OUP user menu

Seasonal changes in ribosomal RNA of sulfate-reducing bacteria and sulfate reducing activity in a freshwater lake sediment

Jian-hua Li , Kevin J. Purdy , Susumu Takii , Hidetake Hayashi
DOI: http://dx.doi.org/10.1111/j.1574-6941.1999.tb00558.x 31-39 First published online: 1 January 1999


Seasonal changes in the population of sulfate-reducing bacteria (SRB) in the profundal sediments of a freshwater lake, Lake Kizaki, Japan, were directly determined using oligonucleotide probes complementary to the 16S ribosomal RNA of the major phylogenetic groups of SRB. The results based on the hybridization indicated that relative 16S rRNA abundance (RNA index) of Bacteria and Gram-negative mesophilic SRB detected by probes EUB 338 and SRB 385 with respect to all known organisms as detected by probe UNIV 1400 in surface sediments (0–6 cm) were 71% and 4.1% on average for 15 months, respectively. The RNA indexes of the major SRB genera to all bacteria in the 0–3-cm and 3–6-cm layers, on average, were 1.4% for Desulfobulbus, 0.6% for Desulfobacterium, and 0.5% for Desulfovibrio. The RNA indexes of Desulfobulbus showed relatively high values in those of detected SRB during almost all of the study periods, while Desulfobacterium and Desulfovibrio exhibited low relative abundance. The RNA index of Desulfobulbus correlated with the rate of sulfate reduction in the sediment. Therefore, Desulfobulbus appears to be dominant in the active SRB population in the surface sediment.

  • Freshwater lake
  • Sulfate reduction rate
  • Sulfate-reducing bacterium
  • Oligonucleotide probe
  • Population dynamics

1 Introduction

Sulfate-reducing bacteria (SRB) are a distinctive and ubiquitous group of anaerobic prokaryotes. They are commonly encountered in marine and freshwater sediments, anoxic waters, soils, biofilms, and intestinal contents [1, 2]. Ecological studies of the diversity and abundance of SRB communities have been carried out by traditional culture methods. However, conventional methods such as most-probable-number estimations with selective enrichment media underestimate SRB population densities by as much as a 1000-fold [3, 4]. A recent study, with improved enrichment media, reported most-probable-number values for SRB more comparable with measured rates of sulfate reduction [5]. Yet, the in situ dynamics of SRB at the genus level remained mostly unknown due to the limitations in methodology.

Ribosomal RNA (rRNA), particularly small subunit rRNA, sequence data now give a means to circumvent many of these difficulties by providing unambiguous features (nucleotide sequences) of identity. Over the past decade, reliable methods for determining microbial phylogeny based on gene sequence information [6] and methods for recovering genes directly from environmental samples [7, 8] have been developed for studies in both determinative and environmental microbiology. Using probes based on these sequence data, it is possible to identify and classify the members of a microbial community in situ [910, 11]. This general approach was first applied to ecological studies of ruminal bacterial populations [12], and is well suited to the analysis of microbial populations whose members, such as SRB, are difficult to isolate or grow in the laboratory.

The application of 16S rRNA-targeted oligonucleotide probes (except in situ hybridization) is based on the direct extraction of nucleic acids from environmental biomass [7, 13]. Recently, various methods for direct extraction of nucleic acids from environmental samples of soils or sediments were developed [1418], but a standard method has not yet been established. In the present study, a newly developed hydroxyapatite (HTP) spin column method as described by Purdy et al. [19] was used. The method has a relatively high efficiency for nucleic acid extraction and has been demonstrated to be suitable for nucleic acid extraction from aquatic sediments.

The aim of the study is to analyze the population dynamics of SRB in profundal sediments of Lake Kizaki using 16S rRNA-targeted oligonucleotide probes. At the same time, sulfate reduction rate was measured along with environmental factors and its relationship to the results of probing for SRB is discussed.

