Bacterial communities in sediments of the shallow Lake Dongping in China

Authors

  • H. Song,

    1.  Department of Microbiology, College of Life Science, Shandong Agricultural University, Taian, Shandong Province, China
    Search for more papers by this author
  • Z. Li,

    1.  Department of Microbiology, College of Life Science, Shandong Agricultural University, Taian, Shandong Province, China
    Search for more papers by this author
  • B. Du,

    1.  Department of Microbiology, College of Life Science, Shandong Agricultural University, Taian, Shandong Province, China
    Search for more papers by this author
  • G. Wang,

    1.  Shenzhen Engineering Laboratory for Algal Biofuel Technology Development and Application, School of Environment and Energy, Peking University Shenzhen Graduate School, Shenzhen, China
    2.  Department of Microbiology, University of Hawaii at Manoa, Honolulu, HI, USA
    Search for more papers by this author
  • Y. Ding

    1.  Department of Microbiology, College of Life Science, Shandong Agricultural University, Taian, Shandong Province, China
    Search for more papers by this author

Yanqin Ding, Department of Microbiology, College of Life Science, Shandong Agricultural University, Taian, Shandong Province 271018, China.
E-mail: dingyq6885@163.com and Guangyi Wang, Department of Microbiology, University of Hawaii at Manoa, Honolulu, HI, USA.
E-mail: gywang@pkusz.edu.cn or guangyi@hawaii.edu

Abstract

Aims:  The purpose of this study was to discuss how the environmental inputs and anthropogenic activities impact bacterial communities in the sediments of a shallow, eutrophic and temperate freshwater lake.

Methods and Results:  Sediment cores were collected from Lake Dongping, located in Taian, Shandong, China. All samples were processed within 4 h of collection. Total nitrogen, total phosphorus (TP), total organic carbon, ammonium nitrogen and nitrate nitrogen content of samples were measured by Kjeldahl determination, sulphuric acid–perchloric acid digestion and molybdenum blue colorimetry, potassium dichromate titration, Nessler’s reagent colorimetric and the phenol disulphonic acid colorimetric method, respectively. Seasonal and temporal diversity of sediment bacterial communities at six stations in Lake Dongping were investigated using molecular approaches (terminal restriction fragment length polymorphism and 16S rDNA clone libraries). Noticeable seasonal and temporal variations were observed in bacterial diversity and composition at all six stations. Sediment bacterial communities in Lake Dongping belonged to 16 phyla: Proteobacteria (including α-Proteobacteria, β-Proteobacteria, δ-Proteobacteria, ε-Proteobacteria, γ-Proteobacteria), Acidobacteria, Planctomycetes, Bacteroidetes, Firmicutes, Verrucomicrobia, Nitrospira, Chloroflexi, Gemmatimonadetes, Chlorobi, Cyanobacteria, Deferribacteres, Actinobacteria, OP8, Spirochaetes and OP11. Members of β-, δ- and γ-Proteobacterial sequences were predominant in 11 of 12 clone libraries derived from sediment samples. Sediment samples collected at stations 1 and 4 in July had the greatest bacterial diversity while those collected at station 2 in October had the least diversity. TP concentration was significantly correlated with the distribution of bacterial communities.

Conclusions:  Our results suggested that different environmental nutrient inputs contribute to seasonal and temporal variations of chemical features and bacterial communities in sediments of Lake Dongping. TP concentration was significantly correlated with the distribution of bacterial communities.

Significance and Impact of the Study:  This study has an important implication for the optimization of integrated ecosystem assessment of shallow temperate freshwater lake and provides interesting information for the subsequent of the ecosystem.

Introduction

Sediments are a major component of freshwater ecosystems and interact in complex ways with the water body (D’Angelo and Reddy 1994; Fleming et al. 2006) owing to the presence of diverse microbial communities with high biogeochemical activities. Sediment microbes play important roles in nutrient cycling in lake ecosystems (Schallenberg and Kalff 1993) through the transformation of complex organic compounds and minerals in freshwater sediments (Nealson 1997; Jurgens et al. 2000). Therefore, changes in the composition of microbial communities can significantly impact biogeochemical environments of sediments through changes in metabolic transformation of organic and inorganic elements (Wang et al. 2003). For example, sediment microbes have been suggested to influence phosphorus availability through the uptake and release of soluble reactive phosphorus (Gachter et al. 1988). On the other hand, many biogeochemical factors, such as the relative abundance of organic matter in sediments, can influence diversity and the composition of sediment microbial communities (Coolen et al. 2002). Previous studies on factors affecting microbial communities have largely focused on sediments from special environments (Jiang et al. 2006; Li et al. 2008; Percent et al. 2008); freshwater sediments have been poorly studied relatively.

