Unique Communities of Anoxygenic Phototrophic Bacteria in Saline

RESEARCH ARTICLE Unique communities ofanoxygenic phototrophic bacteria in saline lakes of Salar de Atacama (Chile): evidence fora new phylogenetic lineage of phototrophic Gammaproteobacteria from pufLM gene analyses Vera Thiel1, Marcus Tank1, Sven C. Neulinger1, Linda Gehrmann1, Cristina Dorador2 & Johannes F. Imhoff1

1Leibniz-Institut f ¨urMeereswissenschaften (IFM-GEOMAR), Kiel, Germany; and 2Departamento de Acuicultura, Facultad de Recursos del Mar, Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 Universidad de Antofagasta, Antofagasta, Chile

Correspondence: Johannes F. Imhoff, Abstract Leibniz-Institut f ¨urMeereswissenschaften, D¨usternbrooker Weg 20, D-24105 Kiel, Phototrophic bacteria are important primary producers of salt lakes in the Salar de Germany. Tel.: 149 431 600 4450; fax: 149 Atacama and at times form visible mass developments within and on top of the 431 600 4452; e-mail: jimhoff@ifm- lake sediments. The communities of phototrophic bacteria from two of these lakes geomar.de were characterized by molecular genetic approaches using key genes for the biosynthesis of the photosynthetic apparatus in phototrophic purple bacteria Received 6 July 2010; revised 6 August 2010; (pufLM) and in green sulfur bacteria (fmoA). Terminal restriction fragment length accepted 6 August 2010. polymorphism of the pufLM genes indicated high variability of the community Final version published online 24 September composition between the two lakes and subsamples thereof. The communities 2010. were characterized by the dominance of a novel, so far undescribed lineage of

DOI:10.1111/j.1574-6941.2010.00966.x pufLM containing bacteria and the presence of representatives related to known halophilic Chromatiaceae and Ectothiorhodospiraceae. In addition, the presence of Editor: Riks Laanbroek BChl b-containing anoxygenic phototrophic bacteria and of aerobic anoxygenic bacteria was indicated. Green sulfur bacteria were not detected in the environ- Keywords mental samples, although a bacterium related to Prosthecochloris indicum was anoxygenic phototrophic bacteria; Salar de identiﬁed in an enrichment culture. This is the ﬁrst comprehensive description of Atacama; pufL; pufM; fmoA; functional gene. phototrophic bacterial communities in a salt lake of South America made possible only due to the application of the functional pufLM genes.

reviewed in Imhoff (2001). Also, green sulfur bacteria have Introduction been observed in various saline environments mainly based on It has been known for long that the strongly saline environ- microscopic and macroscopic observations (Giani et al., 1989; ment is primarily a domain of prokaryotes and the spectrum Caumette, 1993; Oren, 1993). Species of the genus Prostheco- of eukaryotic species in highly saline biotopes is rather chloris were obtained from marine and saline environments restricted. The dominant primary producers are halophilic and are recognized as halotolerant and moderately halophilic and halotolerant algae and cyanobacteria as well as anoxygenic organisms (Gorlenko, 1970; Imhoff, 2001, 2003; Vila et al., phototrophic bacteria (Imhoff et al., 1979; Truper¨ & Galinski, 2002; Alexander & Imhoff, 2006; Triado-Margarit´ et al., 2010). 1986; Imhoff, 1988, 2001, 2002). A variety of anoxygenic Highly saline lakes in the extremely arid Atacama Desert phototrophic bacteria has been isolated from different hyper- located in northern Chile are characterized by high UV saline habitats, such as marine salterns (Rodriguez-Valera radiation, high salt concentrations and wide diurnal tem- et al., 1985; Caumette et al., 1988, 1991; Caumette, 1993), perature variations. The Salar de Atacama is located at an

MICROBIOLOGY ECOLOGY MICROBIOLOGY alkaline soda lakes in the Egyptian Wadi Natrun (Imhoff & altitude of 2300 m and is the largest evaporitic basin in Chile Truper,¨ 1977, 1981; Imhoff et al., 1978), in Siberia and (2900 km2). It has several permanent hypersaline lakes that Mongolia (Bryantseva et al., 1999, 2000) and from Solar Lake receive waters from the Andes Range (Risacher & Alonso, (Sinai) (Cohen & Krumbein, 1977; Caumette et al., 1997) as 1996). Like other hypersaline environments, the studied

c 2010 Federation of European Microbiological Societies FEMS Microbiol Ecol 74 (2010) 510–522 Published by Blackwell Publishing Ltd. All rights reserved APB of Salar de Atacama using functional genes 511 lakes of the Salar de Atacama (Laguna Chaxa and Laguna associated in a trimeric structure (Fenna et al., 1974). Its Tebenquiche) are inhabited by only a few higher organisms unique occurrence in green sulfur bacteria and the recently such as brine shrimps, some copepods and surrounding described ‘Candidatus Chloracidobacterium thermophilum’ macrophytes (Zu´ niga˜ et al., 1991). Visually, the shallow (Bryant et al., 2007) makes fmoA an appropriate target to Laguna Tebenquiche and Laguna Chaxa exhibit the presence speciﬁcally analyze environmental communities of these of extensive red–purple-colored microbial mats on the sur- bacteria (Alexander & Imhoff, 2006). 13 face of the lake sediments. Based on low d C(HCO3) values of 1.38 for Laguna Chaxa measured in previous studies, biological productivity in these lakes is expected to be high Materials and methods (Boschetti et al., 2007). However, the content of chlorophyll

a was shown to be rather low in previous studies (Demer- Study area Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 gasso et al., 2008), leading to the assumption of a consider- The Salar de Atacama is a closed saline basin within the pre- able impact of anoxygenic phototrophic bacteria on the Andean depression of the Atacama Desert located at 2000300S primary productivity in these habitats. Despite the visual and 6800150W in northern Chile and covers approximately indication, phototrophic bacteria have not been specifically 2900 km2 (Zu´ niga˜ et al., 1991; Demergasso et al., 2004). The studied and almost nothing is known about the diversity Atacama Desert is characterized by extreme aridity and is and composition of the communities of anoxygenic photo- considered to be one of the driest places on earth (Rech trophic bacteria in these lakes. Main studies of the micro- et al., 2006). The average amount of precipitation in the biology of the Salar de Atacama have been focused on the desert and the Salar area reaches o 3 and 25–50 mm year1, cultivable diversity. Heterotrophic strains of moderately respectively. The low precipitation together with an excep- halophilic bacteria have been analyzed by numerical taxon- tionally high evaporation of 1800–3200 mm year1 leads to a omy (Prado et al., 1991; Valderrama et al., 1991) and hyperarid ecosystem (Risacher et al., 2003; Boschetti et al., chemotaxonomic analysis (Marquez et al., 1993) and domi- 2007). Solar radiation is high, especially UV-B radiation, nated the isolation-based studies (Ramos-Cormenzana, which is 20% increased compared with that at sea level 1993; Campos, 1997). The only phototrophic bacteria so (Cabrera et al., 1995). far described in Laguna Tebenquiche (Salar de Atacama) The saline basin of the Salar de Atacama is covered with a were oxygenic cyanobacteria represented by Oscillatoria thick halite crust of several hundred meters (Bobst et al., (Zu´ niga˜ et al., 1991). Recently, the bacterial diversity in 2001). At its edges and in its interior, there are small ponds water samples of Laguna Tebenquiche has been studied by and a number of shallow lakes with high concentrations of ribosomal gene library analysis (Demergasso et al., 2008). salts, which receive streams of fresh subsurface water. However, sequences related to anoxygenic phototrophic Laguna Chaxa and Laguna Tebenquiche (the largest of these bacteria were not recovered, except for a single clone related lakes) are two of these hypersaline lakes receiving freshwater to the aerobic phototrophic purple bacteria of the Roseobac- from the subsurface (Zu´ niga˜ et al., 1991). The pH was only ter clade (Demergasso et al., 2008) with no evidence for any slightly alkaline (Table 1) and despite the shallow character phototrophic potential and activity. of the lakes, low dissolved oxygen concentrations have been In order to specifically study the communities of photo- measured in this (1.2 mg L1) and in previous studies trophic prokaryotes of these habitats, we used molecular (0.6 mg L1,Zu´ niga˜ et al., 1991; Boschetti et al., 2007). genetic analyses with group-specific primers for functional Sodium and chloride are the dominating ions, followed by genes (pufLM, fmoA), which target phototrophic bacterial sulfate (Risacher & Alonso, 1996). communities. Because they represent a physiological group of polyphyletic origin, it is not possible to recover the diversity of phototrophic communities using 16S rRNA Samples gene sequences. The pufLM genes encode for the light (L) and medium (M) subunit of the photosynthetic reaction Sediment samples were taken from two lakes at the Salar de center type II structural proteins of phototrophic purple Atacama, Laguna Tebenquiche and Laguna Chaxa located in bacteria including purple sulfur bacteria, purple nonsulfur the north and in the east of the Salar, respectively (Fig. 1). bacteria and aerobic anoxygenic phototrophic bacteria, as Four sediment samples were taken from Laguna Teben- well as Chloroflexaceae. These genes have been used pre- quiche (23.13S 68.24W; samples SAT1–SAT5) and six from viously to access phototrophic bacteria in environmental Laguna Chaxa (23.29S 68.18W, samples FC1–FC6). The samples (Oz et al., 2005; Hu et al., 2006) and were demon- samples contained differently colored bacterial mats with strated to be suitable phylogenetic markers for purple sulfur supernatant water. They were homogenized and separated bacteria (Tank et al., 2009). The fmoA gene encodes the into aliquots. Samples for DNA extraction were stored at monomer of the FMO protein, which binds Bchl a and is 20 1C until analysis.