2 Materials and methods

2.1 Sediment sampling and analysis

Lake Kizaki is a freshwater mesotrophic lake in central Japan. The lake has a surface area of 1.4 km2 with maximum and mean depths of 29.5 m and 17.9 m, respectively. Stable stratification of lake water occurs from early summer to late autumn every year. Sampling was carried out at the central station (depth: approximately 29 m) of Lake Kizaki from May, 1994 to August, 1995.

The sediment samples for nucleic acid extraction (20 g for each layer) were washed with an equal amount (w/v) of 120 mM sodium phosphate buffer (pH 8.0) three times to remove any extracellular nucleic acid [16]. Subsamples of the washed sediment samples (about 2 g each) were stored in baked aluminum-foil at −20°C until nucleic acid extraction.

2.2 Determination of sulfate reduction rates and sulfate concentration

The sulfate reduction rates in the sediments were measured in triplicate for each layer using a radiotracer method [20]. The concentration of sulfate ions in bottom water and interstitial water was determined by ion chromatography (Ion chromatograph Qic, column: AS4A, Dionex, Sunnydale, CA, USA). Interstitial water in sediments was obtained by centrifugation (9300×g at 4°C) followed by filtration with membrane filters (pore size 0.22 μm).

2.3 Extraction of nucleic acids from sediment samples

Extraction of nucleic acids was performed by the HTP spin column method as described by Purdy et al. [19] with the following pretreatment before bead-beating. Frozen sediment samples were thawed for 10–20 min at room temperature immediately prior to processing. The samples for each sediment layer were divided into 3 aliquots of 0.5 g each, and placed in sterile 2.0-ml screw-top Eppendorf tubes containing 0.5 ml of lysing reagent (50 mM Tris (pH 8.0), 0.01 mM EDTA (pH 8.0), 2.5% sucrose, 10 mM sodium pyrophosphate). The vials were incubated in a water bath at 65°C for 30 min with occasional gentle mixing and then frozen at −80°C for 30 min. The samples were thawed again at 65°C for 10 min, then the mechanical lysis and purification procedures were carried out according to Purdy et al. [19].

The extracts of nucleic acids were quantified and assayed for purity by scanning UV spectroscopy [15]. All solutions used for the extraction were pretreated with 0.1% diethylpyrocarbonate (DEPC) (Sigma, St. Louis, MO, USA) to inactivate nucleases [15].

2.4 Extraction of rRNA from pure culture controls

Pure cultures of sulfate-reducing bacteria were grown as described by Widdel and Bak [21]. These cultures were Desulfovibrio vulgaris DSM 644, Desulfovibrio desulfuricans DSM 1926, Desulfobulbus propionicus DSM 2056, Desulfococcus multivorans DSM 2059, Desulfosarcina variabilis DSM 2060, Desulfobacter latus ATCC 43918, Desulfobacterium autotrophicum DSM 3382. Non-SRB controls (E. coli IAM 1264, B. subtilis NCMB 3610 and Saccharomyces cerevisiae IAM 4206) were grown in nutrient broth for bacteria and malt extract broth for a yeast. Mid-exponential phase cultures (as determined by nephelometry) were sampled (50 ml) and centrifuged at 12 500×g. The resultant cell pellet was resuspended in 120 mM sodium phosphate, pH 8.0+10% (w/v) glycerol and stored in 0.5 ml aliquots at −20°C. rRNA from these aliquots was extracted as described previously [22].

2.5 Immobilization of extracted nucleic acids onto nylon membranes

Known quantities of nucleic acids were slot-blotted onto nylon membranes (Hybond-N, Dupont, Stevenage, UK) using a BioBlot manifold (Bio-Rad, Hercules, CA, USA) after denaturation of rRNA by addition of 2% (w/v) glutaraldehyde [12]. Measured amounts of pure rRNA from pure culture positive controls specific to each probe to be utilized and negative controls were also immobilized on the nylon membranes.

