OUP user menu

Distribution of prokaryotic genetic diversity in athalassohaline lakes of the Atacama Desert, Northern Chile

Cecilia Demergasso, Emilio O. Casamayor, Guillermo Chong, Pedro Galleguillos, Lorena Escudero, Carlos Pedrós-Alió
DOI: http://dx.doi.org/10.1016/j.femsec.2003.12.013 57-69 First published online: 1 April 2004


Athalassohaline lakes are inland saline aquatic environments with ionic proportions quite different from the dissolved salts in seawater. Prokaryotes inhabiting athalassohaline environments are poorly known and very few of such places have been surveyed for microbial diversity studies around the world. We analyzed the planktonic bacterial and archaeal assemblages inhabiting several of these evaporitic basins in a remote and vast area in northern Chile by PCR-denaturing gradient gel electrophoresis (DGGE) and sequencing of 16S rRNA gene fragments. Most systems were springs and athalassohaline ponds in different saltflats of the Atacama Desert region, including Salar de Llamará (in the Central Depression), Salar de Atacama (in the Pre-Andean Depression) and Salar de Ascotán (in the Altiplano). Overall, we analyzed more than 25 samples from 19 different environments with strong gradients of altitude, qualitative ionic compositions and UV influence. Between 4 and 25 well-defined DGGE bands were detected for Bacteria in each sample, whereas Archaea ranged between 1 and 5. Predominant DGGE bands (defined by intensity and frequency of appearance) were excised from the gel and sequenced. Bacterial assemblages were dominated by the Cytophaga–Flavobacterium–Bacteroides (CFB) phylum and a few Proteobacteria. There was a tendency for increasing contribution of CFB with higher salinities and altitude. Thus, CFB accounted for the major fraction of band intensity in the Ascotán samples and for lower percentages in Atacama and Llamará. When the distribution of particular CFB sequences was examined, there were several relatives of Psychroflexus torquis substituting each other as salinity changed in Ascotán. Another set of CFB sequences, very distantly related to Cytophaga marinovorus, was abundant in both Llamará and Atacama at salinities lower than 7%. Archaeal assemblages were dominated by uncultured haloarchaea distantly related to cultured strains mostly obtained from thalassohaline environments. Most of the archaeal sequences did not have a close match with environmental 16S rRNA genes deposited in the database either. Therefore, athalassohaline environments are excellent sources of new microorganisms different from their counterparts in thalassohaline sites and useful tools to relate microbial genetic diversity and environmental characteristics such as changes in salinity (both qualitative and quantitative) and altitude.

  • Athalassohaline lakes
  • Cytophaga
  • DGGE
  • Extremophiles
  • Fingerprinting
  • Haloarchaea
  • Hypersaline
  • Proteobacteria
  • Salar
  • Salt lake
  • 16S rDNA

1 Introduction

The extent of microbial diversity is much larger than the approximately 6000 species that have been formally described up to now [13]. However, the estimates of how many different microbial species exist differ by orders of magnitude. Recognizing this, the international research program Diversitas (http://www.icsu.org/diversitas/) has set “Microbial Ecology” as one of the eleven priority areas for future studies. The aims of the program are to determine how many different microbial species exist, to determine their function, and to provide the know-how to identify environments of high and low microbial diversity.

Hypersaline environments are very useful in studies on microbial diversity. On the one hand, microbial diversity decreases with increasing salinity [47]. On the other hand, hypersaline systems have proved to be an excellent source of new culturable microorganisms [5,811]. Coastal solar salterns have received considerable research attention in the past [12]. These systems consist of a series of ponds in which seawater is progressively concentrated until sodium chloride precipitates and is commercially exploited. FISH analysis of the ponds of highest salinity (around 35%) has shown that these are composed mostly of one uncultured archaeon and one bacterium (Salinibacter rubber), which has been recently isolated [13,14]. However, many other archaea can be isolated from these ponds [15], indicating that the assemblage is dominated by very few phylotypes, but many organisms are present in very low abundance and may grow if conditions change. At lower salinity ponds, on the contrary, a rather diverse assemblage of bacteria and archaea was found [16]. A recent study in a solar saltern in southern Spain revealed a variety of Thermoplasma-related archaea together with many different bacterial groups [17]. Up to now, most studies have been carried out in thalassohaline solar salterns. Thalassohaline waters are concentrated seawaters with NaCl as the major salt. Conversely, athalassohaline waters are saline waters that are rich in anions other than chloride (e.g., sulphate) and frequently in cations other than sodium such as calcium and magnesium (R.M. Bond, cited in [18]). The diversity of aquatic hypersaline habitats is enormous [19,20]. Thus, we decided to carry out a survey of bacterial and archaeal diversity in some of these poorly studied athalassohaline habitats.