Table 1. Samples of Salar de Atacama used in this study, their visual appearance and chemical parameters Chemical data

Sample Description pH Salinity (PSU) Temperature (1C) Redox (mV) FC1 Orange–red salt crust 7.32 o 36.0 34.0 FC2 Purple mat 7.55 o 27.0 48.0 FC3 Greenish surface with a purple layer beneath 7.64 o 32.0 53.0 FC4 Bacterial mat with gas enclosures; black beneath 7.86 193 36.0 67.0 FC5 Salt crust with pink and green layer o o 32.5 o FC6 Pink sand; beneath black sediment 7.99 o 33.7 74.0 SAT1 Red bacterial mat o o o o

SAT2 Pink ﬁlamentous mat with leather skin and gas enclosures 7.00 178 32.0 o Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 SAT3 Orange, leatherlike surface o o o o SAT4 Pink ﬁlamentous bacterial mat with gas enclosures o o o o SAT5 Green mats with gas enclosures 7.96 30 33.5 o

Samples of (Flamingo) Lake Chaxa (FC) were taken on November 24, 2008 at approximate GPS position 23.29S 68.18W; samples of Lake Tebenquiche (SAT) were taken on November 22, 2008 at approximate GPS position 23.13S 68.24W. o, not available.

(a) (b) 0.4

0.2

–0.2

–0.4 Dimension 2 (AU) –0.6

–0.4 –0.2 0 0.2 0.4 –0.6 –0.4 –0.2 0 0.2 0.4 Dimension 1(AU) Fig. 1. Non-metric MDS plot based on T-RFLP fingerprinting data of Laguna Chaxa samples FC 1-6 ( ) and Laguna Tebenquiche samples SAT 1-4 (). Ordination was based on the Bray–Curtis coefficient (a) and the Dice coefficient (aka Sørensen coefficient) (b) derived from the T-RF peak matrix. For clarity, of the three MDS dimensions calculated, only the two with the most conspicuous differences were plotted. The stress value for dimensional downscaling was 0.06 (a) and 0.09 (b), respectively. In the two plots, data points of FC and SAT phototrophic bacterial communities are separated by the axis of dimension 1 (a) and by the main diagonal (upper right to lower left corner) (b), respectively. This indicates that phototrophic bacterial communities of the two sampling locations clearly differed from each other.

DNA extraction DNA for FC samples) and 1 mL of each primer (10 pmol) were used. PCR was carried out using the following condi- DNA was extracted from 250 mg of homogenized sediment tions: initial denaturation (94 1C for 2 min) followed by 35 samples with the PowerSoil DNA isolation Kit (MoBio cycles of primer annealing (40 s at 55 1C), elongation Laboratories Inc.) following the manufacturer’s protocol. (1.5 min at 72 1C) and denaturation (40 s at 94 1C), a final Lysis was performed following the alternative protocol primer annealing (1 min at 42 1C) and a final extension including 2 10 min heating intervals (70 1C) with vortex- (5 min at 72 1C). In PCR for terminal restriction fragment ing before, after and in-between the intervals. length polymorphism (T-RFLP) analysis, the annealing temperature was 56 1C. PCR For amplification of fmoA genes, the primers fmoA_St- Amplification of functional genes was performed using art_mod (50-ATT ATG GCT CTN TTC GGC-30; modified puReTaq Ready-To-Go PCR beads (GE Healthcare) in a from Alexander et al., 2002) and fmoA_1094r (Alexander final volume of 25 mL. For amplification of pufLM genes, the et al., 2002) were used. Amplification of vector inserts was primers 67F and 781R (Tank et al., 2009) were used. 3–5 mL performed with the primers M13f (50-GTA AAA CGA CGG of template (50 ng DNA for SAT samples and 100 ng CCA G-30) and M13r (50-CAG GAA ACA GCT ATG AC-30).

In both cases (fmoA and M13 PCR), an annealing tempera- For phylogenetic analysis, representative clones as well as ture of 50 1C was used. The other PCR conditions were the reference sequences of each phototrophic pufLM-containing same as stated above for pufLM amplification. group including next relatives as determined by BLAST search were used (Supporting Information, Table S1). The evolu- tionary model GTR1I1G used for phylogenetic analyses of Cloning and sequencing nucleotide sequences was determined using the program PCR products were purified by extraction from a 1% TAE MODELGENERATOR, version 0.85 (Keane et al., 2004). Phyloge- agarose gel using the QIAquick Gel Extraction Kit (Qiagen, netic trees were calculated with the PHYML program version Germany). DNA was eluted in 30 mL elution buffer and 3.0 (Guindon & Gascuel, 2003). Chloroflexus aggregans DSM stored at 20 1C until further analysis. DNA was ligated 9485T and Chloroflexus aurantiacus DSM635T served as

