et al., 2003; Maturrano et al., 2006).
DGGE and 16S rRNA gene libraries indicate that Haloquadratum spp. and members of the Bacteroidetes (most of them related to Salinibacter) are presumably the main components of the microbiota in Tuz Lake, as reported in some crystallizers (Elevi Bardavid et al., 2008). The presence of archaeal phylotypes related to sequences recently retrieved from an Australian crystallizer (Burns et al., 2004) or Adriatic salterns is also noteworthy (Pašic et al., 2005). The tolerance of H. walsbyi to high solar irradiance and extremely high magnesium concentrations (Bolhuis et al., 2006) has been used to explain the worldwide success of this archaeon. Indeed, representatives of this phylum have been found in hypersaline environments of different characteristics around the world including Andean salterns as well as Australian and Mediterranean coastal salterns, and now an inland hypersaline environment in Western Asia.
Members of the Bacteroidetes are a very important part of the bacterial community in Tuz Lake. Although not present in all the salterns studied around the world (Maturrano et al., 2006), members of this phylum have been found in coastal salterns (Antón et al., 2000; Benlloch et al., 2002), in the athalassohaline Lake Chaka (32.5% salinity) in Northwestern China (Jiang et al., 2006), as well as in athalassohaline lakes from Los Andes with a wide range of salt concentrations and compositions (Demergasso et al., 2004) and in an evaporite crust on the bottom of salt pools at the Badwater site in Death Valley, CA (Elevi Bardavid et al., 2007). As these authors pointed out, ‘hypersaline waters (both thalassohaline and athalassohaline) apparently constitute an important environment’ for Bacteroidetes. In order to ascertain whether there is a group of Bacteroidetes phylotypes associated with hypersaline environments, we constructed phylogenetic trees including all the sequences retrieved from these kind of systems but did not find a unique cluster exclusively formed by ‘halophilic’ sequences (Fig. 4). However, a group of sequences retrieved only from extreme environments was found; this group included all the Salinibacter sequences as well as sequences from deep sea and from the hypersaline alkaline Mono Lake and could be considered as a monophyletic halophilic branch within the Bacteroidetes phylum.
This work was supported by grant CGL2006-12714-C02-01 (Spanish Ministry of Science and Education) and Anadolu University Research Foundation No. 041016. M.B. was a student of the EEC Erasmus program.