TEXT

Top Abstract Text References

Sulfate-reducing bacteria (SRB) constitute a paraphyletic group of physiologically diverse anaerobes, which all share the ability to obtain energy from dissimilatory reduction of inorganic sulfate. SRB are widespread in nature and, besides being key organisms in the sulfur cycle, also play an important role in the degradation of organic matter in many anoxic environments (12, 13). SRB are also known to play a role in the biodegradation of a number of environmental pollutants (7). Their activity in oil reservoirs results in biocorrosion and other problems of great economic impact in the offshore oil and gas industries (9). Due to their great ecological and economic importance, SRB populations have been intensively studied during the last few decades. Although important for ecological studies, it is generally acknowledged that traditional culturing only recovers a very limited fraction of the natural microbial diversity (1, 8, 32), indicating that new cultivation procedures need to be developed. Recently, a new most-probable-number method was described which greatly improved the growth and recovery of SRB from sludge and sediments (3, 30).

Most of the molecular studies on the diversity of SRB in complex communities, e.g., in granular sludge (23), the water column of a stratified Fjord (28), biofilm (2), microbial mats (27), and sediments (6, 21, 22), have been based on small-subunit (SSU) rRNA gene analysis. However, retrieved SSU rRNA sequences frequently are not related to any cultivated organism, and it thus becomes impossible to infer a likely ecophysiology for the organism containing the gene. SSU rRNA-based phylogeny has revealed that SRB constitute a paraphyletic group with members among the -Proteobacteria, gram-positive bacteria, and even Archaea (4). Unlike monophyletic traits such as oxygenic photosynthesis, some lineages of sulfate reducers are closely related to organisms that are unable to carry out dissimilatory sulfate reduction. This reduces the ability to identify SRB solely based on their SSU rRNA sequence.

An alternative approach to infer physiology from environmental sequences is to retrieve functional gene sequences coding for enzymes that are essential to the target organisms. Dissimilatory sulfite reductase (DSR) is a key enzyme in sulfate respiration, which so far has been found only in dissimilatory SRB. Recently, partial sequences of the gene encoding for DSR were used to evaluate the phylogeny of SRB, and the derived phylogeny was found to be consistent with the SSU rRNA-based phylogeny (31). The PCR primerset used by Wagner et al. (31) amplifies most of the - and -subunits of the DSR gene and allows the detection of members of all known groups of SRB (31). Studies of natural diversity of SRB based on PCR amplification of the DSR gene have recently revealed environmental DSR sequences unrelated to any known SRB gene in a microbial mat (18) and in sediments (29).

In this study, we combined both selective SSU rRNA and DSR primers to investigate the diversity of SRB in brackish-water sediments (Kysing Fjord, Denmark). We compared the phylogeny derived from partial SSU rRNA and DSR sequences retrieved from clone libraries. Subsequently, SRB were cultivated from the site, and their phylogeny was established and compared to that obtained by molecular analysis of the same sample. The aim of this study was to (i) evaluate the diversity in the sample, (ii) determine the ability and/or failure of cultivation methods to retrieve common species, and (iii) evaluate PCR bias by comparing two different primersets.

Sediment samples were obtained from shallow-water sediments in Kysing Fjord (Norsminde, Denmark), previously shown to support high rates of sulfate reduction (11). Surface sediment, used as an inoculum for diversity studies, consisted of a mixture of oxidized (brown) and reduced (black) sediment collected from the top 0 to 4 cm. The organic-rich (~10%, dry weight) sediment mixture was initially stored at 4°C (in situ temperature) in the dark under anaerobic conditions in order to establish a large population of SRB. During storage, a further blackening of the sediment mixture was observed. Prior to making dilutions of the sediment sample, aliquots were frozen at 20°C for subsequent molecular analysis.

Kysing Fjord sediments (hereafter referred to as the inoculum) were 10-fold serially diluted in anoxic natural medium consisting of sterilized top layer sediment (0 to 2 cm) from the site (30). The inoculum was also 10-fold diluted in anoxic synthetic medium (33) containing a mixture of carbon sources (lactate, methanol, pyruvate, acetate, formate, and glucose, each at a 5 mM final concentration) and 0.1 g of yeast extract per liter. The final pH of the medium was ca. 7. Dilution tubes were sampled (0.5 ml) for molecular analysis after 3 weeks of incubation at 25°C. SRB from the highest positive dilution in natural media were reinoculated into different synthetic media, each containing only one of the carbon sources mentioned above (20 mM final concentration). Growth in culture tubes was monitored by measuring H₂S production (5). Purification of the SRB strains was performed by repeated application of the roll tube method (10). Anaerobic Hungate techniques (10, 15) were used throughout this work. Pure cultures and enrichments were characterized by molecular methods.