Molecular methods based on 16S rRNA genes have been proved useful in studies of sediment microbial ecology. Common sediment bacterial sequences come from similar phyla to those typically found in bacterioplankton, including Proteobacteria, Acidobacteria, Planctomycetes, Bacteroidetes, Firmicutes and Verrucomicrobia (Spring et al. 2000). The exceptions are cyanobacterial sequences, which have rarely been reported from sediments although they are frequently recovered from water columns (Lindstrom et al. 2005; Nelson et al. 2007; Röske et al. 2008; Tang et al. 2009). In general, bacterioplankton communities consist predominantly of β-Proteobacteria, while sedimentary bacterial communities are dominated by δ-Proteobacteria in lake ecosystems (Spring et al. 2000).

Lake Dongping (China) is a typical shallow, eutrophic, temperate freshwater lake that receives large amounts of water and nutrients from the Dawen River. Its ecosystem has seriously been impacted by appreciable amounts of nutrients and pollutant inputs from shoreline industries and other anthropogenic activities (e.g. aquaculture) (Fig. 1). These special physical settings and the geological location of Lake Dongping make it an ideal freshwater ecosystem for the study of effects of different environmental inputs on the seasonal variation of bacterial communities in freshwater lake sediments.

Figure 1.

 Schematic representation of Lake Dongping. Note: the number 1, 2, 3, 4, 5, and 6 refer to station 1, station 2, station 3, station 4, station 5, and station 6, respectively; BF, AZ and SOSF refer to the Baliwan floodgate, the aquaculture zone and the sewage outlets of the starch factory, respectively.

In the present study, we used 16S rDNA library construction and terminal restriction fragment length polymorphism (T-RFLP) to investigate the seasonal and temporal diversity of bacterial communities in sediments of a shallow freshwater lake in the north-east of China in comparison with biogeochemical characteristics. This is one of the few report showing how environmental inputs and anthropogenic activities impact bacterial communities in the sediments of shallow freshwater lake in China. The aim of the presented study is to determine the influence of organic matter and nutrients on distribution and community structure of bacteria in Lake Dongping sediment.

Materials and methods

Study sites and sample collection

Lake Dongping (35°30′–36°20′N and 116°00′–116°30′E) located in Dongping County, Shandong Province (China), has a relatively large surface area (627 km2), but is shallow (mean depth about 1–2 m) (Jiang et al. 2002). It is surrounded by about 151 km2 of wetlands in the vicinity of the entrance of the main water source (Fig. 1). The lake receives water inputs from the Dawen River at the south-eastern corner and discharges its overflow into the Yellow River at the northern side (Fig. 1). The Dawen River is a seasonal river with a high water level in the wet season (July) and a low level in the dry season (October) (Jiang et al. 2002). Overall, large volumes of nutrients, sewage effluents and pollutants from the Dawen Basin are carried into Lake Dongping through its water inflows.

To study sediment microbial communities, sediment cores were collected from six sampling stations (Fig. 1). The station locations were chosen based on the lake’s characteristics of the surrounding environment. Samples were collected from the individual stations in July and October 2008. Station 1 (36°02′43·4′′N, 116°13′07·3′′E) is close to the sewage outlets of a starch factory, and Lake Dongping receives large amount of saccharides, starch and proteins from the starch factory. Station 2 (36°06′31·8′′N, 116°12′28·6′′E) is close to the Baliwan floodgate, which controls lake freshwater level with a dike system that discharges overflows into the lower course of the Yellow River. Station 3 (36°02′02·0′′N, 116°11′33·1′′E) is located in an aquaculture zone. Station 4 (35°59′12·1′′N, 116°11′59·6′′E) is at the centre of lake. Station 5 (35°56′00·9′′N, 116°14′14·5′′E) is near the mouth of the Dawen River. Station 6 (35°58′38·9′′N, 116°15′01·0′′E) is near the wetlands of Lake Dongping.

Triplicate sediment cores (0–15 cm depth) were collected randomly within a 2-m2 area at each station using a K-B Corer Sampler (Wildlife Supply Company, NY, USA). Individual sediment cores were kept on ice and transported to the laboratory for analysis. All samples were processed within 4 h of collection. Three sediment cores from the same station were mixed before analysis.