The Atacama Desert area in northern Chile is an ideal place for this kind of studies. The Andes present special geological, geomorphological and climatic characteristics between 14° and 27° South. This region with aridic to semi aridic environments includes many different saline deposits, some of them unique in the world, like the nitrate–iodine ore deposits. In these saline deposits, evaporitic basins called saltflats can be found. A wide range of hypersaline environments with different characteristics is associated to salt deposits [21]. From west to east deposits from the Coastal Range, the Central Depression, the Pre-Andean Depression and the Altiplano Salt Deposits are found in succession (Fig. 1). All these basins receive their main water input from the east, which has a high percentage of salts from the leaching of volcanic rocks. The combination of both arid climate and closed basins in these saltflats provides a high rate of evaporation and the consequent increase in salt content of the brines. Hence, many athalassohaline lakes are found in the area. Water inputs come from the High Andes in the east, and the arid climate has moved with geological time scales from west to east. Thus, the saltflats on the west side are either fossilized because they do not receive any input, or they are in the process of fossilization because they receive only occasional inputs from sporadic flooding. Only one salt deposit is associated with the west Coastal Range (Salar Grande) and is filled exclusively with dry sodium chloride without brines. The Salar Grande did not receive any clastic input and is considered to be a “fossil salt deposit”. Greater variety can be found in the basins of the Central Depression. All of these receive irregular recharge from the east, mostly as underground water and as occasional flooding. Many basins are in the fossilization process. Salar de Llamará is located here. Salt deposits of the Pre-Cordilleran Depression are the oldest and largest (including Salar de Atacama and Salar de Punta Negra). They receive recharge from the north, east and south. Finally, the High Andean Salt Deposits are under development (Salar de Ascotán), showing intermediate stages between saline lakes and salt flats. In addition, some of them have thermal springs. Overall, these ecosystems provide a unique collection of habitats for studying microbial diversity in relation to salinity and altitude. In this paper we carried out a survey on the prokaryotic diversity of 19 different basins in this area at two times of the year (winter and summer) by use of 16S rRNA gene amplification and denaturing gradient gel electrophoresis (DGGE) (e.g., [22]) and sequencing of the most important 16S rDNA DGGE bands (as judged by intensity and frequency of appearance). After the identity of the electrophoretic bands was examined we were able to reveal some patterns between the distribution of prokaryotic diversity and the characteristics of the environment.


A map of the Atacama Desert area in Northern Chile showing the locations of all sampled systems and the characteristics of different salt deposits.

2 Materials and methods

2.1 Description of the aquatic systems studied

A summary of the geographical location, altitude and other parameters of the systems studied is shown in Table 1. We chose one environment from the Central Depression (Salar de Llamará), one from the Pre-Andean Depression (Salar de Atacama) and one from the main Andean Salt deposits (Salar de Ascotán) as our “major” study sites (Fig. 1). Information on the geochemistry of these systems can be found elsewhere [2327]. Each of the major sites was composed of several different subsystems with a range of salinities. Thus, Llamará consists of three different sink-holes (puquios) within the same saltflat. The three pools have an increasing salinity from about 1 up to 25%. We sampled one spot from each pond plus additional spots in the largest of these (the intermediate salinity pond). The Salar the Atacama is a huge system (about 2900 km2) with many different ponds in its interior. We sampled four of these ponds in its northern and eastern edges and several pools in a solar saltern for the precipitation of lithium salts (El Litio). Finally, Salar de Ascotán is fed from several warm springs on its eastern edge. The water trickles through and over, the main body of the saltflat forming several different ponds. We sampled two springs and two ponds. In addition, we took samples from La Rinconada coastal pond, from a slightly saline pond near Chiu-Chiu (salinity>0.3%), and also from the freshwater high altitude Lake Miscanti (Fig. 1).