2.6 Oligonucleotide probes and hybridization conditions

The oligonucleotide probes targeting all known organisms (UNIV 1400) [12], general bacteria (EUB 338) [23] and Proteobacteria delta subgroup containing all mesophilic, Gram-negative SRB genera (SRB 385) were used in these experiments, along with SRB genus specific probes described by Devereux et al. [24](probe 221, Desulfobacterium; probe 660, Desulfobulbus; probe 687, Desulfovibrio; probe 129, Desulfobacter). Oligonucleotide probes were synthesized on an automatic synthesizer (381A DNA synthesizer, Applied Biosystems, Warrington, Cheshire, UK) and end-labeled with γ-32P-ATP (ICN Biomedicals, Costa Mesa, CA, USA) [15]. Hybridizations were performed as described by Stahl et al. [12]. Membranes were sealed into bags with approximately 0.1 ml of hybridization buffer (0.9 M NaCl, 50 mM sodium phosphate, pH 7.0, 5 mM EDTA, pH 8.0, 10× Denhardts and 0.5 mg ml−1 Poly(A) (Sigma-Aldrich, Tokyo, Japan)) per cm2 of membrane added. Membranes were prehybridized at 40°C for at least 2 h before addition of radiolabeled oligonucleotide. Incubation was continued for at least 16 h, after which the membranes were washed once with 1× SSC, 1% SDS for 30 min at 40°C and then in fresh wash buffer at the specific temperature for 30 min [23, 24]. Autoradiographs (XAR-5, Eastman Kodak, Rochester, NY, USA) were exposed for 1–10 days at −70°C and developed manually.

Membranes were stripped after autoradiography in order to reuse the membrane. Boiling 0.5% (w/v) SDS was poured over the membranes to be stripped and then left to cool to room temperature. The stripped membranes were autoradiographed as above to determine if all the bound probe had been removed. If any signal remained after stripping, membranes were stripped once more. If signal was still visible the membrane was discarded. Successfully stripped membranes were reprobed by the method described above.

2.7 Scanning densitometry of autoradiographs and determination of signal levels

Autoradiographs were quantified using densitometry (Shimadzu CS-930, Kyoto, Japan). Probes do not all bind equally well to target sequences and variations in labeled probe specific activity results in different signal levels. Therefore, in order to standardize results of blots probed with different probes, densitometrically measured signals were converted to an amount of rRNA for each individual sample by comparison to the signal from known amounts of rRNA from pure culture control organisms. These standardized results were then expressed as a percentage (RNA index) of the signal (expressed as ng of rRNA detected) for either the universal probe (UNIV 1400) or the general bacterial probe (EUB 338) for each individual sample.

3 Results

3.1 Physical and chemical properties in bottom water and sediment

Temperature in the bottom water remained around 6°C throughout the year except a short period in spring overturn. Dissolved oxygen below a depth of 26 m was depleted from August to early December.

The physical and chemical properties of the sediments are summarized in Table 1. Water content and organic content (ignition loss) in sediments were the highest in the surface layer and decreased with depth. As shown in Fig. 1A, the sulfate concentration in the interstitial water of the sediment also decreased with depth. Seasonally, the concentration in the 0–3-cm layer increased in spring overturn and decreased gradually until the end of stratification. The maximum value of 31.8 μmol l−1 in July, 1995 was due to a flood of record proportions which caused a large amount of sediment to be deposited in the lake [25].

View this table:

Physical and chemical properties of profundal sediments in Lake Kizaki

Depth of sediment (cm)Water content (%)Ignition loss (%)Sulfate concentration in porewater (μmol l−1)
  • a The figures indicate standard error (n= 25).


Seasonal and vertical changes in (A) sulfate concentration (μmol l−1) in interstitial water and (B) sulfate reduction rate (nmol ml−1 day−1) of profundal sediment (0–12 cm) in Lake Kizaki.

3.2 Sulfate reduction rate

The seasonal fluctuation of depth profile in sulfate reduction rate is shown in Fig. 1B. In most cases, the maximal rate was found in the surface layer irrespective of the concentration of dissolved oxygen in the bottom water. The rate decreased with depth and was almost always lower than 0.5 nmol ml−1 day−1 in the layer below 6 cm. In spite of a usually constant in situ temperature (about 6°C), the rate was high from spring to summer and decreased from autumn to winter. The seasonal change in sulfate reduction rate was positively correlated with the sulfate content of sediment [20]. The highest values (ca. 13 nmol ml−1 day−1 in the surface layer) in June and July, 1995 were measured just after the flood mentioned above, and may have been stimulated by sulfate and organic carbon from the floodwater.