Organic matter and nutrients analysis

Biogeochemical analyses were performed in the Soil Biochemical Analysis Laboratory at Shandong Agricultural University. Sediment samples were air-dried, then sieved through a 60-mesh stainless screen. The sieved sediment samples (0·5–1·0 g) were used for total nitrogen (TN) using Kjeldahl method. Fresh sediment samples (10·0 and 50·0 g, respectively) were used for ammonium nitrogen (NH4-N) and nitrate nitrogen (NO3-N) with Nessler’s reagent and the phenol disulphonic acid colorimetric methods, respectively (Hart et al. 1994). For total phosphorus (TP) analysis, air-dried sediment samples were sieved through 100-mesh stainless screen and the sieved sediments (0·5–1·0 g) were used for TP determination using sulphuric acid–perchloric acid digestion and molybdenum blue colorimetry (Olsen and Sommers 1982). Air-dried sediment samples (0·1–1 g) sieved through 60-mesh stainless screen were used for total organic carbon (TOC) determination using potassium dichromate titration (Nelson and Sommers 1996).

Total DNA extraction, PCR amplification and cloning library construction

Total genomic DNA was extracted using a soil DNA Kit (Omega, Norcross, GA, USA). The 16S ribosomal RNA gene fragments used for the clone library construction were amplified using the forward 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and reverse 1492R (5′-GGTTACCTTGTTACGACTT-3′) primers (Lane 1991; Dojka et al. 1998). PCR amplification was performed in 50 μl reaction mixtures containing 1 × PCR buffer, 1·5 mmol l−1 MgCl2, 200 μmol l−1 dNTPs, 0·5 μmol l−1 of each primer, 2·5 U of Taq polymerase (TaKaRa, Dalian, China) and about 20 ng of DNA template. The PCR reaction was performed using the following thermal cycles: initial denaturation at 95°C for 5 min, 30 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 90 s and final extension at 72°C for 10 min. PCR products were purified using the DNA Fragment Purification Kit Ver.2.0 (TaKaRa) and were cloned into a pMD19-T vector (TaKaRa) following the manufacturer’s instructions. The resulting ligation products were transformed into Escherichia coli DH5α competent cells.

Amplified ribosomal DNA restriction analysis (ARDRA) was used to group the random clones (Vergin et al. 2001). A total of 120 random colonies were selected from each clone library. Colony PCR products were digested with restriction enzymes MspI (TaKaRa). The digestion products were separated on a 2·5% agarose gel. Clones that yielded the same ARDRA patterns were grouped into the same taxonomic group. A total of 478 representative clones from the resulting groups were used for sequencing analysis.

T-RFLP analysis

Terminal restriction fragment length polymorphism (T-RFLP) analysis of bacterial 16S rRNA gene amplifications was performed as previously described (Liu et al. 1997; Lueders et al. 2006) using primers 27F-FAM and 1492R. PCR products were digested with restriction endonucleases MspI. Peak scanning of the restriction fragments was performed at Shanghai Jikang Biotechnology using a 3730 automated DNA sequencer.

The percentage abundance (Ap) of each terminal restriction fragment (T-RF) was calculated as Ap = ni/N × 100, in which ni represents the peak area of one distinct T-RF and N is the sum of all peak areas in a given T-RFLP pattern (Lukow et al. 2000). Ap values were determined for all T-RFs detected only in a size range between 50 and 600 bp for a given T-RFLP patterns. In this study, the length of defined T-RFs varied by not more than 1 bp. Only T-RFs with Ap ≥ 1% were included in the analysis. Cluster analysis of the profiles was based on percentage abundance and length of the T-RFs detected in them.

Sequence analysis

All assembled sequences were examined for chimeric artefacts using the Chimera Check software of the Ribosomal Database Project II (RDP II) (Maidak et al. 2000). The resulting sequences were grouped using the FastGroup II program (http://biome.sdsu.edu/fastgroup/fg_tools.htm). Sequences with sequence identity 97% or above were treated as an operational taxonomic unit (OTU). Individual sequences were phylogenetically classified to the taxonomic level of class and subclass according to the hierarchical taxonomy (Garrity et al. 2004) using the RDP Classifier tool release 9.0 (Wang et al. 2007) (Table 2). Sampling sufficiency of each library was determined as described by Kemp and Aller (2004) using the ‘Large Enough’ estimator available online at http://www.aslo.org/lomethods/free/2004/0114a.html (Fig. 2).

