OUP user menu

Microbial diversity and functional characterization of sediments from reservoirs of different trophic state

Axel Wobus, Catrin Bleul, Sebastian Maassen, Carola Scheerer, Markus Schuppler, Enno Jacobs, Isolde Röske
DOI: http://dx.doi.org/10.1016/S0168-6496(03)00249-6 331-347 First published online: 1 December 2003

Abstract

Sediment samples from four reservoirs of different trophic state were compared with regard to chemical gradients in the pore water, composition of microbial communities and extracellular enzyme activities. The trophic state was clearly reflected by steep vertical concentration gradients of ammonium and alkalinity in the pore water. A high concentration of these parameters indicated a high microbial in situ activity in the more eutrophic reservoirs. However, the total number of bacteria in sediments seemed hardly to be influenced by the trophic conditions in the water column. Differences in the microbial composition of the sediments became evident by comparative 16S rDNA analysis of extracted DNA and by fluorescence in situ hybridization. Although a high proportion of the cells detectable with the EUB probe could not be identified at the subdomain level, members of the β-Proteobacteria constituted an important fraction in the sediments of the more eutrophic reservoirs, whereas γ-subgroup Proteobacteria were most frequently detected in sediment samples from the dystrophic Muldenberg reservoir. The assessment of extracellular enzyme activities (esterases, phosphatases, glucosidases and aminopeptidases, respectively) in sediment samples of the four reservoirs revealed specific patterns of metabolic potentials in accordance with the trophic state and characteristics of the catchment.

Keywords
  • 16S rDNA
  • Extracellular enzyme activity
  • Fluorescence in situ hybridization
  • Microbial diversity
  • Pore water
  • Sediment
  • Trophic state

1 Introduction

The bottom sediment of a reservoir is the compartment with the highest concentration of most chemical constituents as well as the highest microbial biomass. Therefore, abiotic and microbial processes in the sediment have a major influence on the nutrient balance in the water body, e.g. the phosphorus cycling [1]. The microbial degradation and transformation of particulate and dissolved organic matter in the sediment are key processes with regard to the carbon cycling in aquatic systems. A strong coupling between the deposition of organic material and bacterial activities in the sediment has been shown for both marine and freshwater systems [24]. According to the fundamental role of extracellular enzymes in the processes of organic matter decomposition, the patterns of their activities provide information on the available organic matter sources in natural systems [5].

Whereas knowledge of the function of sediments in aquatic ecosystems was gained by the determination of enzyme activities as well as by analysis of chemical gradients in the interstitial water, the species composition and structure of bacterial communities in sediments remained largely unknown. Traditional culture-based approaches failed to open the black box of sediment microbial diversity due to selective recovery of only a small subpopulation of the microbial communities. In the last years, the application of molecular biological methods, particularly based on amplification of extracted nucleic acids, enhanced the knowledge of bacterial diversity in different ecosystems (see, for example, reviews [6,7]). A number of studies based on the analysis of 16S rDNA sequences provided an insight into bacterial diversity and composition of marine pelagic and benthic communities (e.g. [812]). In contrast, the microbial composition of lake and other freshwater sediments was investigated only in a few studies. The results of these studies have expanded the knowledge of sequence diversity in different subpopulations like methanogens, methanotrophs or sulfate-reducing bacteria and led to a better understanding of their ecological role [1315]. Moreover, novel bacterial lineages were identified as active constituents of sediment communities [16,17].

Whereas polymerase chain reaction (PCR)-based techniques chiefly provide information on the qualitative community composition, i.e. on the incidence of a distinct genotype, population dynamics can be monitored in a quantitative way only by means of dot- or slot-blot hybridization or fluorescence in situ hybridization (FISH). The latter technique became a powerful tool of microbial ecologists for the investigation of various environments, from waste water treatment to drinking water systems as well as of complex systems like soils (see, for example, reviews [6,18,19]). However, FISH was applied only in a few studies to analyze the microbial communities of sediments [10,2022].

Since sediments are characterized by a high heterogeneity of environmental conditions at very small scales and thus provide various ecological niches, we expected a high microbial diversity in sediment communities. The aim of this study was to raise the bacterial inventory of sediments from reservoirs of different trophic state. Therefore, 16S rDNA clone libraries were established on the basis of extracted total DNA. By the comparison of the bacterial community composition of four reservoir sediments, the influence of trophic conditions in the water on the microbial communities in the sediment should be investigated. Since clone libraries are not quantitative indicators of microbial abundance and may also result in a qualitatively misleading portrait of the microbial diversity, the dominance structure of sediment communities was examined by FISH additionally to the comparative 16S rRNA analysis. Moreover, the sediments were characterized by chemical gradients in the pore water and extracellular enzyme activities in order to provide a comprehensive insight into sediment structure and function.

2 Materials and methods

2.1 Sampling sites

Sediment cores from four different reservoirs in Saxony (Germany) were examined. The reservoirs under investigation represent a broad range of trophic conditions and, with exception of the Quitzdorf reservoir, are used for drinking water supply (Table 1).

View this table:
1

Characteristics of the reservoirs under investigation

ReservoirNeunzehnhain (Neu)Muldenberg (Mul)Saidenbach (Sai)Quitzdorf (Qui)
Trophic stateaoligotrophicoligotrophic; dystrophic, acidicmesotrophichighly eutrophic
Catchment80% forest, 18% agriculture98% forest19% forest, 73% agriculture29% forest, 64% agriculture
Surface area (ha)8.592146750
Max. depth (m)16.523489.3
Secchi depth, range (m)4.7–102.1–3.92.4–7.00.4–2.9
Total phosphorus, range (mg l−1)0.005–0.0140.004–0.0160.007–0.060.04–0.3
Chlorophyll, range (μg l−1)0.7–4.90.3–10.31.5–11.55.9–141
  • aThe trophic state was examined according to German standards including the total phosphorus concentration, the chlorophyll content and the Secchi depth.

The drainage basin of Muldenberg mainly consists of bogs, which cause an increased input of humic substances into the reservoir. The associated brown color of the water entails a reduced Secchi depth. Since the 1970's, acidification by an increased atmospheric load was significant in Muldenberg, which, however, decreased again. Accordingly, the average pH of the water increased from pH 3.5 in 1990 to pH 5 in the year 2000.

The reservoir Quitzdorf has a very shallow mean depth of 2.8 m and is characterized by mass growths of Microcystis during the summer months. Due to its shallow depth, the entire water body within this reservoir warms in the summer months, and no stable stratification occurs.

The reservoirs were examined at intervals of about 2 months. Sediments were sampled at the greatest water depth, using a sediment corer (Kajak Sampler, diameter 6 cm, Uwitec, Mondsee, Austria). In the laboratory, the cores were sliced into layers of 0.5 or 1 cm thickness, respectively. In order to obtain sufficient material for pore water analysis, 10 sediment cores were pooled for each sampling date. The interstitial water was extracted by centrifugation at 11 000 rpm and filtration (pore size 0.2 μm, Pall Corp., East Hills, USA). During the summer months it was impossible to obtain pore water from the Quitzdorf samples by centrifugation without a very high content of Microcystis colonies, therefore, the separation was performed with dialysis samplers according to Hesslein [23] (diaphragm diameter 0.2 μm).