into the pCR4-TOPO vector and transformed into One Shot outgroup species. Sequence similarities were calculated as Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 Competent Escherichia coli cells using the TOPO TA Cloning percentage similarity between all aligned sequences used for Kit (Invitrogen). Inserts were amplified as described above phylogenetic calculation using the PHYLIP 3.63 DNADIST pro- using the M13f/M13r primer set. Insert size was checked by gram (Felsenstein, 2004). agarose gel electrophoresis. Sequencing was performed using the BigDye Terminator v1.1 sequencing kit (Applied Biosystems) on a 3730 DNA Richness estimation analyzer (Applied Biosystems) as specified by the manufacturer. Nucleotide sequences with similarities Z98% were grouped Clone inserts were sequenced using the vector-specific primers using MOTHUR (Schloss et al., 2009). The proportion of T3 and T7 (Invitrogen). Sequences obtained in this study were prokaryotic diversity represented by the clone libraries was deposited in the EMBL database (Kulikova et al., 2004) and estimated by rarefaction analysis combined with nonlinear assigned the accession numbers FN813740–FN813767. regression. Rarefaction analysis calculations were performed applying the algorithm described by Hurlbert (1971) with the Phylogenetic analyses program ARAREFACT-WIN (http://www.uga.edu/strata/software. html) after removal of putative chimera. Rarefaction curves Sequences were edited using the SEQMANII program (DNA- were plotted and regressions were performed using the c STAR). Closest relatives were determined by comparison equation y ¼ að1 ebx Þ as published earlier (Thiel et al., with sequences in the National Centre for Biotechnology 2007) where x is the sample size, y is the observed number of Information (NCBI) GenBank database using the Basic phylotypes, a is the number of phylotypes to be expected with Local Alignment Search Tool (BLAST) (Altschul et al., infinite sample size (i.e. total phylotype richness) (Koellner 1990). Sequences posted as possible chimera using the et al., 2004) and b and c are additional curve fitting para- Bellerophon server (http://foo.maths.uq.edu.au/huber/bel meters. SIGMAPLOT v10.0 (Systat Software Inc.) was used for lerophon.pl, Huber et al., 2004) were further checked for plotting and regression analysis. Coverage was calculated as breakpoints and parental clones manually. Sequences with the ratio of the obtained number of phylotypes to the expected similarity to non-pufLM gene sequences as well as putative number of phylotypes a as calculated by the regression. chimeric sequences were removed from the dataset before richness estimation and phylogenetic analysis. The nucleotide sequences of the pufLM genes were T-RFLP fingerprinting converted into amino acid sequences using BIOEDIT version 7.0.1 (Hall, 1999), aligned with the integrated version of Restriction digests CLUSTAL X (Thompson et al., 1997), and manually refined. For phylogenetic analysis, the corresponding pufLM nucleo- PCR products were purified by excision from a 1% agarose tide sequence alignment was used. Replicate sequences gel in TAE buffer, subsequent extraction with the QIAquick (Z98.0% sequences similarity) were grouped for each Gel Extraction Kit (Qiagen), and eluted from spin columns library using MOTHUR (Schloss et al., 2009). Representative with 30 mL elution buffer. The PCR products were digested sequences were chosen manually considering the maximal with three restriction endonucleases AluI, HinfI and MspI sequence length and quality. Only one representative (New England Biolabs). Ten microliters of the purified PCR sequence each was used for phylogenetic analysis. Represen- product was mixed with 3.5 mL of restriction enzyme master tative sequences with nucleotide similarities Z98.0% were mix containing NEBuffer 4 and 10 U (HinfI; AluI) or 20 U identified as belonging to a single phylotype. As suggested by (MspI), respectively, and filled up with DNA-free H2Oto Tank et al. (2009), pufLM clone sequences of Z86% nucleo- obtain a total volume of 25 mL. Restriction reactions were tide similarity were grouped together in clusters considered incubated for 6 h at 37 1C, followed by 20 min at 65 1C, to equivalent to genera. denature the enzyme.

T-RFLP analysis the T-RFLP-derived distance matrices (based on Bray–Cur- tis and Dice coefficients, respectively) with a second ‘dis- Restriction products were purified by isopropanol precipita- TM tance’ matrix listing ‘0’ for any two samples from the same tion and resuspended in 12 mLofHi–Di formamide sampling site, and ‘1’ for any two samples not from the same (Applied Biosystems). For analysis, the resuspended restric- sampling site. Based on the Spearman correlation coeffi- tion product was mixed with 1 mL of GeneScan – 1000 (ROX) cient, it was tested whether sample pairs of the same site had size standard, and denatured at 95 1Cfor5min.Terminal lower T-RFLP-derived distance values than sample pairs restriction fragments (T-RF) signals were detected by capillary from different sites (one-tailed test). Statistical analyses were electrophoresis on an ABI Prism 310 Genetic Analyzer with TM performed using XLSTAT version 2009.6.02 (Addinsoft). POP–6 Polymer in a 30 cm capillary (Applied Biosystems) under the following conditions: injection time 15 s, injection/ electrophoresis voltage 15 kV, gel temperature 60 1Candrun Results Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 time 90 min. Electropherograms were analyzed using the In this study, we used functional genes as molecular targets to program GENESCAN v2.0.2 (Applied Biosystems). specifically analyze anoxygenic phototrophic bacteria inhabiting two hypersaline lakes of the saline evaporate basin, Salar de Atacama. We thereby took advantage of recently established Statistics phylogenetic congruence between the ribosomal genes and genes for specific structural components of the photosynthetic Peak data of T-RFs between 20 and 928 nt and signal apparatus of green sulfur bacteria (fmoA,Alexanderet al., intensities Z50 AU were exported as tabular data from the 2002; Alexander & Imhoff, 2006) and of phototrophic purple genetic analyzer. A data matrix was created from the bacteria (pufLM,Tanket al., 2009). pufLM genes were combined AluI and MspI peak data with samples as columns successfully amplified from 10 out of 11 samples (SAT5 did and peak positions (T-RF lengths) as rows. Data from the not yield pufLM amplificates), while fmoA genes were not HinfI digestion were omitted from the statistics because amplified from any environmental sample. digestion of sample FC6 with this enzyme was unsuccessful. The area under each peak was used as a measure of T-RF pufLM fingerprinting abundance, standardized as percentage of the total peak area as described by Lukow et al. (2000). Because of rounding In order to get an overview of the diversity of pufLM-contain- errors and minor variations in size determination, the ing organisms between and within the different sampling length of defined T-RFs varied among samples. The varia- locations, we used T-RFLP as a genetic fingerprinting method. tion was empirically determined to be 0.3% of the T-RFLP data were visualized by MDS (Fig. 1). According to fragment length, but at least 1 nt. This meant a size MDS, the phototrophic bacterial communities of Laguna variation of 1 nt for fragment lengths from 20 to 499 nt, Chaxa (FC) and Laguna Tebenquiche (SAT) clearly differed 2 nt for fragment lengths from 500 to 833 nt and 3nt from each other. These results were obtained independently of for fragment lengths from 834 to 928 nt. T-RFs were aligned the method used for distance calculation: In the MDS plot within this range from their expected mean length (Lukow based on normalized peak areas (the Bray–Curtis coefficient), et al., 2000). Two distance matrices were derived from the the separator between FC and SAT phototrophic bacterial peak matrix based on (1) the Bray–Curtis coefficient, which communities was the axis of dimension 1 (Fig. 1a); MDS compares relative abundances of T-RFs shared between any based on peak presence–absence data (the Dice coefficient) two samples and (2) the Dice coefficient (aka Sørensen separated these communities along the main diagonal (upper coefficient), which is an analogue to the Bray–Curtis coeffi- right to lower left corner) of the plot (Fig. 1b). cient, but only takes into account binary data (i.e. presence A highly significant correlation between phototrophic or absence of T-RFs). Nonmetric multidimensional scaling bacterial community composition and sampling location (MDS) was applied to ordinate samples in three dimensions was also indicated by the Mantel test (P = 0.0028 and 0.0053 according to their distances. The number of three dimen- for the distance matrices based on Bray–Curtis and Dice sions was chosen in order to keep the stress value for coefficients, respectively). dimensional downscaling below the recommended thresh- MDS plots also showed conspicuous distances between old of 0.1 for an ideal preservation of the original distances different bacterial populations of the same location (Fig. 1a between samples (Clarke, 1993). For clarity, of these three and b). This was due to the fact that subsamples of photo- dimensions, only the two with the most conspicuous trophic communities from each lake differed considerably in differences were plotted in the MDS diagrams. both the overall number of peaks and the presence or absence To see whether distances within and between sample of specific T-RFs (Table 2). For example, within Laguna groups were connected to the sampling sites, a Mantel test Chaxa, the number of peaks originating from AluI digestion (Mantel, 1967) was performed. This test compared each of varied between 8 and 74 (median, 19) for different