Figure 2.

 Sampling sufficiency estimates of clone libraries from Lake Dongping. The ‘Large Enough’ calculator was used to determine whether individual clone libraries were sampled sufficiently at the class level. If the estimated phylotype richness reached an asymptote, we inferred that the library was large enough to yield a stable estimate of phylotype richness. According to the figure, all samples appeared to have been sufficiently sampled.

Statistical analysis

Coverage was calculated as = 1 − (s/n), where s is the number of unique OTUs and n is the total number of clones in the library. The bacterial diversity, dominance and evenness of each cloning library were estimated at the phylum level using the Shannon, Simpson and Pielou indexes, respectively (Table 3). These parameters were calculated using the bio-dap software package (Thomas and Clay 2000).

Redundancy analysis (RDA) was applied to reveal relationships between the species in the bacterial community and the major environmental variables. Prior to detrended correspondence analysis (DCA) and RDA, environmental data (including TP, TN, TOC, NH4-N, NO3-N, TOC/TN and depth) were z-transformed and square-root transformations were performed on the relative abundance data. RDA was performed using canoco 4.5 software (Biometris, Plant Research International, Wageningen, the Netherlands) with the linear method because DCA run on species variables indicated that the length of the gradient of the first axis was short (<2) (Ter Braak and Prentice 1988; Jongman et al. 1995). Only those species present in at least one of the four libraries in all 12 samples were considered and the remaining rare species were downweighted in this study. Detrending was carried out in segments using the non-linear rescaling method. The significance of the first ordination and canonical axes together was assessed in permutation tests with 499 unrestricted Monte Carlo permutations (P < 0·05).

Nucleotide sequence accession numbers

DNA sequences obtained in this study have been deposited in the GenBank under the accession numbers GQ472311GQ472465, GQ261279GQ261320 and GU208211GU208514.

Results

Environmental characterization

Water depths at the sediment sampling stations were relatively shallow, ranging from 0·88 to 3·72 m (Table 1). No significant temperature and pH variations of lake waters were detected at different sampling stations. The pH of lake water ranged from 6·7 to 7·0. In July and October, lake water temperature ranged from 25 to 28°C and from 17 to 19°C, respectively (data not shown). The other key abiotic and chemical parameters measured at the six stations of Lake Dongping are summarized in Table 1. Organic matter and nutrients parameters at each sampling station displayed unique seasonal changes (Table 1). For instance, NO3-N and TOC content and TOC/TN values were generally higher in the wet season than in the dry season, whereas NH4-N content was generally higher in the dry season than in wet season. It is likely that the high TOC/TN resulted from greater organic matter inputs into the lake sediments from the Darwen River and the surrounding lands, because similar results have been reported for other lakes (Meyers and Ishiwatari 1993). However, primary production also contributed to the high TOC/TN because the chlorophyll content of the water column was higher in the wet season than in the dry season at all stations (data not shown). Overall, significant seasonal and temporal variations were evident in the sedimentary biogeochemical environments in Lake Dongping.

Table 1.   Chemical and physical characteristics of the 12 study samples
SampleNH4-N (mg kg−1)NO3-N (mg kg−1)TN (g kg−1)TP (g kg−1)TOC (g. kg−1)TOC/TNWater depth (m)
MonthStations
  1. Depth, NH4-N, NO3-N, TN, TP and TOC refers to depth of water, ammonium nitrogen, nitrate nitrogen, total nitrogen, total phosphorus and total organic carbon, respectively. The net contents of NH4-N, NO3-N, TN, TP and TOC in sediments were calculated in dry weight. All parameters (except for TOC/TN) are presented as means of three replicates. Statistical analysis (anova) was performed using spss software (IBM, Chicago, IL).

  2. Values with different letters are statistically significant ( 0·05).

July141·41a8·51a2·601a0·635a45·22a17·391·70a
238·40c9·71b1·768b0·610bc45·20bc25·571·62b
340·22cd9·51c3·184cd0·607c58·03d18·233·11c
437·08de12·40d2·376e0·526e59·54e25·063·72d
517·53fg8·52d0·940f0·451e25·19ef26·802·74e
617·49h7·99e1·305h0·605f32·76f25·100·88g
October164·10b3·15b2·442a0·722b29·14ab11·931·30b
235·21c3·26c0·962bc0·312c20·52c21·333·62c
344·31de4·37c3·091d0·512d51·57d16·683·15d
448·54ef4·51d2·722f0·519e55·01ef20·213·63e
540·32g4·29d1·776g0·466f25·45f14·332·77f
644·14h4·12e1·961h0·568g28·83g14·701·40h