For microbiological examination and determination of enzyme activities, aliquots of the sliced fresh material were diluted and homogenized by sonication for 20 s (Sonopuls HD 60, Bandelin electronic, Berlin, Germany). The dry matter content of sediments was determined after drying of the sliced samples overnight at 105°C and the organic matter content was obtained as the subsequent loss of weight after at least 2 h at 550°C.

2.2 Chemical analysis of pore water

The chemical examination of interstitial water was performed according to German standards for water analysis (DIN, 1982–1994): The concentrations of ammonium and phosphate (soluble reactive phosphate, SRP) were analyzed photometrically (U-2000, Hitachi Ltd., Tokyo, Japan), dissolved organic carbon (DOC) was measured using a total organic carbon analyzer (DC-190, Tekmar-Dohrmann, Mason, OH, USA). Inorganic anions (nitrate, nitrite, sulfate) were analyzed by ion chromatography (IC 20 ion chromatograph equipped with eluent generator EG 40, Dionex Corp., Sunnyvale, CA, USA) by means of an IonPac AS15 column (with an ASRS ultra suppressor). Alkalinity was determined by titration with 0.01 M HCl. Vertical profiles of oxygen concentration were measured by means of a Clark-style microelectrode (Kurt-Schwabe-Institut, Meinsberg, Germany) within the intact sediment cores.

2.3 Microbiological methods

To determine the total number of bacteria, sediment samples were preserved immediately after slicing with formaldehyde (final concentration 3.7%) and kept at 4°C for at least 1 h or overnight. Thereafter, the samples were homogenized by a vortexer and were diluted with particle-free and autoclaved water. As nucleic acid stain, SYTOX-Green (Molecular Probes Inc., Eugene, OR, USA) was used at concentrations between 1 and 2.5 μM. The samples were incubated with the stain in the dark at 4°C for about 20 min. The stained samples were homogenized by sonication for at least 20 s (Sonopuls HD70 with an MS73 probe, Bandelin electronic, Berlin, Germany) and then filtered onto black polycarbonate filters (Millipore Corp., Bedford, MA, USA, pore size 0.2 μm) under low vacuum pressure (<30 mm Hg). The dried filters were mounted on microscopic slides with Citifluor (Citifluor Ltd., London, UK) and were examined by using an epifluorescence microscope (Axioskop, Carl Zeiss, Oberkochen, Germany, equipped with a Plan NEOFLUAR 100/1,3 oil objective) under immersion. A combination of specific optical filters was used for counting the SYTOX-Green-stained cells (B 450/40 exciter filter, FT 510, LP 515/50). For statistical evaluation, at least 20 randomly selected microscopic ocular grid fields per sample were examined and a minimum of 200 cells was counted. For all samples, bacteria were enumerated on two replicate filters.

For FISH, aliquots of the sliced and homogenized sediment samples were fixed for at least 3 h with freshly prepared paraformaldehyde–phosphate-buffered saline (PBS) (final formaldehyde concentration about 3%) and with absolute ethanol, respectively (according to [24]). The latter procedure was used in order to permeabilize cells of Gram-positive bacteria prior to hybridization. Subsequently to fixation, the samples were washed and stored in PBS–ethanol (1:1) at −20°C.

The whole-cell hybridization of sediment samples was performed according to standard protocols as described by Manz et al. [24] and Snaidr et al. [25]. Prior to immobilization on glass slides, samples were diluted and homogenized by mild sonication (see above). The oligonucleotide probes were labeled with the Cy3 fluorochrome at the 5′ end (Interactiva Biotechnologie GmbH, Ulm, Germany). Final probe concentrations were 5 ng μl−1. The probes used for examination of sediment samples are given in Table 2. For the probes specific to the β- and the γ-subclass of the Proteobacteria, unlabeled competitor probes have been used to improve the specificity of the hybridization. In situ hybridization was performed in a moisture chamber for 90 min. Subsequently, the slides were washed for 20 min in washing buffer of probe-dependent stringency. After removal of washing buffer and air drying, slides were stained with SYTOX-Blue (Molecular Probes Inc., Eugene, OR, USA) at a final concentration of 10 μM (for 30 min at 4°C in the dark) to determine the total cell number.

View this table:
2

Oligonucleotide probe data

ProbeSpecificitySequence (5′–3′)rRNA targetReference
EUB338bacteria5′-GCTGCCTCCCGTAGGAGT-3′16S[26]
ALF1bα-subclass of Proteobacteria5′-CGTTCGYTCTGAGCCAG-3′16S[27]
BET42aβ-subclass of Proteobacteria5′-GCCTTCCCACTTCGTTT-3′23S[27]
GAM42aγ-subclass of Proteobacteria5′-GCCTTCCCACATCGTTT-3′23S[27]
SRB385sulfate-reducing bacteria of the δ-subclass of Proteobacteria, and some other members of δ-subclass of Proteobacteria and some Gram-positives5′-CGGCGTCGCTGCGTCAGG-3′16S[26]
HGC69aGram-positive bacteria with high G+C content of DNA5′-TATAGTTACCACCGCCGT-3′23S[28]
CF319aCytophaga–Flavobacterium cluster of CFB5′-TGGTCCGTGTCTVAGTAC-3′16S[29]
PLA46Planctomycetales5′-GACTTGCATGCCTAATCC-3′16S[30]
ARCH915archaea5′-GTGCTCCCCCGCCAATTCCT-3′16S[31]

The slides were inspected by epifluorescence microscopy as described above. Filter sets for Cy3 were exciter filter HQ 535/50 and HQ 610/75 for emission, and for SYTOX-Blue were D 436/20, exciter filter and 480/40, emission filter. For each probe and sample, between 600 and 1000 SYTOX-Blue-stained cells and the respective amount of hybridized cells in 10–20 randomly selected microscopic ocular grid fields were counted.

2.4 Construction of 16S rDNA libraries and comparative sequence analysis

For analysis of the 16S rDNA sequences, total community DNA was extracted from pooled sediment samples (0–5 cm depth) using the FastPrep® instrument and the FastDNA® spin kit for soil (Qbiogene) according to the manufacturer's instructions.