3.3 Extraction of nucleic acid from sediment

The pretreatment by freeze-thaw treatment greatly improved the permeability of the extracts from sediments through the HTP spin column and the purity of extracted nucleic acids. Bands corresponding to 16S and 23S rRNA were often clearly visible by ethidium bromide stained agarose gel electrophoresis. While A260/A280 and A260/A230 ratios in some extracts were lower than 1.8 and 1.6, respectively, this appeared to have no effect on hybridization.

The average amount of total 16S rRNA in sediments estimated from hybridization with UNIV 1400 probe showed the highest value in the 0–3-cm layer, and decreased with depth (Fig. 2), although the fluctuation is large (mean 2.0 μg g−1 wet sediment, S.E.=0.4 (n= 15) in the 0–3-cm layer). This indicates that microbial communities were the most active in the surface layer.


Vertical distribution of total small subunit rRNA extracted in profundal sediment (0–12 cm) of Lake Kizaki. Horizontal bars indicate standard error (n= 15) of the values measured in the study period.

3.4 Abundance and dynamics of SRB rRNA index

As recovery of RNA from the sediments may not be quantitative, probing results are described in terms of relative abundance (rRNA index) [12, 26]. Probes designed to cultured members of a genus may not include the true extent of genetic diversity of that genus, or may cross-react with strains from outside their target genus that are, as yet, uncharacterized [12, 27]. Therefore signal from probes represents the extent of binding of target-identical species which will include a proportion of the targeted genera.

The seasonal changes in rRNA indexes of signal from EUB 338 (Bacteria) and from SRB 385 (Proteobacteria delta subgroup containing all Gram-negative mesophilic SRB) with reference to signal from UNIV 1400 (total rRNA) are presented in Fig. 3. On average, the rRNA index from EUB 338 accounted for 77% and 67% in the 0–3-cm and 3–6-cm layers, respectively, while rRNA from SRB 385 accounted for about 4% in both the layers over the study period. Therefore, SRB rRNA detected by SRB 385 accounted for less than 6% of detected bacterial rRNA on average. Seasonally, relative bacterial rRNA (EUB 338) showed high proportions in summer, although significant fluctuations were observed. Also, the rRNA index of the signal from SRB 385 with respect to the signal from EUB 338 increased from spring to autumn in both years.


Seasonal changes in relative abundances of signal detected by EUB 338 (Bacteria, open circle) and signal detected by SRB 385 (Gram-negative mesophilic SRB, open square) in the 0–3-cm layer (A) and the 3–6-cm layer (B) in nucleic acid extracted from profundal sediment of Lake Kizaki. Data is expressed as a percentage of the total small subunit rRNA detected by UNIV 1400 (all organisms).

Seasonal changes in the SRB rRNA indexes at the genus level with respect to the signal from EUB 338 are shown in Fig. 4. In most cases, the index of Desulfobulbus as detected by probe 660 was the largest compared with those of other SRB genera detected with probes used. Its rRNA index comprised 1.62% of the EUB 338 signal in the 0–3-cm layer and 1.20% in the 3–6-cm layer on average. The maximum value reached was 4.82% in July, 1995 when the highest rate of sulfate reduction was recorded (Fig. 1B). The RNA index of Desulfobacterium detected by probe 221 was second largest on average, 0.69% and 0.46% in the 0–3-cm and 3–6-cm layers, respectively. rRNA of Desulfovibrio as determined by probe 687 was also detected always in the study period, but its rRNA index was lower compared with the above two genera in the 0–3-cm layer. It was lower in the 0–3-cm layer (0.39%) than in the 3–6-cm layer (0.65%), in contrast to the above two genera.