View this table:

Geographical location, physicochemical and biological parameters, and results from DGGE gels for the analyzed samples

EcosystemSampling siteUTM coordinatesAltitude (m)DateCodeSalinity (%)pHTemperature (°C)Chlorophyll aa (μgl−1)Prokaryotes (cells ml−1)DGGE bands
Sitio 1764643043598580010-Aug-99LL4Aug3.298.4822.00.092.64E+0619nd.
Sitio 380010-Aug-99LL3Aug4.808.5010.0nd.2.86E+0618nd.
Sitio 5764815843730480010-Aug-99LL5Aug12.607.9025.01.609.14E+0612nd.
Sitio 780010-Aug-99LL2Aug1.707.9116.01.84nd.20nd.
West Doline744108257837023505-Aug-99At13Aug0.908.9410.61.679.40E+0511nd.
East Doline7441082578370235015-Mar-00At13Mar1.258.5018.01.66nd.14nd.
Burro Muerto742430958419523505-Aug-99At14Aug4.758.0121.59.928.88E+05152
El Litio7387542567622235013-Oct-00At23Oct36.406.2623.9nd.nd.05
Spring 10 brine761630057810039007-Aug-99As8Aug2.667.871.12.507.74E+0710nd.
Spring 10390017-Mar-00As9Mar0.507.1418.53.70nd.6nd.
Spring 117606800579200390017-Mar-00As10Mar0.206.7431.30.391.55E+054nd.
Coast PondLa Rinconada7404559345456022-Mar-00Ri16Mar6.068.2624.112.15.13E+0611nd.
FreshwaterChiu Chiu7515870541337254020-Mar-00Ch15Mar0.388.3218.71.597.34E+0521nd.
  • aValues in italics are average values obtained in different sampling visits. nd., not determined.

2.2 Sampling and measurements

Two sampling expeditions were conducted in August 1999 (winter) and March 2000 (summer) visiting several environments that ranged in altitude from sea level (Rinconada) to 4140 m above sea level (Lake Miscanti, visited in an additional expedition in October 2000). The environments showed a variety of salinities and other physicochemical conditions (Table 1). Sampling was carried out with a bucket at the end of a pole, avoiding corners and dead ends of ponds. Samples were transferred to plastic 2-l bottles and kept in an icebox until further processing. An Orion model 290 pH meter was used to measure temperature and pH. Salinity was measured with an Orion model 115 conductivity meter. Prokaryotic cell counts were determined with flow cytometry in 1.8 ml samples fixed with 200 μl of paraformaldehyde:glutaraldehyde (1% and 0.05% final concentrations, respectively) in criovials. Vials were frozen until processing in the laboratory. We followed the protocol of Gasol and del Giorgio [28]. Briefly, 100 μl aliquots were stained with Syto13 (Molecular Probes), a suspension of fluorescently labeled beads was added, and the samples were counted in a FACScalibur flow cytometer (Becton and Dickinson). Samples for chlorophyll analysis were filtered through Whatman GF/F glass fiber filters. The filters were placed in aluminum foil and kept frozen. Chlorophyll a concentration was determined with fluorescence of acetone extracts [29] using a Turner Designs Fluorometer.