Table 2. T-RFLP peak statistics Enzyme

Number of peaks found in . . . AluI HinfIÃ MspI All samples (minimum–median–maximum) 6–17–74 7–28–90 4–29–102 All FC samples (minimum–median–maximum) 8–19–74 18–59–90 10–35.5–102 All SAT samples (minimum–median–maximum) 6–12–20 7–17–28 4–12–37 Every sample (percentage of all peaks) 2 (2%) 0 (0%) 0 (0%) 2–9 samples (percentage of all peaks) 41 (40%) 94 (61%) 80 (53%) 1 sample only (percentage of all peaks) 60 (58%) 61 (39%) 71 (47%)

ÃHinfI digest not successful with sample FC6. Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 community samples. Only two (2%) of all T-RFs produced Table 3. Rarefaction analysis and presumed coverage of pufLM clone with AluI were found in every sample, while 60 peaks (58%) libraries occurred only once in a single sample. Similar results were Phylotypes Phylotypes Coverage obtained with the HinfI and MspI digestions (Table 2). Clone library Sequences obtained expected (%) FC2 68 14 19 75.0 FC5 58 8 12 69.0 Phototrophic purple bacteria SAT3 5 1 ND ND Clone sequences were retrieved from 2 selected samples each SAT4 28 5 7 71.9 of Laguna Chaxa (FC2, FC5) and Laguna Tebenquiche (SAT3, ND, not determinable. SAT4). In total, four clone libraries of pufLM genes were constructed and 159 clone sequences grouped into 25 different phylotypes were analyzed. The number of phylotypes differed lineage phylotypes (Novel-6, Table 4) was present in both considerably between the clone libraries, with values between lakes, Laguna Chaxa and Laguna Tebenquiche (Table 4). 1 (SAT3) and 14 (FC2) (Table 3, Fig. 3). Rarefaction analyses Others dominated individual clone libraries (i.e. Novel-1, showed relations of detected to predicted numbers of phylo- FC2, 43%, Fig. 3). types of 4 69% for the different clone libraries (Table 3). Nearly one-third (28%) of the phylotypes inhabiting the Phylogenetic analysis resulted in a tree topology with a clear Salar de Atacama were identified as being closely related to separation between the three subclasses, Gamma-, Beta-and genera of the salt-loving Ectothiorhodospiraceae as well as to Alphaproteobacteria, and the outgroup (Fig. 2). The different the halophilic Chromatiaceae genera Halochromatium and lineages within these subclasses (e.g. Chromatiaceae, Ectothio- Thiohalocapsa (Table 4). Of the three different members of rhodospira, Halorhodospira,BChlb-containing bacteria and the Ectothiorhodospiraceae present in the Salar de Atacama, the aerobic Gammaproteobacteria) were supported by high the Ectothiorhodospira variabilis-related phylotype was ob- bootstrap values (Fig. 2, data for Alphaproteobacteria are not tained from both Laguna Tebenquiche samples, while rela- shown). Based on this phylogenetic analysis, all but one tives of Ectothiorhodospira mobilis and of H. halophila were phylotype were assigned to the anoxygenic phototrophic present in the Laguna Chaxa sample FC5 only. The Halo- Gammaproteobacteria (Fig. 2). chromatium salexigens-related phylotype dominated the The composition of the clone libraries demonstrated a Laguna Tebenquiche sample SAT4 clone library and was highly diverse and variable community of anoxygenic obtained from both lakes. Thiohalocapsa marina-related phototrophic bacteria in the Salar de Atacama. Each of the bacteria were only present in the Laguna Chaxa sample FC2. subsamples consisted of a distinct phototrophic community Five phylotypes obtained from Laguna Chaxa samples and the presence of most of the phylotypes was restricted to showed affiliations to the Bchl b-containing purple sulfur only one of the samples. Only three phylotypes bacteria of the Thiococcus/Thioflavicoccus group, but shared were retrieved from multiple samples (Halochromatium-1, sequence similarities of only 80% with them (Fig. 2, Ectothiorhodospira-1 and Novel-6, Fig. 2, Table 4). Table 4). For three of them, phylogenetic analysis (Fig. 2) Almost half of the phylotypes (48%) could not be and signature amino acids (data not shown) clearly sup- assigned to any of the known groups and genera but formed ported affiliation to the group of BChl b-containing photo- a separate, highly supported monophyletic lineage within trophic Gammaproteobacteria (Tank et al., 2009). The other the Gammaproteobacteria (Novel lineage in Fig. 2). This two phylotypes clustered with the Thiococcus/Thioflavicoccus novel lineage is highly diverse (11 phylotypes, Novel-1–11, group more distantly and do not contain signature amino Fig. 2) and prevalent in the Salar de Atacama. Representa- acid insertions and, therefore, were not assigned to the tives were observed in all but one sample. One of the novel group of BChl b-containing Gammaproteobacteria (Fig. 2).

FEMS Microbiol Ecol 74 (2010) 510–522 c 2010 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 . et al a Z50 are V. Thiel (2010) 510–522 74

Gammaproteobacteria Betaproteobacteria Alphaproteobacteri FEMS Microbiol Ecol groups without any sequences obtained in this study as Bchl b-containing bacteria Aerobic phototrophic Gammaproteobacteria Ectothiorhodospiraceae Novel lineage Chromatiaceae sequences. Sequences obtained in this study are printed in bold. Behind the Alphaproteobacteria pufLM , FN257144 FN257159 FN257161 , , FN257152 , , NZ-DS999405 , FN257133 - , , FN257153 , , , FN257141 ATCC31751, AF018955 DSM2111, FN257157 FN257170 ATCC51935, FN257156 FN257173 FN257171 , , , , proposed, AM944100 , FN257151 , AAOA01000014 ATCC700959 , D50650 , FN257162 WN22 , CP000544 DSM4395 , DSM3802 DSM227, FN257142 DSM226 DSM18859 FN257160 DSM228, FN257143 DSM5161 , JA319 , FN257154 KT71 SL1 DSM237, FN257158 H . . DSM6210 MTK6IM088 MTK5IM027 ML1, AY177752 Chromatiaceae-1 (43), clone FC5-1 Chromatiaceae-2 (2), clone FC5-2 . . . DFL-11, EQ973121 -3 (4), clone FC2-9 spp. spp. spp. -1 (2), clone FC2-13 -2 (1), clone FC2-14 - sp. MTK2IM039 b sp . . sp. MTK1IM127, FN257175 sp. JA142 b b b clone . . MTRDDF079 . spp spp BChl spp 100 BChl BChl . . . Marine Gammaproteobacterium HTCC2080, AAVV01000005 Thioflavicoccus mobilis Rhodoferax fermentans Thiococcus pfennigii Thiococcus pfennigii Thiococcus pfennigii spp. Chromatium weissei Novel-1 (29), clone FC2-2 Novel-2 (1), clone FC2-1 Novel-3 (2), clone FC5-4 Novel-6 (1), clone FC5-3 Novel-6 (1), clone SAT4-1 Novel-4 (2), clone SAT4-2 Novel-6 (1), clone FC5-3 Novel-6 (1), clone SAT4-1 Gammaproteobacterium NOR5-3 spp Novel-5 (2), clone FC2-7 Novel-10 (1), clone FC5-7 Allochromatium Halochromatium-1 (19), clone SAT4-4 Halochromatium-1 (19), clone SAT4-4 Halochromatium-1 (3), clone FC2-5 Novel-9 (1), clone FC2-10 Halochromatium-1 (3), clone FC2-5 sequences used as outgroup were withdrawn from the tree graphically. All database sequences used for phylogenetic Novel-7 (4), clone FC2-4 Novel-8 (2), clone SAT4-3 Thiobaca sp. Thiocapsa Halochromatium roseum Ectothiorhodospira shaposhnikovii Thiorhodococcus Halochromatium Ectothiorhodospira shaposhnikovii Ectothiorhodospira haloalkaliphila Ectothiorhodospira variabilis Congregibacter litoralis Thiocystis Ectothiorhodospira-1 (5), clone SAT3-1 Halochromatium Ectothiorhodospira-1 (5), clone SAT3-1 Halochromatium salexigens Ectothiorhodospira-1 (4), clone SAT4-5 Halochromatium Halochromatium Gammaproteobacterium NOR51-B, NZ-DS999411 Ectothiorhodospira-1 (4), clone SAT4-5 -1 (4), clone FC5-8 100 100 71 Thiolamprovum pedioforme Halorhodospira-1 (2), clone FC5-6 Halorhodospira halophila Labrenzia alexandrii Novel-11 (10), clone FC2-6 α 94 Ectothiorhodospira imhoffii Halorhodospira halophila 5 5 Ectothiorhodospira-2 (3), clone FC5-5 Thiohalocapsa-1 (7), clone FC2-3 100 100 100 AAPGPB-1 (1), cloneAAPGPB-1 FC2-12 Lamprocystis purpurea Marichromatium Ectothiorhodospira mobilis 96 Thiohalocapsa halophila AAPGPB-2 (1), cloneAAPGPB-2 FC2-8 AAPGPB-3 (2), cloneAAPGPB-3 FC2-11 100 100 100 100 Thiohalocapsa marina 100 96 Uncultured bacterium clone FJ498863 CEHL-20-WP45, 100 100 Uncultured bacterium clone FJ498847 CEHL-20-WP4, Rhodospirillales Roseateles 100 60 100 Rubrivivax 98 97 100 100 100 79 100 100 97 100 75 100 46 94 100 56 100 67 100 100 73 0.10 98 100 100 100 3 3 93 Chloroﬂexaceae Partly condensed maximum likelihood phylogenetic tree of 2010 Federation of European Microbiological Societies c Published by Blackwell Publishing Ltd. All rights reserved Fig. 2. phylotype name, the number of sequencesonly represented one is of given the in samples. parenthesesshown Only followed at by three the the phylotypes nodes. representative were The clone. obtainedwell Most scale from phylotypes bar as more were indicates the than present 0.1 one in change samplecalculation per (highlighted are nucleotide. listed in in shaded Table boxes). S1. Bootstrap values 516 APB of Salar de Atacama using functional genes 517

