OUP user menu

Diversity of sulfate-reducing bacteria from an extreme hypersaline sediment, Great Salt Lake (Utah)

Kasper Urup Kjeldsen, Alexander Loy, Trine Fredlund Jakobsen, Trine Rolighed Thomsen, Michael Wagner, Kjeld Ingvorsen
DOI: http://dx.doi.org/10.1111/j.1574-6941.2007.00288.x 287-298 First published online: 1 May 2007


The diversity of sulfate-reducing bacteria (SRB) inhabiting the extreme hypersaline sediment (270 g L−1 NaCl) of the northern arm of Great Salt Lake was studied by integrating cultivation and genotypic identification approaches involving PCR-based retrieval of 16S rRNA and dsrAB genes, the latter encoding major subunits of dissimilatory (bi) sulfite reductase. The majority (85%) of dsrAB sequences retrieved directly from the sediment formed a lineage of high (micro) diversity affiliated with the genus Desulfohalobium, while others represented novel lineages within the families Desulfohalobiaceae and Desulfobacteraceae or among Gram-positive SRB. Using the same sediment, SRB enrichment cultures were established in parallel at 100 and at 190 g L−1 NaCl using different electron donors. After 5–6 transfers, dsrAB and 16S rRNA gene-based profiling of these enrichment cultures recovered a SRB community composition congruent with the cultivation-independent profiling of the sediment. Pure culture representatives of the predominant Desulfohalobium-related lineage and of one of the Desulfobacteraceae-affilated lineages were successfully obtained. The growth performance of these isolates and of the enrichment cultures suggests that the sediment SRB community of the northern arm of Great Salt Lake consists of moderate halophiles, which are salt-stressed at the in situ salinity of 27%.

  • Great Salt Lake
  • hypersaline
  • halophilic
  • sulfate-reducing bacteria
  • diversity
  • dissimilatory (bi) sulfite reductase (dsrAB)


Hypersaline aquatic environments are abundant worldwide and include inland lakes, marine coastal areas such as salt marshes and solar salterns, as well as deep-sea and oil-reservoir brines (Oren, 200f2). High primary productivity and high sulfate concentrations are characteristic for many hypersaline aquatic systems (Oren, 2002), favouring the activity of sulfate-reducing bacteria (SRB) in anoxic zones. Indeed, high sulfate reduction rates have been measured in hypersaline brines, microbial mats and sediments at salinities exceeding 20%, demonstrating a quantitative important role of SRB in mineralization processes even at such extreme salinities (Caumette et al., 1994; Brandt et al., 2001; Daffonchio et al., 2006). In contrast, most characterized strains of SRB seem poorly adapted to extreme hypersaline conditions, the highest reported in vitro growth optimum for known species being 10% salinity with upper limits for growth ≤24% (Ollivier et al., 1991; Krekeler et al., 1997; Sass & Cypionka, 2004; Warthmann et al., 2005; Belyakova et al., 2006). Few studies have addressed the diversity of SRB in extreme hypersaline environments (Mouné et al., 2003; Sørensen et al., 2005; Daffonchio et al., 2006); however, all indicated the presence of novel phylotypes of SRB with unknown physiological properties.

The northern arm of Great Salt Lake (Utah) is an extreme thalassohaline hypersaline environment with salinities of around 27% (US Geological Survey [http://www.usgs.gov]). Despite the high salinity, the sediment from this part of the lake was recently reported to sustain significant sulfate-reducing activity (30 nmol cm−3 day−1) and to harbour high numbers (107 cm−3) of viable SRB (Brandt et al., 2001). However, little is known about the identity of the SRB residing in the sediments of Great Salt Lake. The aim of the present study was thus to couple molecular identification and phenotypic data for the SRB community inhabiting the 27% salinity sediment of the northern arm of Great Salt Lake using an integrated cultivation-dependent and cultivation-independent approach. A range of enrichment cultures was set up in parallel at 10 and 19% salinity using different electron donors in order to identify and isolate SRB able to grow at high salinities in vitro. SRB proliferating in stable-growing enrichment cultures as well as those inhabiting the 27% salinity sediment were identified by PCR-based retrieval of dsrAB genes encoding the alpha and the beta subunit of the dissimilatory (bi)sulfite reductase, which serves as a phylogenetic marker for SRB (Karkhoff-Schweizer et al., 1995; Klein et al., 2001; Zverlov et al., 2005). Furthermore, 16S rRNA gene sequences were retrieved from the enrichment cultures using SRB group-selective primers to support and complement the dsrAB-based identification approach.

