Salinibacter representatives have been detected in the environment using different techniques, with different levels of sensitivity that can yield contradictory results even when applied to the same sample (some examples are given below). Therefore, one must be aware of their characteristics in order to compare results obtained using different techniques. Our group has used basically three methods for the detection of S. ruber and relatives in natural samples: FISH, 16S rRNA gene clone libraries and DGGE analysis, and culture. Among these three, fluorescence in situ hybridization, FISH, is the only method for direct quantification in natural samples as it permits the identification of single cells by means of the use of phylogenetic probes. However it has some well known limitations like problems with cell permeation, relatively high thresholds of ribosome content, accessibility to the secondary structure, etc. [18]. One of the major constraints of the technique is the database comprehensiveness, i.e. when new sequences belonging to a given group are discovered, probes should be re-evaluated and redesigned so they target the whole group (see below the example of Salinibacter sequences in Tuz Lake).
Second, a microorganism can be detected in environmental samples by analysis of 16S rRNA gene sequences PCR amplified from environmental nucleic acids, either by clone library construction or denaturing gradient gel electrophoresis (DGGE). DGGE is normally used for community fingerprinting as it allows the separation of different sequences of the same size on the same denaturing gel. Since they include a PCR step in which 16S rRNA genes are amplified using "specific" primers (not always so specific) and environmental DNA as template, both techniques do not provide quantitative data but, obviously, allow for the detection of sequences related to that of S. ruber. In the case of DGGE, there is an extra limitation since the sequences retrieved are partial and therefore almost two thirds of the phylogenetic information is missing.
Finally, the culture approach provides a source of detection and quantification despite it has important limitations derived from the different degree of culturability of the natural microbial populations. In any case, it provides important and accessible information on the isolates.
In addition, Salinibacter spp. have been also monitorized by means of other techniques like sulfonolipid detection [19], pigment composition analyses [20], total melting profiles, and reassociation techniques [21], among others.
Using all these methods, S. ruber or closely related bacteria has been detected all over the world (Figure 1). In Europe, for instance, Salinibacter representatives have been found in crystallizer pond salterns in mainland Spain (Alicante and Tarragona), Balearic (Mallorca and Ibiza) and Canary Islands. In all cases, the bacterium was detected both by culture and molecular methods, including FISH, DGGE and 16S rRNA gene clone library analysis. Additionally, the analysis by electrospray mass spectrometry of lipid extracts allowed the detection of S. ruber sulfonolipid signature peak at m/z 660 in a crystallizer pond in the Margherita di Savoia salterns in Italy [19].
In Asia, Salinibacter close relatives have been found in Turkey and Israel. When analyzing the bacterial community inhabiting the hypersaline Tuz Lake in central Anatolia, Turkey [22], sequences related to Salinibacter dominated bacterial 16S rRNA gene clone libraries and DGGE profiles although FISH counts gave very low numbers. A close look at the new 16S rRNA gene sequences showed that many of them, although clustering with Salinibacter phylotypes (see below and Fig. 2), lacked the signature sequences targeted by the FISH probes. This is an example of how dependent FISH results are on the specificity of the probes. Therefore, caution must be exerted when analyzing by FISH a microbial community since, strictly speaking, probes should be always rechecked against 16S rRNA gene sequences retrieved from the samples being analyzed. In Israel, although S. ruber could not be detected by culture-independent methods in water samples from Eilat salterns, it could be readily isolated [23].
In Africa, sequences related to S. ruber have also been retrieved from water and sediments from three different soda lakes in the Wadi An Natrum depression in Egypt [24]. We also detected S. ruber by FISH in a water sample taken from a shabka in Suez (Fig. 1).
In the Andean Maras salterns, in Peru, S. ruber could be easily isolated in culture from brine samples taken in different years although it could not be detected either by FISH or by analyzing 107 bacterial 16S rRNA gene clones [15]. In fact, the bacterial community there was dominated by members of the class Proteobacteria, and specially representatives of the recently classified Salicola marasensis [15]. Finally, during the last Halophiles meeting, the presence of S. ruber-related sequences has also been reported for other locations in America, such as the Great Salt Lake in Utah (oral presentation by C. D. Litchfield), in the atalassohaline Andean Lake Tebenquiche in Northern Chile (oral presentation by C. Pedrós-Alió), and Guerrero Negro (oral presentation by S. Sabet) salterns in Baja California, Mexico.