Table 4. Cluster of pufLM clone sequences obtained from this study and its afﬁliation regarding BLAST search Representative sequence Afﬁliation and next relatives Phylotypes (no. of clones) Lake Halochromatium Hch. salexigens (86%) Halochromatium-1 FC2-5 (3) Chaxa SAT4-4 (19) Tebenquiche Thiohalocapsa Thc. halophila (96%) Thiohalocapsa-1 FC2-3 (7) Chaxa BChl b-containing bacteria (uncertain) BChl b-1 FC2-13 (2) Chaxa Thiococcus pfennigiiT (79–80%) BChl b-2 FC2-14 (1) Chaxa BChl b-3 FC2-9 (4) Chaxa ChromatiaceaeÃ Chromatiaceae-1 FC5-1 (43) Chaxa Chromatiaceae-2 FC5-2 (2) Chaxa

Ectothiorhodospira Ectothiorhodospira-1 SAT3-1 (5) Tebenquiche Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 Ect. varabilis (92%) SAT4-5 (4) Tebenquiche Ectothiorhodospira Ect. mobilis (97%) Ectothiorhodospira-2 FC5-5 (3) Chaxa Halorhodospira Hlr. halophila (85%) Halorhodospira-1 FC5-6 (2) Chaxa Novel lineageÃ Novel-1 FC2_2 (29) Chaxa Novel-2 FC2_1 (1) Chaxa Novel-3 FC5_4 (2) Chaxa Novel-4 SAT4_2 (2) Tebenquiche Novel-5 FC2_7 (2) Chaxa Novel-6 FC5_3 (1) Chaxa Novel-5 SAT4_1 (1) Novel-7 FC2-7 (2) Tebenquiche Novel-8 FC2_4 (4) Chaxa SAT4_3 (2) Tebenquiche Novel-9 FC2_10 (1) Chaxa Novel-10 FC5_7 (1) Chaxa Novel-11 FC2_6 (10) Chaxa Aerobic anoxygenic phototrophic AAPGPB-1 FC2-8 (1) Chaxa Gammaproteobacteria (uncertain) Tibetan AAPGPB-2 FC2-11 (2) Chaxa saline lake clones CEHL-20-WP4/-WP45 (84–86%) AAPGPB-3 FC2-12 (1) Chaxa Alphaproteobacteria (uncertain) Rubrivivax gelatinosus (77%) a-1 FC5-8 (4) Chaxa

ÃAll next relatives are below 75% sequence similarity.

FC2 FC5 Halochromatium-1 Thiohalocapsa-1 BChl b-1 BChl b-2 BChl b-3 Chromatiaceae-1 Chromatiaceae-2 Ectothiorhodospira-1 Ectothiorhodospira-2 Halorhodospira-1 Novel-1 Novel-2 Novel-3 SAT3 SAT4 Novel-4 Novel-5 Novel-6 Novel-7 Novel-8 Novel-9 Novel-10 Novel-11 AAPGPB-1 Fig. 3. Relative abundance of pufLM phylotypes AAPGPB-2 in the environmental clone libraries. AAPGPB-3 Alpha-1

Additionally, three phylotypes from Laguna Chaxa sam- pufLM and 16S rRNA gene phylogenies of the purple sulfur ple FC2 clustered with aerobic phototrophic Gammaproteo- bacteria (Tank et al., 2009) and the aerobic Gammaproteo- bacteria (Table 4). They showed the closest relation bacteria (M. Tank, unpublished data). Accordingly, a large (84–86%) to pufM clone sequences retrieved from a hyper- number of pufLM phylotypes were afﬁliated with the saline lake in Tibet (Jiang et al., 2010). Because of the low Gammaproteobacteria, a diverse group of a novel monophy- sequence similarities to pufLM sequences of isolated aerobic letic lineage of anoxygenic phototrophic Gammaproteobac- anoxygenic phototrophic Gammaproteobacteria available at teria as well as related to Alphaproteobacteria. the NCBI database (maximum 76%) they cannot reliably be The identiﬁcation of a novel and so far unknown group of assigned to that phylogenetic group. phototrophic Gammaproteobacteria, which inhabits and Only one single clone was found to phylogenetically possibly dominates the extreme habitats of the Salar de