Analysis of T-RFLP profiles

Thirty-three peaks were analysed, including 64·5-, 66·5-, 90-, 118-, 120-, 136·5-, 138-, 140·5-, 148-, 160-, 162·5-, 164-, 185-, 196-, 199-, 224-, 288-, 428-, 437-, 460-, 462·5-, 465-, 470-, 472-, 482·5-, 485-, 488-, 501-, 507-, 509-, 512-, 543- and 558-bp T-RFs. Figure 3 shows dendrogram of cluster analysis of T-RFLP profiles of seasonal differences in the bacterial communities in sediment samples collected from Lake Dongping. Patterns of bacterial communities at station 2 in October were noticeably different from those in the other 11 samples. The other 11 samples contained the 488-bp T-RF in high abundance, as well as the 136·5-, 138-, 148-, 485- and 507-bp T-RFs. However, the 66·5- and 199-bp T-RFs were only detected in October, while the 470-bp T-RF was only detected in July. The abundance of 65- and 488-bp T-RFs was generally higher in July than in October. Sediment bacterial diversity was higher in October than in July. Five additional T-RFs (69·5-, 197·5-, 491·5-, 556- and 567-bp) were only present in station 2 in October. Although the 556- and 567-bp T-RFs were only detected in one sample, they had high percentage abundance (Ap) at 4·9 and 20·2%, respectively. Overall, the dendrogram of cluster analysis of T-RFLP profiles suggested that bacterial communities in sediment samples displayed seasonal variations.

Figure 3.

 Dendrogram of cluster analysis of terminal restriction fragment length polymorphism profiles. Note: the number 1, 2, 3, 4, 5, and 6 refer to station 1, station 2, station 3, station 4, station 5, and station 6, respectively; the letter J and O refer to January and October, respectively.

Clone library coverage and diversity analysis

Twelve independent 16S rRNA gene clone libraries were constructed from sediment samples collected at the six stations. A total of 1427 clones were selected from 12 clone libraries. Of these clones, 485 clones were sequenced. The other clones were excluded from further analysis using ARDRA (Vergin et al. 2001). Seven sequences were identified as chimeric products and were excluded from further analysis.

The resulting 478 sequences are representative of 16 phyla (Table 2). The representatives of β-Proteobacteria were found in all 12 libraries and accounted for about one-fourth of the 16S rRNA clones recovered ranging from 6·7 to 46·4%. Members of δ-Proteobacteria (3·3–29·8%), Acidobacteria (0–18·4%) and γ-Proteobacteria (4·3–16·7%) also comprised a significant component of the bacterial community in sediment bacterial communities. The representatives of OP8 were only found at station 6 in two sampling times. Actinobacteria were only detected in samples from station 1 in October and station 4 in July. Spirochaetes and OP11 were only detected in station 4 in October and July, respectively. Members of β-Proteobacteria dominated eight clone libraries constructed from sediment samples collected, including station 1 (25·0%), station 5 (32·4%) and station 6 (36·6%) in October and station 2 (46·4%), station 3 (23·1%), station 4 (19·5%), station 5 (27·0%) and station 6 (36·1%) in July. Two libraries, derived from samples from stations 3 (24·5%) and 4 (29·8%) in October, were dominated by the representatives of δ-Proteobacteria. Bacterial communities in samples from station 1 were unique and were dominated by ε-Proteobacteria (20·5%) rather than β-Proteobacteria (18·2%) in July. Members of the Firmicutes predominated (40%) in the library derived from station 2 in October, which contained relatively few member of the Proteobacteria (up to 30%).

Table 2.   Distribution of clones in each sample
MonthStationsNo. of OTUs% of bacteria in the following bacterial class in each lake sediment sample
β-Proteobacteriaδ-Proteobacteriaγ-ProteobacteriaAcidobacteriaPlanctomycetesε-ProteobacteriaBacteroidetesα-ProteobacteriaFirmicutesVerrucomicrobiaNitrospiraChloroflexiGemmatimonadetesChlorobiCyanobacteriaDeferribacteresActinobacteriaOP8SpirochaetesOP11
  1. OTU, operational taxonomic unit.