Materials and methods

Sampling site and procedure

Anoxic sediment samples were collected at 3–5 m water depth from the hypersaline northern arm of Great Salt Lake at station 27 (Brandt et al., 2001) and pooled. The porewater of the sediment was characterized by a salinity of 27% (270 g L−1 NaCl), a high sulfate concentration of 200 mM, and a pH of 7.5 (Brandt et al., 2001). The sediment samples were stored at 4°C in butyl rubber stopper-sealed glass bottles, which were filled to capacity until processed for the various experiments.

Enrichment and isolation of SRB

Anoxic enrichment cultures at 10% and at 19% salinity were initiated from the 27% salinity Great Salt Lake sediment. One litre of 10% salinity (19% salinity) medium consisted of 100 g (190 g) NaCl, 10 g (19 g) MgSO4·7H2O, 6 g (11 g) KCl, 0.4 g CaCl2·2H2O, 1.0 g NH4Cl, 0.5 g KH2PO4, 0.2 g yeast extract (Scharlau Chemie S.A., Barcelona, Spain), 50 μL 2% (w/v) resazurin solution, and 1.0 mL each of nonchelated trace element mixture and selenite–tungstate solution (Widdel & Bak, 1992). The following solutions were added to the sterile medium from anoxic sterile stocks: 1/1000 (v/v) vitamin mixture, vitamin B12 and thiamin solution (Widdel & Bak, 1992), 1/100 (v/v) 4 g L−1 MnCl2·4H2O solution, 1/100 (v/v) 12.5 g L−1 FeCl2·4H2O in 10 mM HCl solution, and 3/100 (v/v) 84 g L−1 NaHCO3 solution. Substrates (10 mM final concentration) were added from anoxic sterile stocks according to Table 1. The final pH of the medium was 7.3; further details on its preparation are given in Jakobsen et al. (2006). The media were reduced with 200 μM (final concentration) sterile sodium dithionite solution before inoculation with 5% (v/v) of suspended sediment. All incubations were carried out in the dark at 30°C. Enrichment cultures were reinoculated five to six times into fresh media at 1- to 2-month intervals until they were finally harvested by centrifugation (15 000 g; 20 min; 4°C) for DNA extraction. Isolation of pure cultures was performed as described elsewhere (Jakobsen et al., 2006). The growth of cultures was monitored by phase-contrast microscopy and by measuring sulfide concentrations (Cline, 1969).

View this table:

SRB selective enrichment cultures

CultureSalinity of medium (g L−1 NaCl)SubstrateSulfide (mM*)Number of SRB-affiliated 16S rRNA gene OTUsNumber of dsrAB OTUs
H190Methanol and BES0.362
  • * Total sulfide produced, measured upon harvest of cell material for DNA extraction.

  • OTU, operational taxonomic unit.

  • 2-bromoethanesulfonic acid (BES) was added to inhibit methanogenesis.

DNA extraction

Genomic DNA was extracted from sediment samples and from cell pellets of harvested cultures using bead beating (Fastprep DNA extractor, BIO 101, Vista, CA) and the FastDNA spin kit for soil (Q-Biogene, Carlsbad, CA) according to the manufacturers' instructions. Prior to DNA extraction, sediment samples (0.25–0.5 cm3) were washed with a 100 g L−1 sterile NaCl solution, resuspended in 675 μL DNA extraction buffer, and subjected to enzymatic digestions followed by incubation with sodium dodecyl sulfate (SDS) for 2 h at 65°C, the three latter steps according to the protocol of Juretschko et al. (1998).

Construction of clone libraries

PCR amplification of dsrAB fragments (c. 1.9 kb) was performed from three pooled independently made DNA extracts from the sediment. PCR mixtures included 25 μL of Hot Star Taq Master Mix (Qiagen, Hilden, Germany), 0.5 μL of primer variant mixtures DSR1Fmix and DSR4Rmix (each primer at a concentration of 50 pmol μL−1, MWG Biotech AG, Ebersberg, Germany; Table 2), 1 μL of bovine serum albumin (10 mg mL−1, Amersham Biosciences, Uppsala, Sweden), 22 μL of dH2O (Sigma-Aldrich, Munich, Germany) and 1 μL of template DNA. Thermal cycling consisted of 95°C for 15 min, then 30 cycles of 94°C for 40 s, 52°C for 40 s, and 72°C for 1 min 30 s. Cycling was completed after a final elongation step at 72°C for 10 min.