cluster with the Alphaproteobacteria. However, due to low Atacama, once more indicates that our knowledge of the Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 similarity values of 73% to its next phylogenetic relative, diversity of anoxygenic phototrophic bacteria is far from Labrenzia alexandrii, and of o 77% to all known sequences, complete. This novel lineage is represented by a phylogen- this phylotype could not be reliably assigned phylogeneti- etically diverse number of phylotypes possibly reflecting cally (Table 4). several species and even genera (differentiated at a 86% pufLM sequence similarity level as proposed by Tank et al., Phototrophic green sulfur bacteria 2009) and may be specifically adapted to Chilean salt lakes, similar to a group of Bacteroidetes found in Laguna Teben- Although fmoA genes could be amplified from enrichment quiche (Demergasso et al., 2008; Dorador et al., 2009). cultures of FC2 (Pfennigs media with 5%, 7.5% and 10% The identification of moderately and extreme halophilic NaCl, respectively, data not shown), they were not detected in Chromatiaceae and Ectothiorhodospiraceae corresponds to the environmental samples. From 96 clone sequences of the the hypersaline character of the lakes (Imhoff, 2001). The enrichment, only one phylotype was obtained. BLAST search next relatives of the halophilic Chromatiaceae, Halochroma- and phylogenetic analysis revealed close relationship to Pros- tium salexigens and Thiohalocapsa halophila as well as thecochloris indicum strain 2 K (due to the lack of phylogenetic Ectothiorhodospira species are well known as halophilic information formerly treated as Prosthecochloris aestuarii 2K) bacteria tolerating salinities up to 20%, while Halochroma- with fmoA nucleotide sequence similarity of 98.2%. tium halophila is regarded as extremely halophilic, with salt optima between 13% and 23% NaCl (Caumette et al., 1988, Discussion 1991, 1994; Imhoff, 1992, 1993; Gorlenko et al., 2009). The hypersaline lakes in the Atacama Desert represent Although most Ectothiorhodospira species were isolated unique and extreme habitats clearly dominated by various from alkaline habitats, others are known to tolerate pH forms of microbial life. Microbiological studies have been values as low as 7.5 (Gorlenko et al., 2009) like the only performed only occasionally and almost nothing is known slightly alkaline pH values prevalent at the Salar de Atacama about the phototrophic microbial communities in these (Table 1). The mainly moderately halophilic characteristics habitats. Highly diverse communities of anoxygenic photo- of the identified phototrophic purple sulfur bacteria in the trophic bacteria were present in both Laguna Chaxa and Salar de Atacama is consistent with previous isolation-based Laguna Tebenquiche, but significant differences were found microbiological studies, which were dominated by moder- between the two lakes according to pufLM T-RFLP analysis ately halophilic heterotrophic bacteria while extreme (Fig. 1). In addition, clear differences were also obvious halophiles were not isolated (Ramos-Cormenzana, 1993). between the subsamples of each of the two lakes. Both However, in contrast to those studies, our data also indicate pufLM T-RFLP and clone libraries revealed a high hetero- the presence of extremely phototrophic bacteria related to geneity of the anoxygenic phototrophic bacterial commu- Halorhodospira halophila reflecting variations in habitat salt nities within the different samples of both lakes. Highly concentrations known from the literature (Demergasso patchy and variable conditions were observed in microhabi- et al., 2008) and measured in this study (Table 1). tats including varying salinity, pH, temperature and redox It was not surprising to find phylogenetic evidence for the values (Table 1, Demergasso et al., 2008). These variable presence of novel phototrophic bacteria possibly related to conditions apparently have major impact on the bacterial the BChl b-containing Chromatiaceae group of Thiococcus/ communities of the phototrophic bacteria leading to patchi- Thioflavicoccus species, which was also pointed out by in ness observed as different colors and textures of the micro- vivo absorption spectra for Laguna Chaxa (data not shown). bial mats as well as different molecular patterns in this study. Members of this group have been isolated from freshwater The clear delineation of the phototrophic purple bacteria and marine habitats as well as from alkaline soda lakes from hypersaline lakes into phylogenetic groups in this (Eimhjellen et al., 1967; Bryantseva et al., 2000; Imhoff & study is based on the recently demonstrated congruence of Pfennig, 2001). BChl b has an absorption maximum in the

c 2010 Federation of European Microbiological Societies FEMS Microbiol Ecol 74 (2010) 510–522 Published by Blackwell Publishing Ltd. All rights reserved APB of Salar de Atacama using functional genes 519 infrared part of the spectrum and bacteria with this pigment phototrophic bacteria in the Salar de Atacama. The low are particularly well adapted to sediments covered by water similarity values further suggest that special forms of so far not permanently or only by a thin layer. The lakes of the Salar not recognized aerobic alphaproteobacterial phototrophic de Atacama characterized by shallow water bodies and parti- bacteria have adapted to the habitats of the Salar de Atacama. cularly wet sediments around the lakes might thus be pre- Green sulfur bacteria are quite likely a minor component of ferred habitats of BChl b-containing phototrophic bacteria. the phototrophic bacterial community in the Salar de Ataca- Unfortunately, to date, a pufLM sequence of the only moder- ma, but may develop under appropriate conditions, as demon- ately halophilic BChl b-containing Chromatiaceae known is strated by enrichment cultures (data not shown). However, a unavailable for phylogenetic comparison. Thioalkalicoccus natural abundance below the detection limit of molecular limnaeus was isolated from Siberian soda lakes and tolerates methods is indicated by the lack of ampliﬁcation of fmoA

salt concentrations of up to 6% (Bryantseva et al., 2000; genes. Species of the genus Prosthecochloris are known as salt- Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 Imhoff & Pfennig, 2001). Because of the phylogenetic con- dependent organisms from marine and hypersaline habitats gruency between the 16S rRNA and the pufLM gene in PSB and have been obtained by isolation and molecular methods (Tank et al., 2009), the three pufLM phylotypes from Laguna from different saline habitats (Gorlenko, 1970; Vila et al., 2002; Chaxa possibly represent Thioalkalicoccus limnaeus or relatives Alexander & Imhoff, 2006; Imhoff & Thiel, 2010; Triado-´ and contain BChl b as supported by signature amino acids. Margarit et al., 2010). Detection of the halotolerant Prosteco- ThepresenceofaerobicphototrophicAlpha-andGamma- chloris indica (7% salt tolerance) in the Salar de Atacama proteobacteria in Laguna Chaxa indicates the coexistence of correlates with the recognition of Prosthecochloris species being both aerobic and anaerobic anoxygenic phototrophic bacteria common as representatives of the Chlorobiaceae in saline and in the bacterial mats covering the shallow lake sediments and hypersaline habitats (Alexander et al., 2002; Triado-Margarit´ in its saline waters. Possibly aerobic anoxygenic phototrophic et al., 2010). At the same time, they demonstrate low abun- Roseobacter-like bacteria were also detected by ribosomal genes dance and a possibly quite restricted phylogenetic diversity of retrieved from Laguna Tebenquiche (Demergasso et al., 2008). this group in the salt lakes of the Salar de Atacama. However, using ribosomal genes as target led to the detection of only one single phylotype of the Roseobacter clade, a group comprising a number of heterotrophic Alphaproteobacteria,of Conclusions which only some have photosynthetic potential. Therefore, it The studied hypersaline lakes in the Salar de Atacama are is not clear whether the identified sequence belongs to a unique habitats harboring highly diverse anoxygenic photo- photosynthetically active bacterium. In contrast, the molecu- trophic bacterial communities, including numerous still un- lar approach using photosynthesis genes specifically targeting known pufLM containing anoxygenic phototrophic bacteria. anoxygenic phototrophic bacteria demonstrated a rich diver- The communities have a significantly different composition in sity of anoxygenic phototrophic Proteobacteria, including thetwolakesaswellasinsubsamplesofeachofthelakes. aerobic anoxygenic phototrophic Gammaproteobacteria in Highly variable conditions such as water variability, salt con- the Salar de Atacama. Only few aerobic anoxygenic photo- centrations and light conditions apparently shape the commu- trophic Gammaproteobacteria have been isolated and only one nity structure in microhabitats of the salt lakes. The high described species (Congregibacter litoralis)isknownsofar. number of as-yet-uncultured and -unidentified pufLM phylo- They have been shown to be abundant in marine waters and types retrieved emphasizes the uniqueness of the studied area sediments, but were also found in hypersaline habitats as as well as the need for further studies on phototrophic bacteria, indicated by 16S rRNA gene sequences (Yan et al., 2009), including culture-dependent approaches. Further, this study including a Chilean salt lake geographically close to the Salar demonstrated the great power of molecular methods targeting de Atacama (Salar de Ascotan,´ EMBL entry no. EF632657, the photosynthesis-related functional genes pufLM in studying C.Dorador,I.Vila,K.-P.Witzel&J.F.Imhoff,unpublished natural communities of anoxygenic phototrophic Proteobac- data). Recently, one isolate was recovered from a Canadian teria in environmental samples. hypersaline spring system (Csotonyi et al., 2008). Unfortu- nately, a pufLM sequence of the halophilic isolate is unavailable for comparison and only few pufLM sequences of this Acknowledgements group are retrievable from the databases. Nonetheless, pufM sequences related to those found in the Salar de Atacama were We gratefully acknowledge Andrea Gartner¨ for assistance also present in a Tibetan hypersaline lake (Jiang et al., 2010), with field sampling and Frank Lappe for technical support indicating a possible preference for high-salt conditions in in the lab. This study was partially funded by the German these bacteria (Fig. 2). Academic Exchange Services (DAAD) and its Chilean Part- Phototrophic members of the Alphaproteobacteria appar- ner, the National Commission for Scientific & Technological ently represent a rather insignificant group of anoxygenic Research (CONICYT) (funding no. 2007-224) as well as by