July14418·211·46·82·39·120·56·89·14·50·00·09·12·30·00·00·00·00·00·00·0
22846·421·410·70·07·13·60·00·00·010·70·00·00·00·00·00·00·00·00·00·0
35223·113·55·87·79·67·73·83·83·81·93·80·09·65·80·00·00·00·00·00·0
44119·57·312·214·64·94·94·97·30·00·07·39·80·00·00·00·04·90·00·02·4
53727·018·98·113·50·02·70·05·45·40·05·48·10·02·72·70·00·00·00·00·0
63636·122·211·12·80·00·011·10·00·05·65·60·00·00·00·02·80·02·80·00·0
October13625·019·48·313·98·35·65·60·08·30·00·00·02·80·00·00·02·80·00·00·0
2306·73·316·76·70·00·00·03·340·010·00·00·00·00·06·76·70·00·00·00·0
34922·424·510·218·44·10·04·10·02·08·24·10·00·02·00·00·00·00·00·00·0
44719·129·84·317·00·00·04·30·00·06·46·40·04·32·12·10·00·00·04·30·0
53732·410·85·40·05·42·75·413·50·02·70·016·20·05·40·00·00·00·00·00·0
64136·617·19·84·94·94·90·02·49·80·00·00·04·90·00·02·40·02·40·00·0

Diversity indices of the 12 clone libraries are summarized in Table 3. The lowest diversity and evenness and the highest dominance in the same sampling time were all observed at station 2 (close to the Baliwan floodgate) (Table 3). The bacterial diversity based on the 16S rDNA clone library in most stations was generally higher in July than in October.

Table 3.   Species diversity for the 12 sediment samples
SampleNo. of OTUsCoverageShannon indexPielouSimpson
MonthStations
  1. OTU, operational taxonomic unit.

July14484·52·210·920·11
22885·31·470·820·27
35285·72·360·920·10
44185·42·340·940·09
53786·62·110·880·13
63684·21·810·820·19
October13683·12·080·900·12
23086·41·840·840·19
34980·71·980·860·15
44782·92·030·850·15
53787·12·000·880·15
64187·21·980·820·17

Redundancy analysis

Total variation that could be explained by environmental variation accounted for 1·000, as indicated by the sum of all four eigenvalues. Concerning the variance of species data, the first axis explained 35·4% of the total variation of the hybridization data, the first and the second axes explained 56·6% and all four axes explained 71·3%. Species–environment correlations were high (data not shown), especially for axes 1 and 2 (0·986 and 0·992), indicating a relationship between species and environmental variables.

NO3-N, NH4-N, TP, TN and TOC were positively correlated with each other and were negatively correlated with water depth and the values of TOC/TN (Fig. 4). Monte Carlo variables tests indicated that only one forward-selected environmental variable (TP) was statistically significant (P = 0·004, F-ratio = 2·98, with 499 permutations). Permutation tests also indicated that TP was an important determinant of bacterial assemblages. Planctomycetes showed the highest positive correlation with concentration of TP (Fig. 4). β-Proteobacteria, δ-Proteobacteria, ε-Proteobacteria, Bacteroidetes and Gemmatimonadetes were all positively correlated with TP. In contrast, Acidobacteria, γ-Proteobacteria, α-Proteobacteria, Firmicutes, Verrucomicrobia, Nitrospira, Chloroflexi, Chlorobi, Cyanobacteria and Deferribacteres were all negatively correlated with TP. Especially, Firmicutes showed negatively significant correlation with TP.

Figure 4.

 Redundancy analysis biplots. Triangles indicate bacterial communities, and arrows indicate environmental variables. Note: β-P, β-Proteobacteria; δ-P, δ-Proteobacteria; γ-P, γ-Proteobacteria; Acb, Acidobacteria; Pla, Planctomycetes; ε-P, ε-Proteobacteria; Bcd, Bacteroidetes; α-P, α-Proteobacteria; Fir, Firmicutes; Ver, Verrucomicrobia; Nit, Nitrospira; Chx, Chloroflexi; Gem, Gemmatimonadetes; Chb, Chlorobi; Cya, Cyanobacteria; Def, Deferribacteres. Depth, NH-N, NO-N, TN, TP, TOC, and TOC/TN refer to depth of water, ammonium nitrogen, nitrate nitrogen, total nitrogen, total phosphorus, total organic carbon, and the rabio of C : N respectively. Only phyla detected in at least one quarter of the 12 samples were considered. Arrows indicate the direction of increasing values of the respective variable, and the lengths of the arrows indicate the degree of correlation of the variable with community data. The angles between arrows indicate correlations between individual environmental variables. Extending a specific environmental vector in both directions and projecting a line perpendicular from the species onto the axis can assess the relationship of the species to a particular environmental variable; the closer the species is to the vector, the stronger the relationship between the particular species and the environment relationships.