View this table:

16S rRNA gene- and dsrAB-targeted primers

PrimerSequence (5′–3′)Annealing temp. (°C)SpecificityReference
16S rRNA gene-targeted primers
26FAGAGTTTGATCCTGGCTCA57Most BacteriaHicks et al. (1992)
1492RGG(CT)TACCTTGTTACGACTT57, 60, 63*Most Bacteria and ArchaeaLoy et al. (2002)
ARGLO36FCTATCCGGCTGGGACTA60Archaeoglobus spp.Loy et al. (2002)
DEM116FGTAACGCGTGGATAACCT63Most Desulfotomaculum spp.Stubner & Meuser (2000)
DEM1164RCCTTCCTCCGTTTTGTCA63Stubner & Meuser (2000)
DFSPOS219FCGATTATGGATGGACCCG60Desulfosporosinus spp.This study
DSBAC355FCAGTGAGGAATTTTGCGC63Most Desulfobacterales and SyntrophobacteralesScheid & Stubner (2001)
DSBB280FCGATGGTTAGCGGGTCTG60DesulfobulbaceaeThis study
DSV682FGGTGTAGGAGTGAAATCCG60Desulfovibrionales and other DeltaproteobacteriaModified from Devereux et al. (1992)
dsrAB-targeted primers
DSR1Fmix62–52Most SRB
DSR1FAC(GC)CACTGGAAGCACGWagner et al. (1998)
DSR4Rmix§62–52Most SRB
  • * 57°C when combined with 26F; 60°C when combined with ARGLO36F and DFSPOS219F; 63°C when combined with DSBAC355F.

  • Equimolar mixture of DSR1F, DSR1Fa, and DSR1Fb.

  • During touchdown PCR the temperature was decreased from 62 to 52°C.

  • § §Equimolar mixture of DSR4R, DSR4Ra, DSR4Rb, and DSR4Rc.

PCR was performed on DNA extracts of enrichment and pure cultures using 16S rRNA gene-targeted bacterial and SRB group-selective primers as well as the dsrAB-targeted primer variant mixtures DSR1Fmix and DSR4Rmix (Table 2). PCR mixtures of 50 μL included 38.5 μL of dH2O (Sigma-Aldrich), 5 μL of 10 × Ex Taq™ buffer (Takara Bio Inc., Otsu, Shiga, Japan), 4 μL dNTP (2.5 mM each, Takara Bio Inc.), 0.5 μL of Ex Taq™ polymerase (5 units μL−1, Takara Bio Inc.), 0.5 μL of forward and reverse primer or primer mixture (each primer at a concentration of 50 pmol μL−1, MWG Biotech AG) and 1 μL of template DNA. Thermal cycling for 16S rRNA gene amplification was carried as described above, except that the initial step at 95°C was substituted by a step of 1 min at 94°C, and the annealing temperature was varied between 57 and 63°C (depending on the primer pair; see Table 2). For PCR amplification of dsrAB fragments, a touch-down PCR protocol was used and all PCR reactions were initiated by a ‘hot start’, i.e. by adding DNA polymerase to the individual reactions after the initial 1-min denaturation step. Touch-down PCR thermal cycling conditions included 20 initial cycles where the annealing temperature was decreased by 0.5°C per cycle from 62 to 52°C, followed by 15 cycles with an annealing temperature of 52°C; otherwise, thermal cycling was performed as described above.

For the generation of clone libraries (see Loy et al., 2004; Wagner et al., 2005 for details), PCR products were separated by agarose gel electrophoresis, and amplicons of the expected size were excised and cloned into chemically competent Escherichia coli cells using the TOPO XL cloning kit (Invitrogen Corp, Carlsbad, CA) as specified by the manufacturer. Clones originating from enrichment cultures were screened for unique inserts by restriction fragment length polymorphism (RFLP) analysis by digesting M13 primer pair-generated (primers supplied with the cloning kit) amplicons with the frequently cutting enzyme MspI (occasionally the enzyme ALU1 was additionally applied for confirmation), as recommended by the manufacturer (New England Biolabs GmbH, Frankfurt am Main, Germany). Restriction enzyme digests were separated on 2% (w/v) agarose gels and visualized by ethidium bromide staining. 16S rRNA gene and dsrAB nucleotide sequences were determined as described earlier by Mogensen et al. (2005) and Purkhold et al. (2000), respectively.