FEMS Microbiol Ecol 74 (2010) 510–522 c 2010 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 520 V. Thiel et al. the National Fund for Scientiﬁc & Technological Develop- Clarke KR (1993) Non-parametric multivariate analysis of ment (FONDECYT) (funding no. 11080228). changes in community structure. Aust J Ecol 18: 117–143. Cohen Y & Krumbein WE (1977) Solar Lake (Sinai). 2. Distribution of photosynthetic microorganisms and primary References production. Limnol Oceanogr 22: 609–620. Csotonyi JT, Swiderski J, Stackebrandt E & Yurkov VV (2008) Novel Alexander B & Imhoff JF (2006) Communities of green sulfur halophilic aerobic anoxygenic phototrophs from a Canadian bacteria in marine and saline habitats analyzed by gene hypersaline spring system. Extremophiles 12: 529–539. sequences of 16S rRNA and Fenna–Matthews–Olson protein. Demergasso C, Casamayor E, Chong G, Escudero L & Pedros-´ Int Microbiol 9: 259–266. Alio´ C (2004) Distribution of prokaryotic genetic diversity in Alexander B, Andersen JH, Cox RP & Imhoff JF (2002) Phylogeny athalassohaline lakes of the Atacama Desert, Northern Chile. of green sulfur bacteria on the basis of gene sequences of 16S Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 FEMS Microbiol Ecol 48: 57–69. rRNA and of the Fenna–Matthews–Olson protein. Arch Demergasso C, Escudero L, Casamayor EO, Chong G, Balague V Microbiol 178: 131–140. & Pedros-Alio C (2008) Novelty and spatio-temporal Altschul SF, Gish W, Miller W, Myers EW & Lipman DJ (1990) heterogeneity in the bacterial diversity of hypersaline Lake Basic local alignment search tool. J Mol Biol 215: 403–410. Tebenquiche (Salar de Atacama). Extremophiles 12: 491–504. Bobst AL, Lowenstein TK, Jordan TE, Godfrey LV, Ku TL & Luo S Dorador C, Meneses D, Urtuvia V, Demergasso C, Vila I, Witzel (2001) A 106ka paleoclimate record from drill core of the Salar de Atacama, northern Chile. Palaeogeogr Palaeocl 173: 21–42. KP & Imhoff JF (2009) Diversity of Bacteroidetes in high Boschetti T, Cortecci G, Barbieri M & Mussi M (2007) New and past altitude saline evaporitic basins in northern Chile. J Geophys geochemical data on fresh to brine waters of the Salar de Atacama Res 114: 35–44. and Andean Altiplano, northern Chile. Geoﬂuids 7:33–50. Eimhjellen KE, Steensland H & Traetteberg J (1967) A Thiococcus Bryant DA, Costas AMG, Maresca JA et al. (2007) Candidatus sp. nov. gen., its pigments and internal membrane system. Arch Chloracidobacterium Thermophilum: an aerobic phototrophic Microbiol 59: 82–92. acidobacterium. Science 317: 523–526. Felsenstein J (2004) PHYLIP (Phylogeny Inference Package) version Bryantseva I, Gorlenko VM, Kompantseva EI, Imhoff JF, Suling¨ J 3.6. Department of Genome Sciences, University of & Mityushina L (1999) Thiorhodospira sibirica gen. nov., sp. Washington, Seattle. nov., a new alkaliphilic purple sulfur bacterium from a Fenna RE, Matthews BW, Olson JM & Shaw EK (1974) Structure Siberian soda lake. Int J Syst Bacteriol 49: 697–703. of a bacteriochlorophyll protein from green photosynthetic Bryantseva IA, Gorlenko VM, Kompantseva EI & Imhoff JF bacterium Chlorobium limicola – crystallographic evidence for (2000) Thioalkalicoccus limnaeus gen. nov., sp nov., a new a trimer. J Mol Biol 84: 231–240. alkaliphilic purple sulfur bacterium with bacteriochlorophyll Giani D, Seeler J, Giani L & Krumbein WE (1989) Microbial mats b. Int J Syst Evol Micr 50: 2157–2163. and physicochemistry in a saltern in the Brittany (France) and in a Cabrera S, Bozzo S & Fuenzalida H (1995) Variation in UV laboratory scale saltern model. FEMS Microbiol Ecol 62: 151–162. radiation in Chile. J Photoch Photobio B 28: 137–142. Gorlenko VM (1970) A new phototrophic green sulphur Campos V (1997) Microorganismos de ambientes extremos: bacterium Prosthecochloris aestuarii nov. gen. nov. spec. Z Allg Salar de Atacama, Chile. El Altiplano: Ciencia y Conciencie de Mikrobiol 10: 147–149. los Andes (Gonzales C, ed), pp. 143–147. Editiorial Artegrama, Gorlenko VM, Bryantseva IA, Rabold S, Tourova TP, Rubtsova D, Santiago, Chile. Smirnova E, Thiel V & Imhoff JF (2009) Ectothiorhodospira Caumette P (1993) Ecology and physiology of phototrophic variabilis sp. nov., an alkaliphilic and halophilic purple sulfur bacteria and sulfate reducing bacteria in marine salterns. bacterium from soda lakes. Int J Syst Evol Micr 59: 658–664. Experientia 49: 473–481. Guindon S & Gascuel O (2003) A simple, fast, and accurate Caumette P, Baulaigue R & Matheron R (1988) Characterization algorithm to estimate large phylogenies by maximum of Chromatium salexigens sp. nov., a halophilic Chromatiaceae isolated from Mediterranean salinas. Syst Appl Microbiol 10: likelihood. Syst Biol 52: 696–704. 284–292. Hall TA (1999) BioEdit: a user-friendly biological sequence Caumette P, Baulaigue R & Matheron R (1991) Thiocapsa alignment editor and analysis program for Windows 95/98/ halophila sp. nov, a new halophilic phototrophic purple sulfur NT. Nucl Acid S 41 : 95–98. bacterium. Arch Microbiol 155: 170–176. Hu YH, Du HL, Jiao NZ & Zeng YH (2006) Abundant presence of Caumette P, Matheron R, Raymond N & Relexans JC (1994) the gamma-like proteobacterial pufM gene in oxic seawater. Microbial mats in the hypersaline Ponds of Mediterranean salterns FEMS Microbiol Lett 263: 200–206. (Salins-De-Giraud, France). FEMS Microbiol Ecol 13: 273–286. Huber T, Faulkner G & Hugenholtz P (2004) Bellerophon: a Caumette P, Imhoff JF, Suling¨ J & Matheron R (1997) program to detect chimeric sequences in multiple sequence Chromatium glycolicum sp. nov., a moderately halophilic alignments. Bioinformatics 20: 2317–2319. purple sulfur bacterium that uses glycolate as substrate. Arch Hurlbert S (1971) The nonconcept of species diversity: a critique Microbiol 167: 11–18. and alternative parameters. Ecology 52: 577–586.