Both clones and isolates very closely related to S. ruber have been recently found in crystallizer ponds from three different salterns in Australia (Dickson Oh and Mike Dyall-Smith, personal communication): one at Dry Creek, South Australia; another in Lara, Victoria, and a third in Bajool, Queensland (Fig. 1). Most interestingly, some of these isolates correspond to the so-far uncultured (see below) cluster EHB-2.
Apart from sequences clustering with S. ruber phylotypes EHB-1 and EHB-2 (from Extremely Halophilic Bacteria 1 and 2 [4], Figure 3), Bacteroidetes sequences more distantly related to S. ruber, were very abundant in 16S rRNA gene libraries constructed with DNA extracted from the different layers of an endoevaporite (crystallized gypsum-halite matrix in near-saturated salt water) from saltworks in Guerrero Negro, Mexico [25]. In addition, partial 16S rRNA gene sequences with similarities of less than 92% to Salinibacter have been retrieved from biofilms colonizing Mayan monuments in Uxmal (Mexico) [26]. Also, similar gene sequences were found in clone libraries obtained from a hypersaline endoevaporite microbial mat from a pond of 20% salinity in Eilat salterns in Israel [27]. These sequences were most abundant in the green layer of the mat. Finally, 16S rRNA gene sequences that could represent distinct novel lineages within the radiation of the genus Salinibacter have been recovered from evaporite crusts in brine pools at the Badwater site in Death Valley National Park, California [28]. Bacteria with similarities between 93 and 94% with S. ruber 16S rRNA gene have been isolated from these samples [28].
As illustrated by all these examples, S. ruber and relatives have been detected by different methods in a wide variety of environments. In some case, the bacterium is abundant, as directly demonstrated by FISH or other methods, or has been retrieved only by cultivation. The relevance of the finding of Salinibacter in a given environment depends on the technique used for its detection: it is possible, as with the example of Maras salterns, that the bacterium is a very minoritary component of the community or, as seen in Mallorca salterns, is a very abundant component, easily detectable. Thus, depending on the analyzed environment, Salinibacter spp. can be either one of the most abundant, and most likely ecologically relevant organisms, or be part of what has been called "the seed bank" [29]. According to Pedrós-Alió [29], most abundant taxa in a given environment would be the "core" species that would be maintained through active growth and fuel carbon and energy flows. This group would be accessible through molecular techniques such as 16S rRNA gene clone library analysis. In addition, there would be a diverse assemblage of rare taxa, or "seed bank", that "will seldom be retrieved" by this molecular approach but could occasionally be recovered by cultivation. In our examples of Mallorca and Maras samples, S. ruber would be core and seed bank, respectively. One thus may hypothesize that in a given environment S. ruber could also change its status as it becomes more or less abundant. However, the preliminary data we have obtained so far with crystallizer ponds from solar salterns indicate that although the bacterium experiences changes in numbers along the year, those are not very dramatic, at least under "normal" environmental conditions (unpublished results).
The abundance of Salinibacter spp. has been thoroughly measured only in a few hypersaline environments and therefore, we do not have a clear picture of how abundant this bacterium is at a global scale. In the places analyzed, Salinibacter spp. ranges form 2 to 30% (see Figure 1) although in some cases, like Tuz Lake, this value is clearly an underestimation since the FISH probes used did not target the whole assemblage of Salinibacter sequences. In Santa Pola salterns, Salinibacter spp. was detected only in ponds with salinities above 22.4%. In fact, the numbers detected by FISH increased with salinity (from 3.5 to 12% in three ponds of 25, 31.6 and 37% total salts) [4]. However, direct proof of the activity of Salinibacter spp. in the highest salinity (37% total salts) ponds has not been obtained so far. In fact, Gasol et al. [30] found evidences that above 32% salinity all the prokaryotic activity was carried out by haloarchaea in the same ponds where S. ruber accounted for up to 18% of the DAPI counts. This observation was based on the assumption that Salinibacter was not inhibited by taurocholate, which is a potent haloarchaeal inhibitor. In a recent work [7], Elevi Bardavid and Oren have shown that taurocholate does not inhibit aminoacid synthesis by Salinibacter. Therefore, the function of Salinbacter in most saturated crystallizers remains unkown.