Reaction mixtures were purified by centrifugation through Microcon YM-100 columns (Amicon). The 16S rRNA genes were amplified from total DNA using universal eubacterial primers corresponding to Escherichia coli positions 8–27 (27F: 5′-AGA GTT TGA TC[A/C] TGG CTC AG-3′) [32] and 1404–1387 (1387R: 5′-GGG CGG [A/T]GT GTA CAA GGC-3′) [33]. Moreover, a second broad range primer set for 16S rDNA PCR 27F and 1525R (5′-AAG GAG GTG TTC CA[G/T] CC[A/G] CC-3′) [34] was tested to investigate the universal amplification and the reproducibility of the results. PCR mixtures contained 5 μl of PCR buffer including 15 mM MgCl2 (Applied Biosystems), 200 μM of each deoxynucleoside triphosphate, 10 pM primer, 1.25 U of Amplitaq Gold DNA polymerase (Applied Biosystems), and 1–50 ng of template DNA in a final volume of 50 μl. The PCR was performed using a GeneAmp PCR system 9600 thermal cycler (Perkin-Elmer). PCR cycle conditions were as follows: 10 min of initial denaturation at 95°C; followed by 35 cycles of 90 s denaturation at 95°C, 90 s annealing at 50°C, and 2 min extension at 72°C; plus an additional final extension step of 10 min at 72°C before cooling to 4°C. Amplified DNA was verified by electrophoresis of 9 μl aliquots in 1.5% agarose gels. The PCR products were purified using QIAquick PCR purification kit columns (Qiagen GmbH, Hilden, Germany) and stored at −20°C.

16S rDNA clone libraries were generated for each reservoir sediment from PCR products from two sampling dates in 2000, respectively. Amplified 16S rDNA was cloned into E. coli using the TOPO TA cloning kit (Top 10 F′ competent cells, Invitrogen Corporation, San Diego, CA, USA) following the manufacturers’ recommendations. Recombinant clones were detected by blue–white colony selection on Luria–Bertani (LB) plates containing 100 μg ml−1 ampicillin, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) and isopropyl-β-d-thiogalactopyranoside (IPTG). White colonies were picked, and the insert size was determined by performing direct PCR on material from colonies using M13 forward and reverse primers. Clones containing an insert with the expected size were analyzed by sequencing using the primer 27F.

The analysis was performed with about 300 recombinant clones from each reservoir clone library using an ABI Prism™ 377 sequencer and BigDye terminator cycle sequencing with Amplitaq FS (PE Biosystems). Database searches of EMBL and GenBank databases were performed using FASTA of the HUSAR software package [32].

The retrieved 16S rDNA sequences were added to the database of the Department of Microbiology at the Technical University of Munich using the software package ARB (Department of Microbiology (http://www.mikro.biologie.tu-muenchen.de)). The resulting alignments were checked and corrected manually, considering the secondary structure of the rRNA molecule. The partial sequences (minimum sequence length 500 bp) were inserted into an existing tree by parsimony criteria, without allowing changes in the overall tree topology [35]. The results based on these analysis correlated with the FASTA output. Sequence data have been submitted to the GenBank database under accession numbers AJ518107 to AJ518809.

2.5 Enzyme assays

Extracellular enzyme activities were determined fluorometrically (RF 5001-PC spectrofluorometer, Shimadzu Europa GmbH, Duisburg, Germany) using the following artificial substrates (Sigma-Aldrich Chemie GmbH, Germany): fluorescein diacetate (FDA) (esterase), methylumbelliferyl (MUF)-phosphate (phosphatase), MUF-β-glucoside (β-glucosidase), and l-alanin-4-methoxy-β-napthylamide (aminopeptidase). The sliced sediment samples were diluted with 0.14 M NaCl and homogenized before substrate addition (four replicates per sample, volume of 3.0 ml each). All activities were substrate saturated at the concentrations used (final concentrations of 0.015 mM FDA, 0.31 mM MUF-phosphate, 0.81 mM MUF-β-glucoside, and 0.34 mM l-alanin-4-methoxy-β-napthylamide).

Incubations were performed in the dark with continuous shaking for 0.5–2 h at 20°C (esterase) and 30°C (other assays), respectively. Substrate blanks as well as controls of the samples were also incubated. After incubation, 0.7 ml of glycine buffer (0.1 M, pH 10) were added to the phosphatase and β-glucosidase assay, and 0.5 ml HEPES (0.1 M, pH 7.5) were added to the esterase and aminopeptidase assay, respectively. Two different incubation times for each assay were applied to check the time dependence of substrate conversion. Sediment solids were removed subsequently by centrifugation at 7000×g at 6°C for 10 min and fluorescence of the supernatant was measured at suitable excitation and emission wavelengths. Quantification was achieved by calibration with standard solutions of fluorescein, MUF and 4-methoxy-β-napthylamine, respectively.

2.6 Statistical analysis

The results of chemical and enzymatic analyses were tested by non-parametric statistics (Mann–Whitney U-test and Kruskal–Wallis analysis of variance (ANOVA)) using the software Statistica (StatSoft, Inc.). A matrix of correlation between enzyme activities and chemical data was established by using multiple linear regression statistics (Statistica). Differentiation between the clone libraries was analyzed on the basis of genetic diversity by means of the FST-test (according to Martin [36]) using the program Arlequin, v. 2.0 [37]. Richness of the obtained clone libraries was calculated by rarefaction using the RAREFACT.FOR program (C.J. Krebs, Department of Zoology, University of British Columbia (http://www2.biology.ualberta.ca/jbrzusto/rarefact.php)).

3 Results

3.1 Chemical gradients in the interstitial water

The chemical analysis of pore water provides information on important biogeochemical processes occurring in the sediment and on the environmental conditions for the microbial communities. Moreover, microbial processes may be examined indirectly by the vertical gradients of dissolved oxygen, redox potential and redox-sensitive chemical parameters. Accordingly, the gradients of the concentration of dissolved phosphate, electron acceptors (dissolved oxygen, NO3, SO42−) and indicators of microbial decomposition like NH4+ or alkalinity were obviously dependent from the trophic state of the reservoir. The maximum content of SRP in the pore water amounted up to 0.1 mg l−1 in the oligotrophic Neunzehnhain reservoir, whereas in the eutrophic reservoir Quitzdorf values up to 5.2 mg l−1 were detected. In the reservoirs Saidenbach and Quitzdorf, the SRP concentration of the interstitial water increased during summer stagnation.

Alkalinity and NH4+ concentration were also highest in the Quitzdorf reservoir reaching values of about 7 mmol l−1 and 18 mg l−1, respectively (Fig. 1). Both parameters showed a distinct increase with sediment depth, especially in the more eutrophic reservoirs Saidenbach and Quitzdorf. In these reservoirs, alkalinity as well as NH4+ concentration achieved a maximum during the summer stagnation period. In the reservoirs Neunzehnhain and Muldenberg, ammonium concentrations were below 1 and 2 mg l−1, respectively.

1

Vertical gradients of (A) ammonium concentration (mg l−1), (B) alkalinity (mmol l−1), (C) sulfate concentration (mg l−1), and (D) DOC concentration (mg l−1) in the interstitial water of the four reservoirs in 2000–2001 (dotted lines: full overturn, black lines: summer stagnation). Sediment depth −1 cm corresponds to the concentrations above the sediment–water interface.

The negative vertical gradients in nitrate concentration increased with the trophic level (data not shown). In all reservoirs under investigation, with exception of the oligotrophic Neunzehnhain reservoir, nitrate was already depleted at 1 cm sediment depth. The decrease of the sulfate concentration with sediment depth likewise reflected the trophic state of the reservoirs (Fig. 1). In the Neunzehnhain reservoir, no significant consumption of sulfate was detected. In the sediments of the other reservoirs under investigation microbial sulfate reduction resulted in an accumulation of sulfide. Sulfide concentrations were particularly high in the Quitzdorf and Saidenbach reservoirs during summer stagnation.