Genomic DNA was extracted by bead beating using the Bio 101 FastPrep Instrument, the FastDNA Spin Kit for Soil (inoculum and serial dilutions), and the FastDNA Kit MH (enrichments and pure cultures), according to the protocol of the manufacturer (Bio 101, Vista, Calif.). DNA extracts were stored at 20°C. PCR amplification was performed on extracted DNA or directly using a Thermocycler (PTC-200 Peltier Thermal Cycler; M. J. Research). SSU rRNA gene was amplified with the primerset 385F and 907R (Table 1). Reaction conditions were as follows: 93°C for 1 min, 25 cycles of 30 s at 92°C, 1 min at 57°C, and 45 s at 72°C, and a final extension of 5 min at 72°C. For denaturing gradient gel electrophoresis (DGGE) analysis, 35 PCR cycles and the primer 385F with a GC clamp (Table 1) were used. DSR gene was amplified with the primerset DSR1F and DSR4R (Table 1). Reaction conditions were as follows: 94°C for 1 min, 30 or 35 cycles of 20 s at 94°C, 1 min at 54°C, and 3 min at 72°C, and a final extension of 10 min at 72°C.

Partial SSU ribosomal DNA (rDNA) amplificates were separated by DGGE according to their melting behavior (19). We used the D-GENE DGGE system from Bio-Rad (Munich, Germany), and 8% acrylamide gels with a denaturating gradient ranging from 25 to 75% (100% denaturant was 7 M urea and 40% [vol/vol] formamide). Electrophoresis was conducted at a constant voltage of 200 V at 60°C for 7 h. Bands of interest were excised from the gel, eluted in 20 µl of sterile distilled water. Each band was subsequently reamplified, and its position was confirmed by DGGE before sequencing. For sequencing, bands were reamplified using the primerset without GC clamp.

PCR products were purified with the QIAquick PCR Purification kit (Qiagen GmbH, Hilden, Germany) and cloned using the TOPO XL Cloning Kit (Invitrogen, San Diego, Calif.). The clone library was screened by direct PCR amplification from a colony. Plasmids containing the insert of the right length were isolated using a QIAprep Miniprep Kit (Qiagen) and kept at 20°C.

Purified PCR products and plasmids were sequenced directly with an ALFexpress automated DNA sequencer, using a Thermosequenase Fluorescent Sequencing Kit (Pharmacia Biotech, Uppsala, Sweden). The primers used for sequencing are listed in Table 1. Partial SSU rRNA sequences were aligned with the sequence editor SeqPup (D. G. Gilbert, SeqPup version 0.6). Only unambiguously aligned sequence positions (485 nucleotides) were exported to the PAUP* version 4.0 program (D. L. Swofford, PAUP* 4th ed.; Sinauer Associates) for phylogenetic analysis. Different phylogenetic algorithms, i.e., maximum parsimony, distance matrix, and maximum-likelihood analysis, were applied to the dataset using default parameters. Partial - and -subunit sequences of the DSR gene were aligned according to the amino acid sequences with the sequence editor GDE implemented in the ARB package (26). Dendrograms were constructed for unambiguous amino acids (301 amino acids) using maximum parsimony, distance matrix, and maximum-likelihood analysis as implemented in the ARB package (26). The robustness of DSR and SSU rRNA trees was tested by bootstrap analysis with 100 resamplings.

Sequence accession numbers. All sequences have been deposited in GenBank under accession numbers AF360632 to AF360642 for SSU rRNA sequences and AF360643 to AF360694 for DSR sequences.