2.3 Nucleic acid analyses, DGGE and sequencing

Microbial biomass was collected by filtration on 0.2 μm pore-size polycarbonate membranes (Nuclepore). Water was pumped until filter clogging decreased the flow rate to a few ml per min. Between 20 and 650 ml of water was filtered, depending on cell concentration and salinity, and the filters were stored at −70 °C. Filters were incubated with lysozyme, proteinase K and sodium dodecyl sulfate (SDS) in lysis buffer as described previously [30]. DNA was extracted with phenol-chloroform-isoamyl alcohol (25:24:1, vol/vol/vol) and precipitated with ethanol. In a recent paper [16] we have shown results on different biomass collection methods and DNA extraction protocols using salt ponds in a coastal solar saltern. Only minor differences were found among the methods that were not significant as compared with inter-pond variability. The extracted genomic DNA was used as target in the PCR to amplify 16S rRNA genes. Bacterial and archaeal fragments suitable for subsequent denaturing gradient gel electrophoresis (DGGE) were amplified with the primer combinations 358fGC-907r and 344fGC-915r, respectively, as described earlier [31]. Both sets of primers are universal for Bacteria and Archaea, respectively, and are widely used by molecular microbial ecologists (e.g., [32,33], and references therein). We used 6% polyacrylamide gel with empirically determined gradient of DNA-denaturant by mixing two stock solutions of acrylamide containing 40% and 80% denaturant (100% denaturant is defined as 7 M urea and 40% deionized formamide). The gradient was overlaid with nondenaturant. About 800 ng of PCR product was loaded for most of the samples and the gels were run at 100 V, 60 °C for 16 h (bacterial PCR products) and 250 V, 60 °C for 5 h (archaeal PCR products) using 1× TAE buffer (40 mM Tris base [pH 7.4], 20 mM sodium acetate, 1 mM EDTA) in a CBS DGGE-2000 system (CBS Scientific Company). The gels were stained with the nucleic acid stain SybrGold for 45 min, rinsed with TAE buffer, removed from the glassplate to a UV transparent gel scoop (Sigma) and visualized with UV radiation in a Fluor-S MultiImager with the Multi-Analyst software (Bio-Rad). High-resolution images (1312×1034 pixels, 12-bits dynamic range) were analyzed using the gel plotting macro tool of the NIH-Image software package version 1.62 (National Institute of Health, USA). After background subtracting, the intensity of each band was measured by integrating the area under the peak and was expressed as percent of the total area in the lane. The error measured among replicates was less than 4%.

Bands were excised from the gels and re-amplified as reported elsewhere [22]. PCR products were purified with the QIAquick PCR-Purification Kit (QIAGEN) and quantified in an agarose gel. About 10–20 ng were directly used for the sequencing reaction using primer 358f without the GC-clamp, with the Bigdye Terminator Cycle Sequencing Kit v2.0 (PE Biosystems) and an ABI PRISM model 377 (v3.3) automated sequencer. Sequences were sent to BLAST search (http://www.ncbi.nlm.nih.gov) to determine the closest relative in the database. A similarity matrix was built using partial sequences (i.e. 500 bp) with the ARB software package (Technical University of Munich, Munich, Germany; http://www.arb-home.de). Partial sequences were inserted into the optimized and validated tree available in ARB (derived from complete sequence data), by using the maximum-parsimony criterion and a special ARB parsimony tool that did not affect the initial tree topology. Nucleotide sequence accession numbers at EMBL are: AJ487523 to AJ487534, AJ566770 to AJ566784, and AJ568003 to AJ568014.

3 Results

Geographical location, physicochemical and biological parameters for all ecosystems sampled in the present study are shown in Table 1. These data will be used in different combinations as needed. We have grouped the samples into three main systems (Llamará, Atacama and Ascotán) and additional data are provided for the other sites (Rinconada, Chiu-Chiu, and Miscanti). Each one of the three main systems comprised a series of water masses with salinities ranging between 1.6% and 24.2% (Llamará), between 0.9% and 36% (Atacama) and between 0.2% and 14.1% (Ascotán). Thus, a wide range of salinity values and compositions (Table 2) could be sampled at each altitude. Chlorophyll a values ranged between 0.06 μgl−1 in Cejas and 14.72 μgl−1 in Cebollar, and prokaryotic counts between 1.6×105 cells ml−1 in a spring and 7.7×107 cells ml−1 in a brine, both in Ascotán (Table 1). Ionic composition was quite similar between Llamará and Atacama; Ascotán, however, was the poorest system for most of the salts but contained the richest spot in arsenic (Table 2).