Similar to alkalinity and NH4+ concentration, the content of DOC was high in the interstitial water compared to the water column above the sediment. High DOC concentrations were measured in the eutrophic reservoir Quitzdorf, but also in the Muldenberg reservoir (Fig. 1). In the latter case, DOC was composed of allochthonous humic substances to a great extent. Especially in the Saidenbach reservoir, the DOC concentration showed an increase with sediment depth and maximum values during the summer stagnation period.

3.2 Quantitative microbiological analysis of sediment communities

The comparison of sediment microbial communities included the determination of total bacterial counts and the evaluation of the dominance structure of the microbial communities by means of FISH with rRNA directed oligonucleotide probes. For epifluorescence enumeration of sediment bacteria, the green fluorescent stain SYTOX-Green was preferred over the widely used 4,6-diamino-2-phenylindole (DAPI) staining due to a considerable reduction of background fluorescence.

The total number of bacteria in sediment samples from the four reservoirs mostly varied in the range from 1 to 6×109 cells per ml, indicating that population density was enhanced by a factor of 103 to 104 compared to the water body (Fig. 2). No significant differences of the cell counts per ml were found between the reservoirs of different trophic state. There were also no significant influences of the sediment depth or the sampling date on the bacterial counts. However, the highest numbers of bacteria were detected with only a few exceptions in May or June. With regard to the dry matter, highest cell counts were mostly detected in the upper sediment layer (0–1 cm). The total number of bacteria per g dry weight varied between 1 and 10×1010 in all sediments, except for the Neunzehnhain reservoir. In this oligotrophic reservoir, bacterial counts were in the range from 0.5 to 5×1010 per g dry matter.

2

Vertical distribution of total number of bacteria (direct count with SYTOX-Green) in sediment samples of the four reservoirs in 2000–2001.

In samples from three reservoirs, a relatively high percentage of total bacterial cells hybridized with the EUB338 probe, whereas the FISH technique failed in case of the Neunzehnhain reservoir (tested for samples from five sampling dates). As shown in Fig. 3, 40–80% of the total number of bacteria in sediments from the meso- and eutrophic reservoirs Saidenbach and Quitzdorf were detected by FISH using the EUB338 probe. In contrast, the oligotrophic reservoir Muldenberg harbored a lower percentage of cells detectable with this universal eubacterial probe. In all sediments the percentage of hybridized cells tended to decrease with sediment depth. The EUB II and EUB III probes recommended by Daims et al. [38] were also applied to samples from all investigated sediments but the counts were not above the detection limit, set at 1% of the total cell counts.

3

Vertical distribution of the fraction of cells hybridizing with probe EUB338 (as percentage of total cell counts determined by SYTOX-Blue staining) in sediment samples from the reservoirs Muldenberg, Saidenbach and Quitzdorf.

By means of the ARCH915 probe, the incidence of Archaea was demonstrated in sediment samples of the three reservoirs. However, only a percentage of less than 4% of total SYTOX-Blue cell counts in sediments from Saidenbach and Quitzdorf was affiliated to the ARCH915 probe. For the Muldenberg reservoir, the abundance of Archaea reached 5–7% of total number of bacteria at maximum only at one sampling date in samples from more than 1 cm sediment depth.

By means of group specific probes less than 30% of the total cells were identified as members of the Proteobacteria (Fig. 4) whereas members of other major groups of bacteria like the high GC group of Gram-positive bacteria, the Cytophaga–Flavobacterium cluster or the Planctomycetales were not detected by FISH. The sum of cells hybridized with the probes for the subgroups of Proteobacteria achieved about 50% of the EUB-positive cell counts at best. In most cases, the highest percentages were obtained in the upper sediment layer (0–1 cm). Among the members of Proteobacteria, the β-subgroup was the most abundant in the sediments of the eutrophic reservoirs Saidenbach and Quitzdorf (Fig. 4). However, the percentage of cells hybridized with the BET42a probe strongly decreased with sediment depth in both reservoirs. In contrast, members of the γ-subgroup were most frequently detected in sediment samples from the Muldenberg reservoir. A relative constant but low percentage of cells affiliated to the δ-subgroup, whereas α-Proteobacteria were rarely observed in sediment samples of the three reservoirs assessed by FISH.

4

Composition of Proteobacteria communities in sediment samples from three reservoirs at different sediment depths. The values obtained with the group specific probes ALF1b, BET42a, GAM42a, and SRB385 were standardized as percentage of total cell counts determined by SYTOX-Blue staining. In case of the Muldenberg and Saidenbach reservoirs, the values from a representative sampling date are shown, values for the Quitzdorf reservoir represent median values of four sampling dates from 2001.

The quantitative examination of microbial sediment communities did not show a significant influence of the trophic conditions in the reservoir on the bacterial density in the sediment whereas different proportions of the total cell counts in the different sediments hybridized with oligonucleotide probes specific for major groups of bacteria. The dominance structure within the Proteobacteria obviously differs between the sediment communities from the different reservoirs. However, a relatively high percentage of unaffiliated cells suggests the importance of bacterial groups not covered by the used probes.

3.3 Comparative 16S rDNA analysis

As the reservoirs were characterized by different pore water profiles as well as by differences in the quantity and quality of depositing organic material, the sediments of the four reservoirs under investigation should be inhabited by different microbial communities. These differences should be visualized by the presence or absence of different bacterial species or phylogenetic groups indicated by comparative 16S rDNA analysis.

For each reservoir sediment, clone libraries were established by amplification and cloning of 16S rDNA genes from extracted total DNA from two sampling dates. Two broad range primer sets 27F/1387R and 27F/1525R were used in parallel to verify the reproducibility of the methods and to investigate the potential influence of different primer sets on 16S rRNA gene amplification. The analysis of the resulting two clone libraries revealed highly similar results. For example, 79 of the 111 phylotypes obtained with the primer set 27F/1387R were also found in the library resulting from amplification of total DNA from Quitzdorf with the primer set 27F/1525R (total of 106 phylotypes). The similarity between two clone libraries was calculated as follows: 2×shared operational taxonomic units (OTUs)/total number of OTUs in both libraries (according to [39]). The similarity between both libraries obtained with the two primer sets was 73% whereas the libraries obtained from Quitzdorf and Saidenbach, using the primer set 27F/1525R were comparable divergent (similarity value of 17%). These results indicate that the 16S rDNA analysis led to reproducible results and was not subjected to random bias. Therefore, we used the more effective primer set 27F/1387R in the ongoing study.