Seasonal changes in relative abundances of three SRB genera targeted rRNA in total bacterial rRNA in nucleic acid extracted from the 0–3-cm layer (A) and the 3–6-cm layer (B) of profundal sediment in Lake Kizaki. Data is expressed as a percentage of the EUB 338 (all bacteria) signal.

As the signal from probe 129 (Desulfobacter) was faintly detected in only two samples, Desulfobacter, an obligate marine SRB genus, was substantially absent in the freshwater sediments.

A positive correlation was found between the rRNA index with respect to the signal from EUB 338 of either the signal from SRB 385 (Fig. 5A, P<0.002) or of the signal from probe 660 (Desulfobulbus) (Fig. 5B, P<0.001) and the measured sulfate reduction rate in the study period.


Relationship between rRNA index to total bacterial rRNA for (A) signal from SRB 385 or (B) signal from p 660 (Desulfobulbus) and sulfate reduction rate in the 0–3-cm layer (solid circle) and the 3–6-cm layer (open circle) of Lake Kizaki sediment.

4 Discussion

Many studies have demonstrated the applicability and usefulness of 16S rRNA targeted oligonucleotide probes described by Devereux et al. [24] and Amann et al. [28] to analyze the distribution and community structure of SRB in anaerobic biofilms, microbial mats and sediments [22, 23, 27, 2931]. The present study, by improving a method to extract nucleic acids from sediments, demonstrated the community structure and dynamics of SRB in lake sediments with low rates of sulfate reduction using 16S rRNA-targeted oligonucleotide probes.

Sulfate reduction rates in profundal sediments of Lake Kizaki were relatively low compared with other mesotrophic lakes due to low concentrations of sulfate in the sediment [20]. However, a large supply of sulfate along with organic matter was introduced by a record flood and successive inflow which occurred in the summer of 1995 [25]. This stimulated the highest rate of sulfate reduction in the sediment (13 nmol ml−1 day−1 in June and July, 1995) observed in five years (Fig. 1B, [20]).

The yield and purity of the nucleic acids extracted by the HTP spin column method [19] were unsatisfactory because of high contents of humic substance in the sediment samples from Lake Kizaki. Furthermore, it was very difficult to run the extracted samples through HTP spin columns. To raise the purity of the nucleic acids and the permeability of the extracts through the column, a pretreatment of a freeze-thaw cycle was introduced along with the addition of DEPC to inactivate nucleases [18]. The pretreatment was shown to greatly improve both the purity of extracted nucleic acids and the running of the sample through the column. Nevertheless, both the purity coefficients (A260/A230 and A260/A280) of some extracts were lower than 1.6 and 1.8, respectively, but the results of the hybridization did not seem to be affected.

The average RNA index of all Gram-negative mesophilic SRB detected by the probe SRB 385 was 4.2% of all organisms small subunit rRNA detected by UNIV 1400 and 5.4% of all bacterial 16S rRNA by EUB 338 in the upper 0–6-cm layer. The values are comparable to the maximum from a depth profile of an estuarine sediment reported by Devereux et al. [27]. However, the absolute amount of SRB rRNA should be much smaller in Lake Kizaki than in the estuarine sediment, because the total small subunit rRNA content in the sediment (about 2 μg g−1 wet weight, Fig. 2) was less than a third of that reported by Devereux et al. [27].

Desulfobulbus rRNA as determined by the signal from probe 660 was found to be predominant in SRB rRNA in both the 0–3-cm (1.62%) and 3–6-cm (1.20%) layers (Fig. 4). The predominance of detected Desulfobulbus rRNA increased in summer and reached 4.82% of signal from EUB 338 in July, 1995. Relative signal from probe 660 (Desulfobulbus) corresponded well with the rate of sulfate reduction in these sediments, indicating a link between sulfate reduction rate and Desulfobulbus activity (Fig. 5B). Desulfobulbus spp. are known to prefer propionate as a substrate and incompletely decompose it to acetate, although they are also capable of using a few other simple substrates like lactate, ethanol and H2 as electron donors [21]. In sediment slurries prepared from the Lake Kizaki added propionate was more rapidly utilized compared with acetate and butyrate (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996). Also, propionate seemed to be used only by SRB, because the addition of molybdate, a specific inhibitor of sulfate reduction [34], completely inhibited propionate utilization in the slurry (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996). Furthermore, two strains of Desulfobulbus isolated from cultures enriched with either propionate or butyrate as a sole carbon source exhibited the highest growth rate on propionate as a sole substrate (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996). These support the probing results which indicate that Desulfobulbus was the most dominant SRB genus detected in the surface layer (0–6 cm) of sediment from Lake Kizaki.