Comparative 16S rRNA gene and dsrAB sequence analyses

Partial 16S rRNA gene and dsrAB sequences were compiled and aligned as described elsewhere (Kjeldsen et al., 2006). Aligned sequences were checked for possible chimeric origin by phylogenetic analysis (see below) of independent 5′ and 3′ end sections of the respective sequences (Hugenholtz et al., 1998) generated by splitting the sequences into halves. Two distinct putative chimeric 16S rRNA gene sequences were removed, while no putative chimeric dsrAB sequences were found. The aligned sequences were grouped into operational taxonomic units (OTUs) based on a 97% and a 90% (see below) identity criterion for 16S rRNA gene and dsrAB sequences, respectively, using the computer program dotur (Schloss & Handelsman, 2005) applying the default sequence assignment algorithm. dotur was also used for rarefaction analysis. Phylogenetic trees based on 16S rRNA gene and deduced DsrAB amino acid sequence datasets were calculated as described previously (Jakobsen et al., 2006). The DsrAB dataset included 460 unambiguously aligned amino acid sequence positions. The 16S rRNA gene sequence dataset was based on sequences >1300 nt, and different filters were applied to select sequence positions for the phylogenetic analyses (see legend to Fig. 4 for specifications). Short sequences were added to the generated 16S rRNA gene- and the DsrAB amino acid-based trees applying maximum-parsimony criteria. Phylogenetic consensus trees were drawn as recommended by Ludwig et al. (1998). A broad range of taxa was included in the phylogenetic analyses; however, some of the reference sequences were removed from the presented phylogenetic trees to enhance clarity.


Phylogenetic consensus trees based on analyses of 16S rRNA gene sequences (≥1300 nucleotides) including sequences of representative clones and isolates (marked in bold) from enrichment cultures A–L (see Table 1). See legend to Fig. 3 for additional details. (a) Phylogenetic relationship of sequences affiliated with the class Deltaproteobacteria. The tree was inferred from neighbour-joining, maximum-parsimony, and maximum-likelihood analysis, using two different sequence conservation filters (the Deltaproteobacteria and the Bacteria filter of the ssu_jan04_corr_opt ARB database [available at http://www.arb-home.de]). Bootstrap values (neighbour-joining) were calculated using the Deltaproteobacteria filter. The bar indicates 10% sequence divergence as inferred from neighbour-joining analysis using the Deltaproteobacteria filter. (b) Phylogenetic relationship of sequences affiliated with the genus Desulfotomaculum. The tree was constructed as described above using the Firmicutes and the Bacteria filter of the ssu_jan04_corr_opt ARB database. Bootstrap values (neighbour-joining) were calculated using the former filter. The bar indicates 10% sequence divergence as inferred from neighbour-joining analysis using the Firmicutes filter.

The distributions of retrieved 16S rRNA gene OTUs and dsrAB OTUs among enrichment cultures A, B, C, D, E, G, I, J, K and L (Table 1) were evaluated by UPGMA-based cluster analysis. Only 16S rRNA gene-OTUs 1–12 (i.e. those affiliated to known SRB families, Fig. 4) were included in the analysis. Binary data matrixes were constructed based on the presence or absence of 16S rRNA gene OTUs and/or dsrAB OTUs in the ten enrichment cultures and subjected to UPGMA analysis in paup* (Swofford, 2003).

Nucleotide sequence data

The sequences determined in the present study were deposited in GenBank under the accession numbers DQ067421, DQ067422, DQ386171–DQ386261 and EF158463–EF158466.


Retrieval of dsrAB sequences from the Great Salt Lake sediment and their grouping into OTUs

A dsrAB clone library consisting of 108 clones was constructed from the sediment. Considerable (micro)diversity was evident among the sequenced clones, and the number of OTUs they could be grouped into was highly dependent on the sequence identity threshold used for the grouping (Fig. 1). Strains of a given prokaryotic species generally share >97% 16S rRNA gene-sequence identity (Stackebrandt & Goebel, 1994; Rosselló-Mora & Amann, 2001), and consequently grouping of environmentally retrieved 16S rRNA gene sequences into OTUs is often based on a 97% sequence identity threshold (e.g. Donachie et al., 2004; Juretschko et al., 2002; van der Wielen et al., 2005). Thus, in order to derive a sequence identity threshold for grouping dsrAB sequences into OTUs conceptually consistent with the 97% threshold for grouping 16S rRNA gene sequences, pairwise 16S rRNA gene and dsrAB sequence identities were plotted against each other for corresponding pairs of dissimilatory sulfate- and/or sulfite-reducing prokaryotes (collectively referred to as SRB below) (Fig. 2). A total of 119 distinct SRB for which concurrent 16S rRNA gene and dsrAB sequence data are available in GenBank were included in the analysis. As seen from Fig. 2, pairs of SRB sharing <97% 16S rRNA gene-sequence identity in general share <90%dsrAB sequence identity, and hence 90% sequence identity was chosen as an arbitrary threshold for grouping the retrieved environmental dsrAB sequences into OTUs. Based on this threshold, the 108 clones could be grouped into 13 OTUs from each of which representative clones were sequenced on both strands. The clone library was dominated by OTU VI-type clones (Fig. 3), accounting for 64% of all retrieved sequences. As evident from rarefaction analysis (Fig. 1), further sampling of clones probably would have resulted in the detection of some additional OTUs.