View this table:

Main ionic composition for several aquatic environments in the Atacama Desert region

EcosystemSiteNa+ (mM)K+ (mM)Ca2+ (mM)Mg2+ (mM)HCO3 (mM)CO32− (mM)Cl (mM)SO42− (mM)SiO2 (mM)B (mM)As (mM)
LlamaráSitio 115163025793.211972621.35.60.06
Sitio 3325357181133.626455071.03.20.14
Sitio 54319.220223.2420891.7
Sitio 74729.525253.6380901.70.02
Burro Muerto9099712710.92.81132330.8100.05
Spring 10411.
Spring 11906.
  • Data also obtained from [25,27,47]. Empty spaces: not determined.

Bacterial and archaeal DGGE gels were run separately (Fig. 2). Archaea were surveyed only in Salar de Atacama but covered a wide range of salinities. The number of DGGE bands per lane for Bacteria ranged between 4 and 25, whereas for Archaea the range was smaller (1–5 bands). The highest numbers of bacterial bands were in Miscanti and Chiu-Chiu, and the lowest in a spring in Ascotán. For Archaea the highest number of bands appeared in a pond of Lake Tebenquiche, and the lowest in Cejas. We also detected up to five archaeal bands in one lithium pond where bacteria could not be PCR-amplified. The main DGGE bands (in terms of intensity and frequency of appearance) were excised, reamplified and sequenced, although, we could not always obtain good sequences. From 67 excised bands, 13 failed reamplification or sequencing. Overall, we found a striking dominance of the Cytophaga–Flavobacterium–Bacteroides (CFB) phyla and of the Haloarchaea (i.e., Halobacteriaceae) (Table 3). These two groups of sequences showed 83–98% similarity to sequences in the database, and the best match with cultured strains was found at 98% with Psychroflexus torquis (CFB isolated from Antartic sea ice, [34]) and at 94% with Haloarcula marismortui (haloarchaea isolated from the Dead Sea). Such levels of similarity are usually assumed to represent different genera when complete 16S rRNA gene sequences (i.e., 1500 bp) are compared. Sequences whose closest relative was a member of the genus Psychroflexus were the most frequently recovered (10 bands, Fig. 3). Haloarchaea were also the most abundant prokaryotic bands in the hypersaline lithium ponds (two sequences distantly related to Halorubrum spp., Fig. 4). Proteobacteria were also detected (one was 97% similar to Alcaligenes sp., two were 94% similar to the Ectothiorhodospiraceae Alkalispirillum mobile, and one more was 93% similar to the colorless sulfur bacteria Thiomicrospira) and only a few bands belonged to other groups, like the high-GC gram-positive (the actinobacterium Leifsonia was the closest relative for three bands), chloroplast of diatoms (one band) and unidentified euryarchaea (six bands were within a cluster containing only environmental sequences).


Negative image of denaturing gradient gel electrophoresis gels containing bacterial (upper gels) and archaeal (lower gel) 16S rDNA fragments. Bands that were cut off from the gel are labeled with the same number as in Table 3, and Figs. 3 and 4. When bands across several lanes could be identified as being the same, they have the same number. Some DGGE bands failed reamplification or sequencing.

View this table:

Closest relatives in the database (BLAST search) for 54 selected bacterial and archaeal DGGE bands shown in Fig. 2

Phylogenetic groupClosest relative% similarityDGGE bands
CFB phylum (total 19 bands)Psychroflexus torquis85–981,2,4,5,14,17,31,33,34,45
Cytophaga marinoflava89–9011,36,41
Unidentified clones83–983,19,23,28,49,50
Proteobacteria phylum (total 15 bands)α-proteobacterium90–9410,24,51
Unidentified clones83–988,12,15,16,20,35,38,42
High-GC gram-positive (total 4 bands)Leifsonia sp.91–9713,25,44
Unidentified clone9640
Algae (total 1 band)Chloroplast diatom9118
Haloarchaea (total 9 bands)Haloarcula marismortui93–9411,15
Unidentified clones84–981,2,9,12,17,18,C
Euryarchaea (total 6 bands)Unidentified clones84–963,4,5,6,13,14
  • CFB phylum=Cytophaga–Flavobacteria–Bacteroides phylum. Some DGGE bands failed reamplification or sequencing. Haloarchaeal DGGE band number 9 was obtained from an additional sampling point in October 2000.