The second most prevalent SRB rRNA was detected by probe 221 which is designed to target Desulfobacterium spp. Desulfobacterium spp. utilize various kinds of simple organic compounds along with H2 and completely oxidize acetate to CO2. Although most of Desulfobacterium spp. require a high concentration of NaCl, a freshwater strain has been reported [21]. The Desulfobacterium detected may oxidize acetate in situ, because acetate utilization in the slurry experiments was severely inhibited by addition of molybdate (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996). However, the isolation of any members of this genus was not successful. Although Desulfotomaculum acetoxidans is known as a freshwater acetate utilizer, no probe specific to the species is available at present.

The third dominant rRNA was from Desulfovibrio as detected by probe 687. Unlike the above two genera, Desulfovibrio rRNA was relatively less abundant in the 0–3-cm layer than in the 3–6-cm layer and was most abundant during the mixing period in spring. Desulfovibrio has been reported to be the dominant species in marine and freshwater sediments utilizing electron donors such as lactate, ethanol, pyruvate and H2[5, 32, 33]. Devereux et al. [27] and Trimmer et al. [31] also reported that Desulfovibrio rRNA was dominant in SRB rRNA in estuarine sediments. However, the present study demonstrated that Desulfovibrio rRNA occupied rather a minor part of SRB rRNA in the sediment of Lake Kizaki. In slurry experiments the rapid utilization of lactate was only slightly inhibited by molybdate addition (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996) as reported by Purdy et al. [22] for sediment slurries from a tidal river, which suggests that Desulfovibrio species are not the sole lactate utilizers in the sediment of Lake Kizaki, although pure cultures of Desulfovibrio utilizing lactate were obtained from the sediment samples (J.-H. Li, PhD thesis, Tokyo Metropolitan University, 1996). Purdy et al. [22] showed by 16S rRNA probing that added lactate along with other volatile fatty acids did not stimulate the growth of Desulfovibrio in sediment slurries from a tidal river.

A relatively high value of RNA index in signal from SRB 385 or in signal from probe 660 (Desulfobulbus) was often found in summer just after stratification. This tendency agreed well with the peak of the rate of sulfate reduction. According to statistical regression analysis, a positive correlation was found between the RNA index of signal from SRB 385 (P<0.002) or of Desulfobulbus from probe 660 (P<0.001) to general bacterial signal from EUB 338 and the sulfate reduction rate (Fig. 5). When the sulfate reduction was not observed in Fig. 5, signal from probe 660 was not detected, while considerable signal from SRB 385 was detected. This suggests that rRNA from the sediments detected by SRB 385 contained those of other members of Proteobacteria delta subgroup than SRB [34]. Also the dominant active SRB in the sediment seems to be Desulfobulbus. Generally the rRNA content of a cell increases with increasing growth rate [35, 36]. And so at higher rates of sulfate reduction greater levels of rRNA would be expected. However, the correlation between sulfate reduction rate and sulfate-reducing bacterial numbers enumerated by the MPN method was not obtained in a previous study in this lake (J.-H. Li, Master thesis, Shinshu University, 1993). Thus, compared with conventional cultural methods, this approach, using 16S rRNA targeted oligonucleotide probes, is recognized as a powerful method for studying the distribution and dynamics of active bacterial populations in the environment.


This work was partly supported by Grants-in-Aid (no. 07680560 and 09680509) of the Ministry of Education, Science and Culture, Japan.


  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].
View Abstract