Rarefaction curves for dsrAB sequences derived directly from the sediment. The five curves represent datasets in which the sequences were grouped into OTUs based on different nucleotide identity threshold values ranging from 85% to 98%.


16S rRNA gene and dsrAB sequence identity plot for 119 pure cultures of sulfite- and/or sulfate-reducing prokaryotes. Dotted lines indicate 97% 16S rRNA gene and 90%dsrAB sequence identity thresholds. Arrow a: Desulfovibrio desulfuricans (dsrAB: AJ249777, 16S rRNA gene: AF192153) vs. Desulfovibrio intestinalis (dsrAB: AB061539, 16S rRNA gene: Y12254). Arrow b: strain EtOH3 (dsrAB: DQ386236, 16S rRNA gene: DQ067421) vs. strain Benz (dsrAB: DQ386234–5, 16S rRNA gene: DQ386218). Arrow c: Desulfosarcina cetonica (dsrAB: AF420282, 16S rRNA gene: AJ237603) vs. strain oXyS1 (dsrAB: AF482465, 16S rRNA gene: Y17286). The cluster of data points in the lower left corner represents comparisons between members of the archaeal genus Archaeoglobus and bacterial species.


Consensus tree showing the phylogenetic relationship of deduced DsrAB amino acid sequences including sequences of isolates and representative clone sequences derived from enrichment cultures A–L (see Table 1) and directly from the sediment (all shown in bold). Parentheses show the number of sequences out of a total of 108 originating from the sediment assigned to the various OTUs. Square brackets show the OTU numbering. Members of lineages ah are shaded in grey, and the congruent 16S rRNA gene lineages are shown in Fig. 4. Short sequences (<460 amino acids) marked by (*) were added without changing the overall tree topology using maximum-parsimony criteria. Branches defined solely by short sequences are indicated by dotted lines. Nodes supported by bootstrap values >50%, calculated by distance-matrix analysis, are indicated by open circles (not shown for terminal groupings including short sequences). The consensus tree was inferred from distance-matrix, maximum-parsimony, and maximum-likelihood analysis; polytomies were introduced to connect branches, which were ambiguously resolved by the various treeing methods. The bar represents 10% sequence divergence as inferred from distance-matrix analysis.

Establishment of halophilic SRB enrichment cultures

A previous study of the Great Salt Lake 27% salinity sediment failed to derive enrichment cultures of SRB, which grew consistently at in situ salinity (K. Ingvorsen and K.K. Brandt, unpublished findings). Consequently, 19% salinity was chosen as the upper salinity limit for the enrichment cultures in the present study. At this and at 10% salinity, stable-growing sulfide-producing enrichment cultures were established from Great Salt Lake 27% salinity sediment for all substrates mentioned in Table 1. In general, the 19% salinity enrichment cultures reached lower cell densities and produced smaller amounts of sulfide than the 10% salinity enrichment cultures (Table 1). Enrichment cultures set up with methanol at 19% salinity often developed an overpressure, probably as a result of the formation of methane. In order to prevent methanogens from competing with SRB for substrates, parallel enrichment cultures amended with bromoethanesulfonic acid (BES), a specific inhibitor of methanogenesis (Oremland & Capone, 1987), were made.

Retrieval of 16S rRNA gene and dsrAB sequences from enrichment cultures

16S rRNA gene fragments were PCR-amplified from DNA extracts from enrichment cultures A–L (Table 1) with various primer pairs targetting a broad range of recognized SRB lineages. In summary, PCR amplicons of the expected size were obtained from all enrichment cultures using Desulfobacterales/Syntrophobacterales-selective primer pairs, and from all but one culture using Desulfovibrionales-selective primer pairs (Table 2). In contrast, only a few enrichment cultures produced PCR amplicons of the expected size using Desulfobulbaceae-, Desulfotomaculum-, and Desulfosporosinus-selective primer pairs, and no PCR amplicon was obtained from any of the enrichment cultures using a primer pair selective for the archaeon Archaeoglobus (Table 2). The PCR amplicons were subsequently cloned, and 25–30 clones from each clone library were screened for unique inserts by RFLP analysis. At least one member of each RFLP type from each clone library was later sequenced. The sequenced 16S rRNA gene clones grouped into 23 distinct OTUs using a 97% sequence identity threshold for grouping the sequences (Table 3 and Fig. 4). Twelve OTUs were affiliated with 16S rRNA gene sequences of known SRB (Fig. 4), while the remaining eleven OTUs were affiliated with lineages not known to contain SRB and therefore probably were the result of unspecific primer binding during PCR (Table 3). Notably, the primer pairs selective for Desulfotomaculum and Desulfovibrionales (Table 2) appeared highly specific, giving rise to none or very few clones affiliated with nontarget groups. In general there was <1.0% divergence among the sequences constituting the individual OTUs, the only exception being OTU 7 (Fig. 4), which, based on analysis of 625 unambiguously aligned nucleotide positions of 14 clone sequences, included four distinct sequence clusters differing by 1.5 to 2.9% identity (results not shown).