SRB growth. When Kysing Fjord sediment was serially diluted in natural media (sterilized sediments from the site), the growth of SRB was detected after 3 weeks of incubation down to the 10¹⁰ dilution, using both the SSU rRNA and the DSR primerset. No amplification was obtained using these selective primers in dilutions higher than the 10¹⁰ dilution or in samples from uninoculated natural media even after 12 weeks of incubation. When serial dilutions were made in synthetic media containing a mixture of common carbon sources for SRB, growth occurred only down to the 10⁶ dilution. These results, confirmed by subsequent chemical analysis of H₂S production, agree with recent studies which show that viable counts of SRB were up to 4 orders of magnitude higher when sediment media were used for enumeration (3, 30). Recently, Sass et al. (25) reported that only a minor fraction of the total community of SRB in a lake sediment was recovered by cultivation in a synthetic medium. Natural media seem to improve the recovery of both cultured and uncultured SRB from complex environments. However, the growth of SRB down to the dilution 10¹⁰ was unexpected since recovery of SRB in marine sediment is usually much lower (reference 30 and references therein). There could be several explanations for the high recovery we observed: (i) enrichment of SRB in the organic-rich sediment during the preincubation period, (ii) the presence of "hot spots" of SRB, and (iii) an event of low probability so that a SRB was diluted out to the 10¹⁰ dilution despite a small population size. High rates of sulfate reduction (up to 4 µmol cm³ day¹) have been reported for this organic-rich sediment (11). Using literature values of specific sulfate reduction rates of 10¹⁴ to 10¹⁵ mol of sulfate cell¹ day¹ (30) would yield population sizes of 4 × 10⁸ to 4 × 10⁹ cells cm³ operating at their V_max, indicating that the number of active cells may have been significantly higher.

DGGE analysis of serial dilutions. DGGE analysis of partial SSU rRNA gene amplificates showed only three bands in the last positive dilution (10¹⁰) (Fig. 1). Phylogenetic analysis showed that only one sequence (band 3) was related to the SRB within the -Proteobacteria. The sequence was highly similar (>99% similarity) to the SSU rDNA sequence of Desulfomicrobium baculatum (Fig. 2B). The sequences of bands 1 and 2 were affiliated with the gram-positive bacteria but were not related to any Desulfotomaculum sp. They are thus most likely not SRB, i.e., false positives. The DGGE pattern showed a general reduction in diversity for increasing dilutions; however, band 1 and band 3 were detected in the 10¹⁰ dilution but not in the lower dilutions (Fig. 1). In particular, band 3 corresponding to the D. baculatum sequence did not appear in lower dilutions. It is unlikely to be due to a preferential PCR amplification of SSU rDNA of organisms corresponding to the prominent bands in lower dilutions, since mixing DNA extracts from the dilutions 10¹⁰ and 10⁷ or 10⁹ did not inhibit the amplification of SSU rDNA of D. baculatum (data not shown). D. baculatum was probably overgrown by other organisms in the lower dilutions and therefore was present in too low numbers to be detected by DGGE analysis of SSU rDNA amplificates. These fast-growing organisms could have been less abundant in the inoculum than D. baculatum and therefore absent and unable to inhibit its growth in the 10¹⁰ dilution.

View larger version (84K): [in this window] [in a new window]   FIG. 1.   DGGE gel containing partial SSU rRNA gene amplificates obtained after serial dilutions of inoculum in natural media. Lanes show PCR amplificates of DNA extract from serial dilutions of the inoculum after 3 weeks of incubation. Denaturant concentration varied from 25% at the top to 75% at the bottom of the gel. The numbers indicate bands that were excised from the gel, reamplified, and sequenced.

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Comparison of the phylogenetic trees derived from retrieved DSR sequences (A) and SSU rRNA sequences (B). Species present in both trees are connected by horizontal lines. Environmental sequences and sequences from newly isolated strains are shown in boldface. The numbers on branching points are bootstrap values with 100 replicates (values of <50 were not included).

SRB diversity. The phylogeny of retrieved amino acid sequences (DSR gene) and nucleotide sequences (SSU rRNA gene) was estimated by distance matrix, parsimony, and maximum-likelihood analysis. Similar tree topologies were obtained with all phylogenetic methods. Figure 2 shows a comparison of DSR and SSU rRNA sequence-based trees obtained by distance matrix analysis. From the inoculum, 10 DSR clones and 23 SSU rRNA clones were sequenced. We retrieved a cluster of nine DSR sequences closely related to Desulfosarcina variabilis (Fig. 2A). One DSR sequence (INOC-DSR13) was deeply branching and was only distantly affiliated with two Desulfotomaculum species (Fig. 2A). However, it was closely related to a cluster of 14 DSR sequences previously retrieved in another brackish sediment in Denmark (Mariager Fjord) (data not shown). Only 6 of the 23 partial SSU rRNA sequences retrieved were branching within the -Proteobacteria (Fig. 2B). Five of these six sequences were related to known SRB, the closest known relative being Desulfosarcina variabilis (1 to 4.4% sequence similarity for partial rDNA sequence). The last -Proteobacteria sequence (INOC_RNA29) was only distantly related to Syntrophus buswelii (<90% sequence similarity for partial rDNA sequence).

10

7

7

10

10

5

6

2