A total of 2541 recombinant clones were analyzed by sequencing the plasmid inserts. The sequences were assigned to individual OTUs based on their phylogenetic positions and the 97% sequence similarity criterion. 1515 sequences, that is about 60% of the total clones, had a similarity of more than 97% to 16S rRNA sequences of cultured bacteria or clone sequences determined in other diversity surveys. In this paper, the results of the analysis of these sequences are discussed. About 60% of these clones fell into 15 known divisions (including the five subdivisions of Proteobacteria), all other clones showed a high homology to hitherto unidentified bacteria which are unclassified and not affiliated to certain phylogenetic clusters (Fig. 5). Thus, between 19.8% (Muldenberg reservoir) and 54.6% (Saidenbach reservoir) of the analyzed clones remained unaffiliated.

5

Phylogenetic diversity on the level of major groups of bacteria in 16S rDNA clone libraries from sediments of the four reservoirs. The bars represent the percentage of clones in each library affiliated to one of the major groups of bacteria due to analysis of 16S rDNA sequences with a minimum of 97% similarity to known 16S rRNA sequences (CFB cluster). For the Proteobacteria, the percentages of clones affiliated to the α-, β-, γ-, and δ-subgroups are shown. The percentage of sequences similar to hitherto unclassified bacteria was 46.3% for Neunzehnhain, 19.8% for Muldenberg, 54.6% for Saidenbach, and 29.7% for Quitzdorf.

Altogether, the comparative 16S rDNA analysis revealed 528 different phylotypes which were grouped into 293 OTUs. About 59% of these OTUs were found only in one library. The number of unique OTUs was similar in the four libraries, ranging from 31.4% of the total number of OTUs (Neunzehnhain) to 36.8% (in the Quitzdorf library) (Table 3). Only 32 OTUs (about 11% of all OTUs) were found in libraries from all reservoirs under investigation. The similarity values obtained from pairwise comparisons of the OTUs detected in the different libraries (according to [39]) did not considerably differ between the different libraries. Two reservoirs shared between 54 and 64 OTUs in each case resulting in similarity values ranging between 43.6 and 48.4%.

View this table:
3

Number of phylotypes and diversity indices obtained for rDNA clone libraries from four different reservoir sediments

ParameterClone libraries
NeuMulSaiQui
Total no. of phylotypes457248424386
Total no. of OTUs137119124133
No. of unique OTUs43384349
No. of OTUs recorded only once67626768
Coverage (%)85.375.484.782.6
OTU richnessa98.8 (4.1)119.0 (n.a.)90.7 (3.9)105.6 (3.7)
Chao-1 richness estimatora220.1 (21.4)192.9 (19.8)210.3 (22.2)239.9 (27.3)
Shannon index (H′)4.0984.5393.6254.444
Shannon evenness index (E)0.8330.9500.7520.909
  • aOTU richness was calculated by rarefaction [40] using RAREFACT.FOR program (C.J. Krebs, Department of Zoology, University of British Columbia (http://www2.biology.ualberta.ca/jbrzusto/rarefact.php) for a standardized sample size of 248 clones. The same program was used to calculate Chao-1 estimator of total number of OTUs in the samples [41]. The values in parentheses are standard deviations.

Considering sequences without a homolog of ≥97% similarity (i.e. OTUs with only one sequence) to be unique, the homologous coverage within each library was estimated (according to McCaig et al. [42] and Juretschko et al. [43]). By this means it was shown that the retrieved 16S rRNA gene clones cover between 75% (Muldenberg) and 85% (Neunzehnhain) of the expected OTU diversity. Thus, the minimum number of bacterial species in the sediment communities (inferred from the coverage values and the number of different OTUs detected) may range between 146 (Saidenbach) and 161 (Neunzehnhain and Quitzdorf). However, a higher OTU richness in all sediments is indicated by means of the Chao-1 estimator (Table 3). The estimated OTU richness for a standardized sample size of 248 clones (calculated by rarefaction, see Table 3) indicates the clone library from Saidenbach to inhabit the lowest number of OTUs whereas the Muldenberg sediment obviously had the highest level of OTU richness.

The number of phylotypes was quite different within the main phylogenetic groups. For all four reservoirs, the highest aggregate numbers of sequences were found in the division of Proteobacteria (Fig. 5). Members of this division accounted for 25.4% of the clones from Neunzehnhain, 46.0% of the clones from Muldenberg, 18.7% of the clones from Saidenbach, and 28.0% of the clones from Quitzdorf. The clones were relatively even distributed among the α-, β-, γ-, and δ-subgroups, respectively. The most Proteobacteria sequences from the clone library of Saidenbach affiliated to the δ-subgroup, whereas sequences belonging to the α-Proteobacteria were most frequently detected in the other sediments. A total of 130 clones grouped into 31 OTUs affiliated to the latter group. Only three OTUs were found in all sediments whereas 18 OTUs were restricted to only one library. To assess how divergent the four libraries were with regard to the α-subgroup of Proteobacteria, the differentiation between the libraries was analyzed on the basis of genetic diversity by means of the FST-test (according to [36]). Significant genetic differentiation (P<0.05) became evident between all the sediment communities except Neunzehnhain and Saidenbach. These two libraries did not significantly differ with regard to the (α-Proteobacteria (P=0.207).

High numbers of sequences affiliated to the Firmicutes were found in the clone libraries from the Neunzehnhain and Quitzdorf reservoir, whereas the Cytophaga–Flexibacter–Bacteroides (CFB) cluster comprised a relatively high percentage of clones from the Muldenberg and Saidenbach library, respectively. The ‘unidentified bacteria’ do not form a separate phylogenetic group since this group represents sequences in databases which are not affiliated to one of the main phylogenetic lineages. However, a fraction of these organisms may constitute an independent phylogenetic group. This group comprised 136 different phylotypes (25.8% of the total number of phylotypes), with 108 sequences detected only in one library.

As expected, Cyanobacteria species represented an important fraction of the microbial communities in the eutrophic reservoirs Saidenbach and Quitzdorf. According to the mass growth of Microcystis during the summer months in Quitzdorf, high abundances of Microcystis colonies were also detected in the sediment [44] and 18 sequences clustered into three OTUs belonging to this genus were found in the 16S rDNA library from Quitzdorf (Figs. 5 and 6). The comparison of two different sampling dates (spring and summer) showed that the observed late-summer blooms of cyanobacteria in the Quitzdorf reservoir resulted in an increased number of Microcystis sequences detected in the 16S rDNA library (data not shown). Only the clone library from the Saidenbach reservoir contained sequences affiliated to Synechococcus species (Fig. 6).

6

Phylogenetic relationships of cyanobacterial rRNA clones from sediments from the Saidenbach reservoir (Sai) and from the Quitzdorf reservoir (Qui). Sequences with more than 97% similarity were assigned to individual OTUs. The abundance of these OTUs in a library is given in parentheses. The scale bar represents 10% estimated sequence divergence.