View this table:

16S rRNA gene sequence-based OTUs identified in enrichment cultures A–L not representing SRB

OTU no.*Closest relativeEnrichment culture§
1398%, Urania Basin clone U50II_21p1% (AY547953)G
1488%, Cascadia Margin clone Hyd89-72 (AJ535258)D, H, I, K
1592%, Guaymas Basin clone B01R010 (AY197381)D
1694%, Anaerobic digester clone AB16 (AF275926)A, C, E
1794%, Kalahari Shield subsurface water clone EV818BHEB5102502DRLWq27f023 (DQ256327)A, D, I
1899%, Halanaerobium saccharolyticum (X89070)B, F
1989%, Halanaerobium lacusrosei (L39787)D, L
2090%, Halocella cellulosilytica (X89072)L
2198%, Acetohalobium arabaticum (X89077)J, L
2299%, Bacillus mycoides (Z84583)I
2395%, Antarctica Vestfold Hills clone EKHO-5 (AF142889)D
  • * The OTU classification was based on a threshold sequence identity of 97%.

  • The sequences were classified into the respective lineages based on BLAST searches and phylogenetic analyses (see text).

  • The percentage sequence identity was determined by comparison with the GenBank database using BLAST searches.

  • § Enrichment cultures from which clones classified into the respective OTUs were retrieved (see Table 1).

  • Including the isolated strains Lac3 and Benz1.

PCR amplicons of dsrAB were obtained from all twelve enrichment cultures (A–L) and subsequently used for the construction of separate clone libraries. For each clone library, 30–40 clones were screened by RFLP analysis and selected for sequencing as described above. In total, ten distinct OTUs were detected among the sequenced clones using a 90% sequence identity threshold for grouping the sequences (Fig. 3). Up to 9% sequence divergence was evident among the sequences assigned to the individual OTUs. Most of the dsrAB sequences from the enrichment cultures were closely related to sequences retrieved from the original sediment sample, but only few were identical (Fig. 3).

Distribution of OTUs among enrichment cultures

A cluster analysis of the distribution of SRB-affiliated 16S rRNA gene OTUs among the enrichment cultures (F and H excluded) revealed no clear pattern (results not shown). Thus, the SRB communities of the 10% salinity enrichment cultures did not appear to be more closely related to each other than to those of the 19% enrichment cultures. Furthermore, enrichment cultures amended with the same substrate but differing in salinity did not result in a similar SRB-affiliated 16S rRNA gene-OTU composition. Equivalent results were obtained when analysing the distribution of dsrAB OTUs and a concatenated 16S rRNA gene- and dsrAB-OTU dataset (results not shown).

Isolation of pure cultures

Four distinct sulfate-reducing bacterial strains (PropA, EtOH3, Benz, and Lac2), and two fermentative bacterial strains (Benz1 and Lac3) were isolated from the enrichment cultures (Figs 3 and 4, Table 3). The deltaproteobacterial strains EtOH3 and PropA were characterized in detail and found to represent a novel species (now named Desulfohalobium utahense) of the genus Desulfohalobium and a novel genus within the family Desulfobacteraceae, respectively (Jakobsen et al., 2006; T.F. Jakobsen, K.U. Kjeldsen and K. Ingvorsen, unpublished findings). Strain EtOH3 grows at salinities ranging from >0 to 240 g L−1 NaCl, with an optimum around 100 g L−1. The corresponding values for strain PropA are >0–200 and 60 g L−1. Strain Benz had NaCl growth characteristics similar to those of its close relative, strain EtOH3. Strain Lac2, a close relative of the moderately halophilic Gram-positive SRB Desulfotomaculum halophilum (Tardy-Jacquenod et al., 1998), did not grow at 190 g L−1 NaCl. The two fermentative strains shared ≥99% 16S rRNA gene-sequence identity with the halophilic fermentative bacterium Halanaerobium saccharolyticum, which grows at salinities of up to 300 g L−1 NaCl (Zhilina et al., 1992).