Discussion

Organic matter and nutrient analyses revealed large seasonal and spatial differences in the biogeochemical properties of sediment samples collected from Lake Dongping. Particularly, the TOC/TN values of the lake sediment samples (Table 1) suggested that the proportion of land-derived organic matter was higher in sediment samples collected in July than in October (Meyers and Ishiwatari 1993). Total organic contents ranged from 2·5 to 5·8% (July) and from 2·1 to 5·5% (October) and were comparable to those reported for some sites of Lake Geneva (Haller, 2011).

Because of its intrinsic disadvantages (for example, low taxonomic resolution), T-RFLP provides only limited insight into the actual diversity of bacterial communities in environmental samples (Blackwood et al. 2007). Therefore, both T-RFLP and clone library construction were used in combination with phylogenetic characterization of the distinct bacterial assemblages in different sediment samples. Results of this study indicated that sediment bacterial communities displayed significant seasonal and temporal variations in their structure and composition (Table 1). These variations are likely ascribed to the different anthropogenic and environmental inputs in different locations (Tables 1–3). Redundancy analysis suggested that only TP was statistically important in determining the bacterial community structure of sediment samples from Lake Dongping (Fig. 4). Also, it is worth to point out that noticeably higher nutrient levels (Table 1) and greater bacterial diversity were observed at stations 3 and 4, which were close to the aquaculture zones. Because the complex anthropogenic activities and environmental inputs surrounding the lake, the dynamic variations of bacterial communities in responding to these inputs should be a interesting model for the further study of microbial ecological function in the lake ecosystems.

Significant distinctions were observed among the clone libraries derived from in different sediment samples at phylum levels (Table 2). Members of β-, δ- and γ-Proteobacteria were the dominant phylogenetic groups at of all 12 sediment samples. This observation was in agreement with 16S rRNA gene analysis of other lake sediments and bacterioplankton (Mallet et al. 2004; Nelson et al. 2007; Percent et al. 2008; Haller et al. 2011). As reported in the sediments of Lake Geneva (Haller et al. 2011), the relative proportion of the different proteobacterial subdivision did vary among different sediment samples collected in July and October. The β-Proteobacterial phylotypes were the predominant lineages in all clone libraries. In the eight of 12 sediment samples examined, the β-Proteobacterial phylotypes accounted for over 22% of the clones (Table 2). Thus, results of this study seem to support the previous report of the ability of β-Proteobacteria to degrade complex organic macromolecules (Kirchman 2002). Members of δ-Proteobacteria were more frequently retrieved from sediments than from the water columns (Spring et al. 2000) and were only present as the predominant phylotypes at stations 3 and 4 (in the aquaculture zone and the centre, respectively) in October and stations 2 and 6 (sewage discharge and wetland, respectively) in July. Deltaproteobacteria play major roles in anoxic sittings like meromictic lakes, sediments and anaerobic digesters (Lehours et al. 2007; Karr et al. 2005; Riviere et al. 2009). The predominance of δ-Proteobacteria in these stations likely reflected their relatively anaerobic environments.

Furthermore, proteobacterial sequences accounted for the vast majority (>80%) of the clones in sediment samples collected in October from sites 1, 3, 4, 5 and 6 with the exception of site 2 (43·3% only). In October, the lake received relatively little environmental perturbations because of little water inputs from the Dawen River at the south-eastern corner, no overflow discharges at the northern side in that season, and low-level aquaculture activities. On the other hand, proteobacteiral sequences contribute to relatively low percentage of the clones in sediment samples collected in July from sites 1, 3, 4, 5 and 6 with the exception of site 2 (over 78%). These five sites received significant environment or anthropogenic inputs during the summer. Together with other reports (Haller et al. 2011), results of this study seem to support that variation in proteobacterial populations may reflect the level of environmental impacts on the lake ecosystems. Overall, the predominance of Proteobacteria in sediment communities suggests that proteobacterial metabolism is actively involved in the functioning and processes of freshwater lake sediment ecosystems.