Further noteworthy differences between the 16S rDNA libraries on the level of main phylogenetic groups regarded the incidence of the green sulfur bacteria and of members of the CFB cluster. The incidence of anoxygenic phototrophic green sulfur bacteria was restricted to the sediments from Quitzdorf and Neunzehnhain. After the Proteobacteria and Firmicutes, the CFB cluster comprised the highest number of sequences (81 clones) affiliated to one main phylogenetic group. In contrast to the other sediments, the clone library from the Neunzehnhain reservoir harbored only few sequences affiliated to the CFB cluster. As shown in Fig. 7, two OTUs (OTU 10, similar to the uncultured bacterium PHOS-HE62 [45] and OTU 13, similar to the uncultured eubacterium WCHB1-53 [46]) were found in each library, the OTU 2 and OTU 3 (similar to sequences isolated from a benzene mineralizing consortium [47] and a deep-sea sediment [48], respectively) were missing only in the clone library from Neunzehnhain. These four OTUs comprised about 70% of all clones affiliated to the CFB cluster. The similarity between the clone libraries from Muldenberg, Saidenbach and Quitzdorf with regard to the CFB group is supported by the FST-test within this group indicating no significant differences between the three libraries.

7

Phylogenetic tree showing the relationship of OTUs from four reservoir sediments (Neunzehnhain, Neu; Muldenberg, Mul; Saidenbach, Sai; Quitzdorf, Qui) to selected reference sequences of the CFB cluster. The values in parentheses represent the frequency of clones in the library. The scale bar represents 10% estimated sequence divergence.

Clones affiliated to the Verrucomicrobia were relatively even distributed among the four libraries. More than 50% of the sequences belonged to one OTU abundant in all the four clone libraries. This ubiquitous clone was similar to the uncultured verrucomicrobium DEV020 (AJ401128) extracted from Elbe river biofilms by Felske and co-workers (unpublished).

The 16S rDNA clone libraries established for the four reservoir sediments differed only slightly with regard to their OTU richness as well as to the distribution of the revealed phylotypes among the main phylogenetic groups of bacteria. Notwithstanding the similarities on the level of broad taxonomic units, the similarity values obtained from pairwise comparisons of the OTUs of about 50% indicate that each sediment harbors a characteristic microbial community. This conclusion is also supported by the fact that one third of the OTUs from each library was found only in this library (unique OTUs).

3.4 Enzyme activities

Extracellular enzymes play an important role in the decomposition of organic matter in sediments and their activities in the sediment are considered to be dependent on the input of organic matter by sedimentation [5]. Therefore, the vertical distribution of esterase, phosphatase, β-glucosidase and aminopeptidase activities was studied for sediments from the four reservoirs sampled at different dates.

The measured rates varied exceedingly at different sampling dates, but did not depend on the sediment depth (Kruskal–Wallis ANOVA, P>0.05). For example, rates of enzymatic hydrolysis of FDA between 0.2 and 1.4 μmol h−1 ml−1 were measured, with noteworthy temporal variation especially in the oligotrophic reservoirs. A significant influence of sediment depth was only detected on the phosphatase activity (per ml) in the Saidenbach reservoir and on the aminopeptidase activity (related to dry matter) in the sediment of the Muldenberg reservoir (P=0.0003 and P=0.0201, respectively). Contrarily, the significant influence of the sampling date on the four enzyme activities became evident for all reservoirs by Kruskal–Wallis ANOVA (P<0.01), except for the phosphatase activity per ml in the Saidenbach sediment (P=0.103).

Remarkable differences between the four reservoirs appeared by the pairwise comparison of all measured extracellular enzyme activities, related to dry matter, regardless of sediment depth or sampling date, respectively (Fig. 8). Significant differences between all reservoirs under investigation were detected with regard to the phosphatase activity. Thus, the trophic state of the reservoirs had the expected effect on phosphatase activity. Highest rates of MUF-phosphate conversion were found in the sediments of the oligotrophic Muldenberg reservoir, whereas the more eutrophic reservoirs Saidenbach and Quitzdorf were characterized by low phosphatase activities (Fig. 8). On a dry matter basis, the meso- and eutrophic reservoirs Saidenbach and Quitzdorf differed significantly with regard to only this extracellular activity. The oligotrophic reservoirs Neunzehnhain and Muldenberg showed similar β-glucosidase activities (P=0.563 after Mann–Whitney U-test) whereas the aminopeptidase activities differed with P=0.023 (Mann–Whitney U-test). No significant similarities were found between these reservoirs and the more eutrophic reservoirs with regard to the levels of extracellular enzyme activities.

8

Comparison between the four reservoirs (Neunzehnhain, Neu; Muldenberg, Mul; Saidenbach, Sai; Quitzdorf, Qui) with regard to the activities of phosphatase and l-alanin-aminopeptidase determined as conversion rates (per g dry matter) of MUF-phosphate and l-alanin-4-methoxy-β-napthylamide (l-Ala-MNA), respectively. The upper box value is the 75th percentile, the lower box value is the 25th percentile, values higher than 1.5 of box height referred to outliers, extreme values are those which are outside the 3 box length range from the upper and lower value of the box. Differences between the phosphatase activities in the different reservoirs were significant (P<0.001, except for Neu and Sai: P=0.026, Mann–Whitney U-test). P levels for differences with regard to the aminopeptidase activity were <0.001 with exception of Neu vs. Mul (P=0.023) and Sai vs. Qui (P=0.352).

The enzymatic hydrolysis of FDA was likewise highest (on a volume basis) in the oligotrophic reservoirs Neunzehnhain and Muldenberg, indicating a high hydrolytic potential in these sediments even though the sedimentation of particulate material was low. On the other hand, activities related to the decomposition of proteins (aminopeptidase) or cellulose (β-glucosidase) were relatively more important in the sediments of the eutrophic reservoirs (Fig. 8). Thus it appears that the sediments of reservoirs of different trophic state feature specific patterns of extracellular enzyme activities.

If the activities of sediment samples of all reservoirs were pooled, the phosphatase activity correlated negatively with the aminopeptidase and the β-glucosidase activity (Pearson correlation coefficients −0.28 and −0.23, respectively, P=0.0005 for aminopeptidase, P=0.0015 for β-glucosidase). Between phosphatase and esterase activity (on a dry matter basis), a significant positive correlation was found (r=0.43, P<0.00001, n=182). Moreover, to detect factors influencing the enzyme activities, a correlation analysis of enzyme activities and chemical parameters of the interstitial water as well as bacterial density and organic matter was carried out. The esterase activity correlated positively with organic matter content, but negatively with the total number of bacteria in the sediments (both significant with P<0.0001). The enzymatic conversion of MUF-β-glucoside was also clearly related to both the bacterial density and the organic matter content (Pearson correlation coefficients 0.215 (P=0.012, n=135) and 0.226 (P=0.006, n=145), respectively), whereas the aminopeptidase activity showed a positive correlation to the concentration of ammonium and SRP in the interstitial water, to the alkalinity and to the SRP:DOC ratio (P levels<0.01, with exception of SRP (P=0.035)). As expected, the activity of phosphatase correlated negatively with SRP (P=0.005) as well as with the SRP:DOC ratio (P=0.005), but also with indicators of microbial decomposition like ammonium concentration or alkalinity (each with P<0.001).