Imhoff JF (1988) Halophilic phototrophic bacteria. Halophilic planted with transgenic and non-transgenic potato plants. Bacteria (Rodriguez-Valera F, ed), pp. 85–108. CRC Press, FEMS Microbiol Ecol 32: 241–247. Boca Raton, FL. Mantel N (1967) The detection of disease clustering and a Imhoff JF (1992) Anoxygenic phototrophic bacteria. generalized regression approach. Cancer Res 27: 209–220. Photosynthetic Prokaryotes (Mann NH & Carr NG, eds), pp. Marquez MC, Quesada E, Bejar V & Ventosa A (1993) A 53–92. Plenum Press, New York. chemotaxonomic study of some moderately halophilic Imhoff JF (1993) Osmotic adaptation in halophilic and Gram-positive isolates. J Appl Bacteriol 75: 604–607. halotolerant microorganisms. The Biology of Halophilic Oren A (1993) Ecology of extremely halophilic microorganisms. Bacteria (Vreeland RH & Hochstein LJ, eds), pp. 211–253. The The Biology of Halophilic Bacteria (Vreeland RH & Hochstein CRC Press, Boca Raton, FL. LJ, eds), pp. 25–54. CRC Press, Boca Raton, FL. Imhoff JF (2001) True marine and halophilic anoxygenic Oz A, Sabehi G, Koblizek M, Massana R & Beja O (2005) Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 phototrophic bacteria. Arch Microbiol 176: 243–254. Roseobacter-like bacteria in Red and Mediterranean Sea Imhoff JF (2002) Phototrophic anoxygenic bacteria in marine aerobic anoxygenic photosynthetic populations. Appl Environ and hypersaline environments. Encyclopedia of Environmental Microb 71: 344–353. Microbiology (Bitton G, ed), pp. 2470–2489. John Wiley & Prado B, Delmoral A, Quesada E, Rios R, Monteolivasanchez M, Sons Ltd, New York. Campos V & Ramos-Cormenzana A (1991) Numerical Imhoff JF (2003) Phylogenetic Taxonomy of the family taxonomy of moderately halophilic gram-negative rods Chlorobiaceae on the basis of 16S rRNA and fmo isolated from the Salar de Atacama, Chile. Syst Appl Microbiol (Fenna–Matthews–Olson protein) gene sequences. Int J Syst 14: 275–281. Evol Micr 53: 941–951. Ramos-Cormenzana A (1993) Ecology of moderately halophilic Imhoff JF & Pfennig N (2001) Thioflavicoccus mobilis gen. nov., sp bacteria. Biology of Halophilic Bacteria (Vreeland RH & nov., a novel purple sulfur bacterium with bacteriochlorophyll Hochstein LJ, eds), pp. 55–86. CRC Press, Boca Raton, FL. b. Int J Syst Evol Micr 51: 105–110. Rech JA, Currie BS, Michalski G & Cowan AM (2006) Neogene Imhoff JF & Thiel V (2010) Phylogeny and taxonomy of climate change and uplift in the Atacama Desert, Chile. Chlorobiaceae. Photosynth Res 104: 123–126. Geology 34: 761–764. Imhoff JF & Truper¨ HG (1977) Ectothiorhodospira halochloris sp. Risacher F & Alonso H (1996) Geochemistry of the Salar de nov. – a new extremely halophilic phototrophic bacterium Atacama. Part 2: water evolution. Revista Geologica´ de Chile 23: containing bacteriochlorophyll b. Arch Microbiol 114: 115–121. 123–134. Imhoff JF & Truper¨ HG (1981) Ectothiorhodospira abdelmalekii Risacher F, Alonso H & Salazar C (2003) The origin of brines and sp. nov., a new halophilic and alkaliphilic phototropic salts in Chilean salars: a hydrochemical review. Earth-Sci Rev bacterium. Zbl Bakt Mik Hyg I C 2: 228–234. 63: 249–293. Imhoff JF, Hashwa F & Truper¨ HG (1978) Isolation of extremely Rodriguez-Valera F, Ventosa A, Juez G & Imhoff JF (1985) halophilic phototrophic bacteria from the alkaline Wadi Variation of environmental features and microbial Natrun, Egypt. Arch Hydrobiol 84: 381–388. populations with salt concentrations in a multi-pond saltern. Imhoff JF, Sahl HG, Soliman GSH & Truper HG (1979) Wadi Microb Ecol 11: 107–115. Natrun - chemical composition and microbial mass Schloss PD, Westcott SL, Ryabin T et al. (2009) Introducing developments in alkaline brines of eutrophic desert lakes. MOTHUR: open-source, platform-independent, community- Geomicrobiol J 1: 219–234. supported software for describing and comparing microbial Jiang HC, Deng SC, Huang QY, Dong HL & Yu BS (2010) communities. Appl Environ Microb 75: 7537–7541. Response of aerobic anoxygenic phototrophic bacterial Tank M, Thiel V & Imhoff JF (2009) Phylogenetic relationship of diversity to environment conditions in saline lakes and phototrophic purple sulfur bacteria according to pufL and Daotang River on the Tibetan Plateau, NW China. pufM genes. Int Microbiol 12: 175–185. Geomicrobiol J 27: 400–408. Thiel V, Neulinger SC, Staufenberger T, Schmaljohann R & Keane TM, Naughton TJ & McInerney JO (2004) Imhoff JF (2007) Spatial distribution of sponge-associated ModelGenerator: amino acid and nucleotide substitution model bacteria in the marine sponge Tethya aurantium. FEMS selection. National University of Ireland, Maynooth, Ireland. Microbiol Ecol 59: 47–63. Koellner T, Hersperger AM & Wohlgemuth T (2004) Rarefaction Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F & Higgins method for assessing plant species diversity on a regional scale. DG (1997) The CLUSTAL_X windows interface: flexible strategies Ecography 27: 532–544. for multiple sequence alignment aided by quality analysis Kulikova T, Aldebert P, Althorpe N et al. (2004) The EMBL tools. Nucleic Acids Res 25: 4876–4882. nucleotide sequence database. Nucleic Acids Res 32: D27–D30. Triado-Margarit´ X, Vila X & Abella CA (2010) Novel green sulfur Lukow T, Dunfield PF & Liesack W (2000) Use of the T-RFLP bacteria phylotypes detected in saline environments: technique to assess spatial and temporal changes in the ecophysiological characters versus phylogenetic taxonomy. bacterial community structure within an agricultural soil Antonie van Leeuwenhoek 97: 419–431.

Truper¨ HG & Galinski EA (1986) Concentrated brines as habitats Zu´ niga˜ LR, Campos V, Pinochet H & Prado B (1991) A for microorganisms. Cell Mol Life Sci 42: 1182–1187. limnological reconnaissance of lake Tebenquiche, Salar de Valderrama MJ, Prado B, Del Moral A, Rios R, Ramos- Atacama, Chile. Hydrobiologia 210: 19–24. Cormenzana A & Campos V (1991) Numerical taxonomy of moderately halophilic Gram-positive cocci isolated from the Supporting Information Salar de Atacama (Chile). Microbiolog´ıa (Madrid, Spain) 7: Additional Supporting Information may be found in the 35–41. online version of this article: Vila X, Guyoneaud R, Cristina XP, Figueras JB & Abella CA (2002) Green sulfur bacteria from hypersaline Chiprana Lake Table S1. Database sequences used for phylogenetic analysis. (Monegros, Spain): habitat description and phylogenetic Downloaded from https://academic.oup.com/femsec/article/74/3/510/586152 by guest on 30 September 2021 relationship of isolated strains. Photosynth Res 71: 165–172. Please note: Wiley-Blackwell is not responsible for the Yan S, Fuchs BM, Lenk S, Harder J, Wulf J, Jiao NZ & Amann R content or functionality of any supporting materials sup- (2009) Biogeography and phylogeny ofThe NOR5/OM60 plied by the authors. Any queries (other than missing clade of Gammaproteobacteria. Syst Appl Microbiol 32: material) should be directed to the corresponding author 124–139. for the article.