Additionally, the structure of bacterial communities in sediment samples collected at station 2 in October was uniquely different from those in the other samples. Members of Firmicutes were the predominant phyloytpes accounting for 40% of the clones in this station (Table 2). To the best of our knowledge, this is the first report of a predominance of Firmicutes in freshwater lake sediment. Obviously, the lack of organic matter and nutrients and increased anoxic environments may have contributed to the low abundance of Proteobacteria (Tables 1 and 2, and Fig. 4). According to the result of RDA, the members of Firmicutes in Lake Dongping sediments were not dependent on organic matter and nutrients of higher concentration (Fig. 4). Because Firmicutes are common in soil and in the human body (Felske et al. 2000; Ott et al. 2004; Aas et al. 2005; Lu et al. 2006), the large population of Firmicutes likely derives from the water inputs of the Dawen River, which maybe carry significant amount of soil, particulate organic matter, pollutants and fertilizers derived from the Dawen Basin. It is noteworthy that the water level of Lake Dongping has been regulated through drainage, which results in newly formed sediments near the floodgate. The nutrient levels at that station were considerably lower than those at the other stations in October. Furthermore, the low degree of evenness suggested that the composition and function of bacterial communities were unstable in October (Wittebolle et al. 2009). Accordingly, the dominance of Firmicutes was temporal in the fresh sediment. A further important characteristic of freshwater lake sediments is that they are not spatially separated from their adjacent habitats, but are a part of more complex ecosystems (Spring et al. 2000). So, future analysis of bacterial composition of the water of the Dawen River may help us to understand this interesting microbiological phenomenon.

The results of RDA suggest that the concentrations of principal nutrients were the most important environmental factors that influenced the structure of bacterial communities in sediments of Lake Dongping (Fig. 4). In freshwater lacustrine sediment ecosystems, greater amounts of nutrients are thought to support richer and more diverse bacterial communities through increased niche partitioning (Dykhuizen 1998). As a major nutrient for aquatic ecology, phosphorus has been recognized as the most critical nutrient limiting lake productivity (Doricha et al. 1984; Jin et al. 2005). In this study, the samples from station 1 in two seasons have most abundant phosphorus. And, they also have highest evenness of the clone libraries (Tables 1 and 3). It suggests that abundant phosphorus might keep bacteria community stability in Lake Dongping sediments. One of the most important factors determining the water phosphorus concentration is phosphorus release from underlying sediments (Wang et al. 2006). In Lake Dongping, phosphorus arose primarily from industrial and agricultural waste water (Jiang et al. 2002) reaching the water column and accumulating in the sediment. It is clear that the phosphorus inputs from outside ambient had significant impacted the bacteria community structure and stability in Lake Dongping sediments. The Monte Carlo variables tests in RDA indicated that other environmental variables were not statistically significant (P > 0·1, with 499 permutations) in RDA. However, an influence of synergistic reaction of these nutrients on the distribution of bacterial communities still cannot be excluded.

The results of the present study support the idea that environmental changes could be evaluated using the composition variation of bacterial communities. In the present study, the lowest degree of evenness (i.e. the highest dominance and the lowest diversity) was observed in four samples from stations 2 and 6 and indicated unstable microbial communities at these two stations. This shows that the environmental conditions were always changing in these two regions. Actually, sediments at stations 2 and 6, which are close to the wetlands and Baliwan floodgate, respectively, were severely impacted by anthropogenic activities. And, the wetland was readily influenced by fisheries and shipping because of its shallow water body. The sediments near the floodgate underwent resuspension, while the overflow of lake water was drained through the gate. Additionally, photosynthetic bacteria (including Chloroflexi, Cyanobacteria and Chlorobi) were detected in two-thirds of the clone libraries, indicating that the structure of bacterial communities in the lake sediments was influenced by the lake water transparency and possibly by eutrophication.

In summary, bacterial community compositions in Lake Dongping sediments displayed measurable seasonal and temporal variations (Tables 2 and 3 and Fig. 3). β-, δ- and γ-Proteobacteria were the predominant bacterial lineages recovered in the sediments of Lake Dongping, with variations occurring in different individual samples. Although it is premature to conclude how the diversity and composition of bacterial communities respond to variations in organic matter and nutrients features of sediment, our results indicated parallels in variability in physical and geochemical features and bacterial community composition and diversity in Lake Dongping sediments. Clearly, further studies are needed to determine the actual environmental inputs (e.g. river inputs, land run-off and aquaculture). Finally, our results clearly suggest a potential significant impact of TP on the structure of bacterial communities in the sediments of Lake Dongping.

Ancillary