4 Discussion

The investigated reservoirs differed with regard to parameters of the water body like chlorophyll content or total phosphorus as well as to chemical parameters of the interstitial water of the sediment. Both compartments are connected by intensive interactions. The level and the gradients of alkalinity and ammonium concentration but also the depletion of electron acceptors (oxygen, nitrate and sulfate) in the pore water (Fig. 1, see also [44]) show that the bacterial activity in the sediments was closely linked to the trophic conditions in the water body, i.e. the resulting sedimentation of particulate organic matter. For example, high ammonium concentrations in the interstitial water of the reservoirs Saidenbach and Quitzdorf indicate high rates of microbial decomposition of organic nitrogen compounds in these sediments. Therefore, the microbial in situ activities obviously increased with trophic state although the bacterial densities in the sediments from the different reservoirs did not significantly differ.

Since a high proportion of the carbon loss from POC decomposition in the sediments remains unassimilated, the DOC content of the interstitial water should be considered as a result of microbial activity [49]. With exception of the dystrophic reservoir Muldenberg, our findings suggest that a higher input of detrital organic material in the more eutrophic reservoirs enhanced the microbial activity in the upper sediment layers and therefore resulted in higher DOC concentrations in the pore water. This became also evident from the comparatively high β-glucosidase activities in the sediments from the Saidenbach and Quitzdorf reservoir, respectively. However, as unassimilated DOC can also be exported into the water column, the DOC standing stock represents a highly dynamic pool of organic substrates and refractory material [49].

In contrast to the other reservoirs, the interstitial water of the sediment from Muldenberg consisted mainly of allochthonous humic substances. Therefore, the DOC content of the pore water did not result from high microbial decomposition activities. According to Sinsabaugh and Findlay [49], a low carbon quality (i.e. accumulation of humic substances and a lack of polysaccharides) is indicated by the low activities of β-glucosidase and aminopeptidase in the Muldenberg sediment.

Whereas in laboratory experiments a concurrent increase of bacterial biomass in response to the addition of seston suggests a strong coupling between changes in the water column and abundance and activity of microorganisms in the sediment [4], the bacterial abundance in the different sediments does not seem to be influenced by the different conditions in the water body and the different C-input by detrital material. However, it has to be expected that the species composition and microbial activities in the sediment reflect differences in the trophic conditions as well as characteristics of the catchment, respectively.

The high number of different phylotypes found in the clone libraries obtained after DNA extraction from sediment samples of the four different reservoirs is consistent with the results of several studies on soil or sediment samples [9,12,39,42,48]. The calculated OTU richness (Table 3) is similar to the number of OTUs recorded in clone libraries (N=236) from a zinc-contaminated and a control soil [40]. Moreover, Shannon indices (H′) estimated for the sediments were also in the same range as calculated in the study mentioned above. The diversity as indicated by the Shannon index was lowest in the mesotrophic Saidenbach reservoir. Although diversity may be underestimated due to incomplete coverage [40], from the comparison of H′ it can be inferred that the sediment communities in Muldenberg and Quitzdorf are more diverse. However, it has to be taken into consideration that this general diversity index is influenced by both richness and evenness [40]. Since neither the estimated OTU richness nor the diversity indices tended to increase or decrease with trophic state, the diversity of sediment communities does not seem to depend upon the trophic state of the reservoir. However, the comparison of the diversity of clone libraries is limited by the uncertainty of definition of bacterial species [50]. Thus, the species diversity might be higher, since sequences with a similarity of the sequenced gene fragment of ≥97% can belong to different bacterial species. Moreover, the frequency of sequences amplified by PCR does probably not reflect the abundance of the corresponding bacterial species in a community because of PCR drift and selection [5153]. The comparison of clone libraries by means of estimated richness or diversity indices is further limited by the fact that bacterial species (or OTUs) are counted equivalently regardless their phylogenetic relationships [36].

In contrast to clone libraries obtained from soils which were dominated by clones affiliated to the α-Proteobacteria[42] or by Acidobacterium sequences [39], respectively, Firmicutes and the α-, β-, γ-, and δ-subgroups of Proteobacteria were the bacterial groups comprising the most phylotypes in the libraries of the four reservoirs. This is in accordance with 16S rRNA analysis of lake bacterioplankton [35,54] which revealed sequences affiliated with the Proteobacteria (α-, β- and γ-subgroups), with the CFB group, and with the high GC Gram-positives, respectively, as the most abundant. Although the communities of the different reservoirs only weakly differed on the level of main phylogenetic lineages (Fig. 5), it became evident by similarity indices between 44 and 48%, respectively, that each reservoir harbors a specific sediment microbial community. This is also illustrated by the high percentage of OTUs (about 59%) found in only one library. As mentioned above, the comparison of clone libraries on the basis of OTUs does not account for the information about the disparity among the sampled sequences. By means of the FST-test (according to Martin [36]: genetic diversity within each community compared to the total genetic diversity for all samples), a significant genetic differentiation between the sediment communities became evident with regard to the α-subgroup of Proteobacteria. Thus, the conclusion can be drawn that the sediment communities from the reservoirs under investigation mainly differ with regard to the representation of members of the main phylogenetic groups highly abundant in the clone libraries.

If the phylogenetic groups that account for differences between the communities can be identified, inferences about the environmental conditions driving the community structure become possible. However, the majority of OTUs revealed by 16S rDNA analysis of the four sediment communities represented uncultured bacteria with unknown physiological capabilities. Moreover, the bias associated with community DNA extraction, PCR amplification and cloning strategies results in clone libraries that are no quantitative indicators of microbial abundance and therefore may also result in a qualitatively misleading portrait of the microbial diversity [5153,55]. For these reasons, analysis of bacterial community diversity based on amplification of 16S rRNA genes should be combined with cultivation techniques (for further physiological characterization of members of the community), and an assessment of the dominance structure by techniques like FISH, respectively.

The FISH with rRNA-targeted oligonucleotide probes is the most promising tool for the quantitative analysis of community structure. However, high background fluorescence and unspecific binding to non-bacterial particles can hamper the enumeration of hybridized cells in sediments as well as a low fluorescence intensity of cells due to low ribosome content of inactive cells [18], respectively. Sediment samples from the Neunzehnhain reservoir were characterized by relatively high amounts of iron and aluminum, low organic matter and many non-bacterial particles which hamper the microscopic examination by bright unspecific signals and high background fluorescence and therefore cause the failure of FISH technique in these samples. In the sediments of the meso- and eutrophic reservoirs Saidenbach and Quitzdorf, respectively, the relatively high percentages of cells hybridized with EUB338 probe (Fig. 3) indicate that a considerable part of the microbial communities down to a depth of at least 50 mm was metabolically active (according to Karner and Fuhrmann [56]). This active subpopulation tended to decrease with sediment depth and seemed also to depend on the trophic state, as indicated by the lower percentage of EUB-positive cells in the sediment from Muldenberg. A similar influence of sediment depth on the percentage of cells detectable by FISH was obtained by Llobet-Brossa [20] for wadden sea sediments. These authors reported a large fraction (up to 73%) of DAPI-stained cells hybridized with the EUB probe, too.