Phylogeny of Great Salt Lake SRB

An interpretation of the combined 16S rRNA gene and dsrAB data from the original sediment sample and the enrichment cultures revealed 11 main SRB lineages (ak), roughly corresponding to eight distinct genera (Figs 3 and 4). The class Deltaproteobacteria harboured seven Great Salt Lake lineages, of which four branched within the family Desulfobacteraceae, two within the family Desulfohalobiaceae, and one within the family Desulfovibrionaceae. The remaining four lineages belonged to the family Peptococcaceae in the phylum Firmicutes, which comprises Gram-positive bacteria with a low G+C content.


Although dissimilatory sulfate reduction is considered an important (in some cases even the most important) process for the mineralization of organic matter in sulfate- and salt-rich environments (Caumette et al., 1994; Sørensen et al., 2004; Scholten et al., 2005), the diversity and phenotypic properties of halophilic SRB are largely unknown. The present study aimed at combining selective cultivation with 16S rRNA gene- and dsrAB-targeted molecular identification approaches in order to reveal the identity and physiology of SRB thriving in the extreme hypersaline (27% salinity) sediment of the northern arm of Great Salt Lake. The dsrAB-based profiling of the original sediment and the enrichment cultures uncovered a total of 20 distinct dsrAB OTUs, demonstrating the presence of a diverse SRB community in this inhospitable environment (Fig. 3). Many of the dsrAB sequences retrieved directly from the sediment were closely related to those obtained from enrichment cultures or isolates (Fig. 3). This indicates that mostly SRB representative for the in situ SRB community were cultured. Thus, as discussed below, phenotypic data, for example in vitro halotolerance, are now available for representatives of most SRB lineages detected in the extreme hypersaline sediment of Great Salt Lake.

Most of the dsrAB OTUs identified both in the sediment and in the enrichment cultures belonged to a single lineage (lineage d; Fig. 3), clustering within the family Desulfohalobiaceae, which also included the SRB Desulfohalobium retbaense and Desulfovermiculus halophilus as well as the strains EtOH3 and Benz. These SRB all share remarkably similar NaCl growth characteristics, representing the most halotolerant sulfate reducers described so far with upper limits and optima for growth at around 230–240 and 100 g L−1 NaCl, respectively (Ollivier et al., 1991; Belyakova et al., 2006; see above). In contrast to Desulfohalobium retbaense and strains EtOH3 and Benz, Desulfovermiculus halophilus is able to grow autotrophically (Belyakova et al., 2006), demonstrating that considerable metabolic diversity may exist among the members of lineage d despite a possibly conserved NaCl tolerance. Such metabolic diversity may help to explain the apparent coexistence of closely related yet genetically distinct members of this lineage within both the sediment and the enrichment cultures (Fig. 3). Because dsrAB is currently reported to exist only in a single copy within the genomes of sulfate reducers (Klenk et al., 1997; Larsen et al., 2001; Heidelberg et al., 2004; Rabus et al., 2004; Zverlov et al., 2005), the observed genetic diversity probably reflects true organismal diversity. It is likely that some of the OTUs identified may represent populations occurring at low densities in the sediment and in the enrichment cultures. For instance, strains EtOH3 and Benz were detected in enrichment cultures C and G containing acetate (Figs 3 and 4), despite probably being present in low numbers because neither of these strains is able to use acetate as sole carbon source for growth (Jakobsen et al., 2006; T.F. Jakobsen, K.U. Kjeldsen and K. Ingvorsen, unpublished findings). Furthermore, 69 out of the 108 dsrAB sequences retrieved directly from the sediment belonged to a single OTU of lineage d (no. VI, Fig. 3), which might suggest that a single population quantitatively dominated the SRB community of the sediment.

Members of lineage d appear widely distributed in extreme hypersaline systems, indicating an ecological significance of this lineage. Besides Desulfohalobium retbaense isolated from Lake Retba of 34% salinity (Ollivier et al., 1991), lineage d contains environmental 16S rRNA gene or dsrAB sequences originating from Mediterranean Sea brines of 20–26% salinity (van der Wielen et al., 2005), from a 20% salinity pond of a solar saltern at the Red Sea (Sørensen et al., 2005), and from a 25–32% salinity pond of the Salin de Giraud solar saltern (Mouné, 2003) (Figs 3 and 4). A similar widespread abundance is evident for members of lineage a, which includes several environmental 16S rRNA gene and dsrAB sequences originating from other hypersaline environments with salinities ranging from 8% to 29% (Eder et al., 2002; Loy et al., 2002; Donachie et al., 2004; Sørensen et al., 2005) (Figs 3 and 4). Strain PropA, the only cultured representative of lineage a, is slightly less halophilic than the cultured members of lineage d (see above), but nevertheless represents the most halophilic member of the family Desulfobacteraceae described so far.