Phylogenetic affiliation of the 16S rDNA partial sequences (c.a., 500 bp) from the excised bands (see Fig. 2) belonging to haline environments within the Cytophaga–Flavobacterium group. Scale bar=0.10 mutations per nucleotide position. The code includes “NCh” for Northern Chile, the band number as in Fig. 2 (upper gels), and the code for the natural environment from which the 16S rRNA gene sequence was originally obtained.


Phylogenetic affiliation of the 16S rDNA partial sequences (c.a., 500 bp) from the excised bands (as in Fig. 2, lower gel) of Salar de Atacama (ATA) within the Archaea. Scale bar=0.10 mutations per nucleotide position. Sequence ATA-A09 was obtained from an additional sampling point in October 2000.

Fig. 5 shows the general composition of bacterial and archaeal assemblages from the three main systems studied, using the relative DGGE band intensities as an estimate of relative abundance. An average of 29% from the total bacterial band intensity in each lane could not be sequenced. In a few cases, an important fraction of the total band intensity in each lane could not be identified (>40% unidentified in five samples). The conclusions must carefully take this unknown into account. For Archaea, more than 90% of the total band intensity could be identified in all cases. We found considerable heterogeneity in the different sampling stations within each system. Thus, CFB sequences were abundant in most Llamará samples, but they were not detected in Sitio 5. Likewise, Actinobacteria (HGC in Fig. 5) were relatively important at Sitio 7, but not detected or minor components at the remaining sampling stations. Similar differences in composition were found between different sampling spots in the other two systems (Fig. 5). On the other hand, differences within the same spot between winter and summer samples seemed to be smaller than those due to space. For example, Cebollar and Spring 10 brine both sampled in Ascotán had a virtually identical composition in summer and winter. In other cases, however, differences were more apparent for certain groups (e.g., haloarchaea and unidentified euryarchaea shifted in Lake Tebenquiche). When all samples were compared, two trends could be observed. First, a reduction in the number of bacterial DGGE bands occurred from the low altitude Llamará (average 14 bands) to the high altitude Ascotán (average 8 bands). Second, the relative importance of CFB group was greater in Ascotán (average 65% of total band intensity) than in Atacama or Llamará (32% and 26%, respectively).


Distribution of the sequences retrieved from denaturing gels of Salar de Llamará, Salar de Atacama, and Salar de Ascotán. For each ecosystem the proportions of the total band intensity ascribed to a particular group of microorganisms are indicated (CFB: Cytophaga–Flavobacteria, PRO: Proteobacteria, HGC: high-GC gram-positive bacteria, ALG: chloroplast from Algae, HAL: Haloarchaea, EUR: uncultured Euryarchaea, NIB: non-identified bacteria, NIA: non-identified archaea). Relative proportions of different bacteria and archaea are quantified separately. UTM coordinates are shown for the center of each system. Note differences in scale among the three systems.

When we correlated the most abundant phyla (CFB and Proteobacteria) to the physicochemical properties of each sample, most relationships were found insignificant. However, the relative contribution of CFB to the total bacterial assemblage increased with salinity (slope 5.6, R2=0.584) whereas Proteobacteria did not. The increase of CFB with salinity was independent of the trend from Llamará to Ascotán, since both low and high salinity samples were present in the three sites. However, the distribution of different CFB sequences along the range of salinity indicated that different CFB populations were adapted to different ranges of salinity. Thus, in Ascotán Psychroflexus-like sequences were the most abundant, however, we noted a substitution of populations with salinity: the population represented by DGGE bands 1–5 was present at salinities below 5% whereas the populations represented by bands 31–34 and 33 appeared at higher salinities. We also detected that sequences 11, 36 and 41 (distantly related to Cytophaga marinoflava) were present at different salinities and sites. Such bands were present in Atacama and Llamará but not in Ascotán, at a range of salinity between 1% and 7%. When salinity increased the CFB assemblage changed its composition (band 2 in Llamará and bands 17 and 19 in the Atacama salt flat).