Although we applied a set of seven probes for major phyla within the domain Bacteria, more than 50% of detectable cells, i.e. the EUB338 counts, remained unaffiliated. Only the probes specific for the subgroups of Proteobacteria gave counts above the detection limit, set at 1% of the total cell counts. Thus, the diversity of sediment communities was not sufficiently reflected by the results of FISH and the high percentage of unaffiliated cells suggests that other significant bacterial populations exist. It was unexpected that only few cells hybridized with the CF319a or the HGC69a probe, respectively, since the clone libraries contained numerous sequences affiliated to these main phylogenetic groups. The absence of positive hybridization signals with the HGC69a probe may be explained by an insufficient permeabilization of bacterial cell walls. Moreover, it has to be taken into consideration that the clone libraries do not reflect the real composition of microbial communities. But, the Cytophaga–Flavobacterium cluster is shown to be frequently underrepresented in 16S rDNA clone libraries, compared with FISH data [57].

The incidence of members of the Cytophaga–Flavobacterium cluster in clone libraries from three reservoir sediments corresponded to the relatively high abundance of clones related to the genus Cytophaga in clone libraries obtained from marine snow [8] as well as from particle-attached microbial communities in an estuary [58]. Using FISH, Llobet-Brossa et al. [20] found members of the Cytophaga–Flavobacterium cluster to be the most abundant bacterial group in wadden sea sediments. The unexpected failure of the CF319a probe suggests a less importance of this group in the sediment communities under investigation. However, the abundance of members of the CFB cluster has to be checked by new probes since they possibly do not contain a perfectly matching target site for the CF319a probe. The absence of cells detected with the CF319a probe may also be attributed to low levels of rRNA per cell and/or a lower sensitivity of the probe (compared to higher sensitivities of other probes, e.g. EUB338) [59].

An unknown portion of the detectable but unaffiliated cells may belong to other main groups of bacteria (e.g. the Fibrobacter/Acidobacteria group, the Verrucomicrobia and the green non-sulfur bacteria) for which hitherto no group specific probes were available or these probes have not yet been applied. Members of these phyla may represent an important part of the sediment microflora since their presence was shown in the clone libraries from the four sediments under investigation. Within all the three main groups, two or three OTUs, respectively, were found in all clone libraries indicating the presence of widely distributed bacterial species or groups within these phyla. Although these main groups have been frequently detected in various environments like soil, sediment, freshwater aggregates and activated sludge [12,39,43,60], only limited physiological information is available about these organisms and their potential environmental significance remains almost unpredictable [61].

Notwithstanding the incompleteness of the probe set used, differences between the reservoirs Muldenberg, Saidenbach and Quitzdorf were detected by means of FISH probes with regard to the dominance structure within the Proteobacteria phylum. The members of the Proteobacteria represent a high proportion of the sequences revealed by 16S rDNA analysis as well as a considerable part of the microbial communities indicated by FISH. Up to 30% of the active bacteria (EUB+ cells) in the upper sediment layers of the reservoirs Saidenbach and Quitzdorf belonged to the β-Proteobacteria, indicating that the higher input of detrital organic material possibly favors the growth and activity of members of this subgroup. The clear decrease of the number of (active) β-Proteobacteria with sediment depth also suggests an influence of the availability of ‘high-quality’ organic carbon and/or of electron acceptors like oxygen or nitrate. Our results are in accordance with the dominance of β-Proteobacteria in microbial assemblages of lake snow aggregates reported by Weiss et al. [62].

Conversely, the relatively high abundances of γ-Proteobacteria in sediments from the dystrophic Muldenberg reservoir (up to 24% of the EUB+ cells) is possibly linked to the high content of humic substances. In a recent study, humic substances were shown to be oxidized by different microorganisms also under anaerobic conditions [63]. In order to prove the importance of γ-Proteobacteria and their function in the Muldenberg sediment, a combined approach using oligonucleotide probes specific to selected clones of the 16S rDNA library and cultivation techniques to isolate dominant members of the microbial community and evaluate their metabolic potential seems to be favorable.

Another characteristic of the Muldenberg sediment was the high phosphatase activity. The (alkaline) phosphatase activity is considered as an indicator of phosphorus deficiency [6466]. Consequently, higher phosphatase activities were observed in the oligotrophic reservoirs and the phosphatase activity was negatively correlated with the SRP content of pore water as well as with the ammonium concentration and alkalinity. In the dystrophic reservoir Muldenberg, the phosphatase activity in the sediment seems to be not only linked to a phosphorus deficiency but also to the relatively high concentration of humic substances (as shown by the good correlation with the SRP:DOC ratio). Our observation is in good agreement with the results of studies on microbial activities of the plankton in acidic humic lakes [67,68]. Large amounts of humic substances [67,68] as well as of dissolved aluminum [65] were discussed to diminish phosphorus availability and accordingly to cause high phosphatase activities. In contrast to the studies concerning planktonic activities, there is no evidence that algae and flagellates significantly contribute to the measured extracellular enzyme activities in the sediments. It is supposed that the specific pattern of enzyme activities in the Muldenberg sediment is linked to a specialized microbial community adapted to the local environmental conditions, especially the main carbon sources and the restricted availability of phosphorus. The use of the artificial substrate ELF™-PO4 which is converted into fluorescent precipitates by phosphatase activity in combination with phylogenetic probes (as proposed by Van Ommen Kloeke and Geesey [69]) should facilitate the identification of the microorganisms responsible for the high phosphatase activity in the Muldenberg sediment.

In conclusion, our results suggest that the trophic character and the differences between the reservoir catchments affect the composition but not (or to a less extent) the bacterial density of microbial communities in the sediments. However, the diversity of microbial sediment communities seems to be hardly influenced by the trophic state. Thus, the detection of ‘key species’ of a sediment type or bacterial clusters which are important for certain functions is the challenge for further investigations. The in situ detection of these microorganisms by means of specific oligonucleotide probes, possibly in combination with microautoradiography [70], seems very promising to identify and quantify these members of the microbial communities. Moreover, PCR-based fingerprint techniques like DGGE and T-RFLP are currently used to investigate temporal changes and spatial gradients of community structure and to relate them to changes of the chemical parameters. Data on community structure of sediments from other reservoirs should be included to elucidate differences between microbial communities attributed to the trophic state. The specific patterns of extracellular enzyme activities obtained for the four sediment microbial communities suggest that these parameters can be used as indicators for the metabolic response of sediment microorganisms to the organic matter supply from the water column. Thus, a combined sediment characterization including estimation of microbial diversity, chemical composition and enzyme activities of sediment samples may contribute to the evaluation of the quality status of a reservoir. Moreover, molecular techniques will be beneficial to a risk assessment of possible remobilization of pathogens from the sediment.

Acknowledgments

This work was supported by a grant from the BMBF (German Ministry for Education and Research, grant 02WT9997/2). We are grateful to the laboratories of the Saxonian Reservoir Administration for providing unpublished data. We gratefully acknowledge Elke Vieweg, Astrid Käppler, and Doreen Große for their technical assistance. We further thank Dietrich Uhlmann for helpful discussions and Kerstin Röske for critically reading the manuscript.

References

  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]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
View Abstract