Additional lineages of deltaproteobacterial SRB for which 16S rRNA gene and/or dsrAB sequences were recovered included lineage e belonging to the family Desulfohalobiaceae, lineage i, the representative sequence of which was almost identical to the 16S rRNA gene sequence of Desulfobacter halotolerans, a slightly halophilic SRB originating from Great Salt Lake (Brandt & Ingvorsen, 1997), as well as lineages b and j, representing novel SRB affiliated with the family Desulfobacteraceae and the genus Desulfovibrio, respectively (Figs 3 and 4). Moreover, the sediment dsrAB clone library contained a sequence (OTU V; lineage c) identical to the dsrAB sequence of Desulfocella halophila, another halophilic SRB originally isolated from Great Salt Lake (Brandt et al., 1999) (Fig. 3). This species was not detected in the enrichment cultures, probably because it is unable to utilize any of the six substrates used (Table 1, Brandt et al., 1999). The only nondeltaproteobacterial SRB lineages (f, g, h and k) identified were affiliated with the Gram-positive, spore-forming genus Desulfotomaculum (Figs 3 and 4). Lineages f, g and k probably represent novel SRB, although these sequences might theoretically also derive from non-sulfate-reducing syntrophs (Imachi et al., 2006) or sulfite-reducers.

The salinity (27%) of the Great Salt Lake sediment exceeds the upper in vitro salinity limit reported for growth of all known SRB species isolated from or inferred to be present in the sediment by the clone libraries (Figs 3 and 4). Thus, the sediment may contain niches of lower salinity (e.g. freshwater seeps), allowing these species to proliferate, or, alternatively, their in vitro halotolerance may differ from their in situ halotolerance. For example, Desulfovibrio halophilus is able to accumulate compatible solutes from the surrounding medium, thereby relieving salt stress (Welsh et al., 1996), a mechanism that might contribute to differences between in vitro and in situ growth performance at high salinities. The SRB community composition of the enrichment cultures did not correlate with the type of substrate or the salinity as judged by the cluster analysis of the distribution of dsrAB and 16S rRNA gene OTUs (see above). Furthermore, enrichment cultures at 19% salinity invariably produced lower amounts of sulfide than their respective 10% salinity counterparts (Table 1). These findings suggest that the cultivable members of the SRB community of the Great Salt Lake sediment were better adapted to growing at 10% than at 19% salinity, thus being moderate halophiles. This probably also applies to the entire SRB community, given the fact that the cultivable SRB cover most of the diversity detected by cultivation-independent methods in terms of close phylogenetic relatedness (Fig. 3).


This study represents the first reported investigation of the identity of SRB inhabiting the extreme hypersaline sediment of the northern arm of Great Salt Lake. The results demonstrate the presence of a diverse SRB community in this highly saline sediment, including novel genera and species, some of which were isolated in pure culture and subjected to physiological tests. All SRB identified in this study were affiliated with known SRB families, and many were furthermore closely related to isolates and phylotypes found in other hypersaline environments. As inferred from the in vitro data, the SRB inhabiting the Great Salt Lake sediment appeared poorly adapted to the high in situ salinity of 27%. In this respect, it is important to note that the northern arm of Great Salt Lake was characterized by more moderate salinities until it was separated from the southern arm by an embankment built in the late 1950s (US Geological Survey [http://www.usgs.gov]). Perhaps this time span is too short for the SRB populations to have fully adapted to the present-day high salinities of the northern arm. However, as pointed out by Oren (1999), it may simply be impossible for SRB to grow optimally at such high salinities owing to the low energy yield of dissimilatory sulfate reduction and the high energetic cost of maintaining the osmotic balance, especially if compatible solutes are used for osmoregulation.


The authors thank Christian Baranyi, Pernille V. Thykier, Britta Poulsen and Tove Wiegers for their excellent technical assistance and J. W. Gwynn from Utah Geological and Mineral Surgery for assistance in obtaining sediment samples. This research was financially supported by the Danish Technical Research Council under the framework program ‘Activity and Diversity in Complex Microbial Systems’. A.L. and M.W. were supported by the European Community (Marie Curie Intra-European Fellowship within the 6th Framework Program) and the bmb+f (project 01 LC 0021A-TP2 in the framework of the BIOLOG II program).


  • Present address: Trine Rolighed Thomsen, Section of Environmental Engineering, Aalborg University, Sohngaardsholmsvej 57, DK-9000 Aalborg, Denmark.

  • Editor: Gary King


View Abstract