4 Discussion

The combination of different environmental conditions and salt composition in the water masses of the Atacama Desert results in the existence of a wide range of environments that allow for exploring 16S rRNA gene sequences within a comparative environmental framework. We selected a fingerprinting technique (PCR-DGGE) and sequence analysis of the resulting 16S rRNA gene bands as the appropriate way to carry out a survey on the genetic diversity in a wide set of remote and unknown aquatic environments from this region. We are aware of potential PCR biases [35], especially those that affect the final ratios of PCR amplicons, thus compromising quantitative interpretations. This is why we do not attempted to give absolute values for the presence of different phylogenetic groups. Working with mixtures of pure cultures and with natural samples, Muyzer et al. [36] and Casamayor et al. [22,37], respectively, found, a highly significant positive correlation between the bands intensities and the predominant community members. Several other authors also reached the same conclusion [3841]. Overall, it appears that relative band intensities in DGGE fingerprints are informative for comparative purposes and can be used to follow relative changes of particular populations, but may be poor indicators of absolute abundances. DGGE is a powerful tool for survey studies, which require sampling of many different points, but the use of general domain-level primers fail to notice less abundant populations (e.g., [22]). Other techniques (such as cloning, sequencing and FISH) can now be used to obtain more detailed and quantitative information. A new sampling scheme in the Atacama region is in progress to complete such studies.

The number of bands in the DGGE fingerprints offered useful information. Bacterial richness in the spots where both domains were tested (Salar de Atacama) was higher than archaeal richness. Archaeal sequences were widely distributed within both the haloarchaea and the cluster of environmental unidentified euryarchaea. We found the same trend in a coastal solar saltern [16,17]. We also noted a trend of lower numbers of bacterial bands in higher altitude systems. Thus, the average number of bands was very similar in Llamará and in Atacama (14.4 and 12.6, respectively) and only 8.4 in Ascotán. The latter value was significantly different from the other two. Freshwater systems, however, did not show this trend. Spring waters from el Tatio at 4300 m, for example, had between 11 and 21 bands (data not shown) while Lake Miscanti at 4140 m had 25. These numbers of bands are not significantly smaller than those from Chiu-Chiu at 2540 m (21 bands) or from the lower salinity ponds in Llamará (between 8 and 20 bands). Thus, the differences among the three main ecosystems cannot be attributed to more extreme conditions related to altitude (such as UV radiation or frequent freezing). Because the lower number of bands in Ascotán was common to both hot water springs and hypersaline lagoons, a possible explanation for the differences among systems could be the mineral concentration and composition of the incoming waters. In Ascotán we found the highest concentration of arsenic. Different ionic ratios are also important for biology because the osmotic strength of a solution at a given level of total dissolved solids changes with the valence of the dominant ions.

The sequences recovered from the predominant prokaryotic community members indicated also interesting trends. Only three sequences from a total of 54 (i.e., 5.5%) had 97% similarity or higher with cultivated strains (Alcaligenes sp., and Psychroflexus torquis, respectively). At 95% similarity we only found five sequences (9.2% of the total). The remaining sequences (including all archaeal sequences) were only distantly related to well-characterized microbial groups. A similar approach in coastal salt ponds yielded a much higher number of sequences related to cultivated microorganims [17]. Thus, in these athalassohaline environments the diversity of the 16S rRNA genes would mostly correspond to new organisms above the genus level and culturing efforts such as those done before in coastal salterns will probably bring these microorganisms into culture. Surprisingly, we were recovering only four phylogenetic bacterial branches: Proteobacteria (alpha, gamma, and to a lesser extent beta subgroups), CFB, Chloroplast and high-GC gram-positive bacteria. Given the wide range of aquatic environments sampled here, we expected a large coverage in the phylogenetic tree. However, the range was similar to that in solar salterns [17] where bacterial members of four groups were retrieved by cloning (alpha- and gamma-Proteobacteria, CFB, Cyanobacteria and high-GC gram-positive bacteria). Our study sites were shallow and well-mixed ponds. Our experience in similar environments has shown that significant differences do not exist for planktonic bacteria within the same pond [16,42] therefore excluding the possibility of poor sampling. Archaeal sequences from Atacama were rather distantly related to those from thalassohaline environments. Thus, the predominant haloarchaeal sequence we found in the hypersaline pond from El Litio (ATA-A17) was very distantly related to the clone SphT found in hypersaline coastal ponds and that corresponded to highly abundant square archaea [13]. Therefore, the same phylogenetic groups could be the predominant inhabitants of widespread planktonic salty environments, but with probably different species or even genera.

The best-adapted group in the wide range of environmental conditions in Atacama Desert seemed to be CFB. They are one of the most widely distributed and abundant groups of bacteria in aquatic systems (see [43] for a review). Together with the known ability of their members to degrade high molecular weight compounds, our finding has generated considerable interest for studying their diversity and ecology. It has been suggested that the initial PCR-dependent cloning surveys underestimated the abundance of CFB in natural systems [44]. Thus, very few clones of CFB appeared in libraries of 16S rRNA genes from marine environments [45]. In contrast, when challenged with probe CF319a in FISH counts, CFB were one of the most abundant groups [46]. In our DGGE studies, we always found CFB forming significant proportions of the bacterial assemblages: 14% on average (range 7–21%) of the total band intensity in a coastal marine environment [41], and up to 40% in a tropical reservoir [31]. CFB were not detected in freshwater Lake Cisó, but they were the main group of heterotrophic bacteria in neighboring Lake Vilar [22,37]. FISH counts with probe CF319a revealed that CFB accounted for 10–32% of the total DAPI count in Lake Vilar and the three DGGE bands that affiliated with the CFB group accounted for about 40 to 50% of the total band intensity [37]. It is clear, therefore that, unlike the primers that were initially used in cloning, those used in DGGE do not select against CFB.

Recently, CFB have been recognized as the main bacterial inhabitants of some hypersaline environments. The bacterium Salinibacter ruber was essentially the only bacterial inhabitant of crystallyzer ponds (37% salinity) in solar salterns [14]. We carried out a detailed study of the microbial diversity along the whole salinity gradient (seawater to crystallizers) in coastal solar salterns with a variety of molecular methods 15,16]. DGGE showed that CFB were present throughout the gradient and their relative importance increased with salinity. The trend for a larger representation of CFB with increased salinity can be seen also in the Ascotán samples. Here, the water from springs on the eastern side flows over and under the salt crust and forms ponds towards the western side that have become hypersaline. As seen in Fig. 5, Proteobacteria were the dominant bands in the DGGE patterns from the spring 11 but the assemblage composition shifted dramatically, and was dominated by CFB as soon as springs became salty (Spring 10 brine) and in the salt lagoons (Cebollar). The pattern of large proportion of CFB in the more saline samples was also clear in Llamará and Atacama. However, none of the retrieved sequences was close to S. ruber. In fact, S. ruber became dominant in the coastal salterns at salinities above 20%, and in Atacama Desert we could sample only a few systems with salinities above 15% so that the presence of this bacterium in Atacama cannot be ruled out. Alternatively, S. ruber might be biogeographically limited to coastal solar salterns or it may not be able to grow in athalassohaline environments. A new sampling expedition in Northern Chile is in progress and general and specific FISH probes will be used to detect and quantify different CFB in these systems. Anyhow, hypersaline waters (both thalasso- and athalassohaline) apparently constitute an important environment for planktonic CFB and a great natural laboratory for studying the ecology of this large group of bacteria.


Sampling and measurements carried out in Chile were funded by projects FONDEF D97F1078 and D99I1026. Measurements carried out in Barcelona were funded by project MicroDIFF (REN2001-2120/MAR) from the Spanish Ministerio de Ciencia y Tecnología. Exchange agreements between Chilean CONICYT (#1999 2 02 136) and the Generalitat de Catalunya (ACI99-27) and CSIC (2001CL0020) provided funds for travels between Spain and Chile. We are grateful to Sociedad Chilena de Litio for permission to work at El Litio. We thank Vanessa Balagué for molecular analyses, Isabel Ferrera for archaeal ARB analysis, and Juan José Pueyo for comments and unpublished observations. EOC was supported by the Programa Ramón y Cajal from the Spanish Ministerio de Ciencia y Tecnología.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
View Abstract