River pollution is one of the most challenging environmental issues, but the effect of river pollution levels on the biofilm communities has not been well-studied. Spatial and temporal distribution characteristics of environmental parameters and the biofilm communities were investigated in the Qinhuai River basin, Nanjing, China. Water samples were grouped into three clusters reflecting their varying pollution levels of relatively slight pollution, moderated pollution, and high pollution by hierarchical cluster analysis. In different clusters, the biofilm communities mainly differed in the proportion of Actinobacteria, Firmicutes, and Proteobacteria. As the dominant classes of Proteobacteria, Alpha-, Beta- and Gammaproteobacteria seemed to show an upward trend followed by a small fluctuation in the abundance with the escalation of water pollution level. Results of redundancy analysis demonstrated that temperature, total nitrogen to total phosphorus ratios (TN/TP) and concentrations of ammonia nitrogen (NH3-N) and TN were mainly responsible for the variation in bacterial community structure. The occurrences of Alpha-, Beta- and Gammaproteobacteria were closely associated with higher temperature, higher concentrations of NH3-N and TN and a lower TN/TP ratio. This study may provide a theoretical basis for the water pollution control and ecological restoration in urban rivers under different pollution levels.

INTRODUCTION

It has been well known that urban rivers play a fundamental and important part in both human life and ecological balance. While they are used as the main collector of urban sewers and the corridors for plant dispersal, they are also the main link between terrestrial and aquatic habitats as part of the hydrological and nutrient cycles (Tiquia 2011). Urban rivers are generally surrounded by densely populated areas, and suffer severe environmental pressures resulting from extensive urban, industrial and agricultural activities (Köck-Schulmeyer et al. 2011). Thus they typically show poor water quality with high nutrient levels and tend to be defined as a single ‘poor’ category (Davenport et al. 2001). Research shows that 80% of urban rivers in China have been significantly polluted, particularly in China's populous eastern plains (Qiu 2011; Zhang & Xu 2011).

River biofilms can respond quickly to changes in the environment and are therefore regarded as promising and early-warning biological systems in contact with river pollution (Fechner et al. 2012). In aquatic environments, natural biofilms are also able to provide a wide variety of ecosystem services, involving organic matter processing and retention, energy flow and cycling of nutrients (Allan & Castillo 2007). The bacterial community in river biofilms is a vital point related to the amelioration of water quality, especially in urban rivers suffering heavy pollution. However, river pollution has led to changes in a series of physicochemical parameters of river water (Zhang & Xu 2011), which may influence the biofilm microbial community structure. It is therefore crucial to figure out the dominant physicochemical factors impacting the abundance and distribution of the bacterial community in natural biofilms, as well as reveal the relationship between microbial community structure and the pollution level in urban rivers in order to govern the river ecosystems in an integrative framework, which has not yet been explained clearly.

There are many environmental factors influencing the bacterial community within river biofilms. Temperature, pH and dissolved organic carbon have been determined to be related to the variation within riverine biofilm communities (Anderson-Glenna et al. 2008). Krause et al. (2012) showed that small changes in pH had direct effects on marine bacterial community composition, particularly members of Gammaproteobacteria, Flavobacteriaceae, Rhodobacteraceae, Campylobacteraceae and other less abundant groups. In the study by Ylla et al. (2012), it was shown that the biofilm microbial community responded to higher water temperature by increasing bacterial cell number, respiratory activity (electron transport system) and microbial extracellular enzymes (extracellular enzyme activity). Additional physical parameters such as light, hydrodynamics, and velocity also contribute to the bacterial community by affecting the nutrient levels and organic carbon input (Kaplan & Bott 1989; Villeneuve et al. 2010). Data regarding the fluctuations in the phylogenetic diversity of the bacterial community within biofilms of urban rivers under different pollution load are sparse. In the study by Brümmer et al. (2003), by comparing the abundance of Betaproteobacteria in two rivers of different pollution levels, a higher percentage was detected in the biofilms of the extremely polluted river than the less polluted river. To the best of our knowledge, this is the only report on the distribution of bacterial community in biofilms under different pollution levels.

In this study we regard the Qinhuai River as an important case study to make a through inquiry into the relationship between the bacterial community of biofilms and the pollution levels of urban rivers. The objectives of this study were to analyze the abundance and distribution of the bacterial community within natural biofilms under different levels of pollution in Nanjing, China; to figure out the influencing factors contribute to the variation in bacterial community within biofilms; and to reveal the relationship between microbial community structure in natural biofilms and the pollution levels so as to provide a theoretical basis for the succession of the bacterial community, water pollution control and ecological restoration in urban rivers under different pollution levels.

METHODS

Study area and sample collection

The Qinhuai River in Nanjing is an important tributary of the southern bank of the Yangtze River. It has three branches including the external Qinhuai River, the inner Qinhuai River and Qinhuai New River. A total of 20 sampling sites were selected along the Qinhuai River (Figure 1) for the analysis of a variety of physicochemical and biological parameters of water and biofilms attached along the river bank. Samples were collected in two seasons: summer in July 2014 and winter in January 2015. (In the rest of the paper, sample numbers with no prime symbol indicate summer samples; numbers followed by a prime symbol are the corresponding winter samples.) Three specimens for each water sample were collected and mixed respectively. Water samples taken from each site by submerging 1 L polyethylene bottles were filtered immediately through Whatman GF/F filters (0.22 μm pore size) for laboratory analyses of water quality variables. All water samples were kept at 4 °C before analysis. To raise a sufficient amount of natural river biofilm, 20 parallel polyvinyl chloride tubes (1 m in length) carrying 12 polycarbonate slides (75 × 25 × 1.5 mm) were modified for this study to hold artificial substrata for biofilm growth at the surface. The collectors were previously cleaned and sterilized by autoclaving and then submerged 1 m below the river surface parallel to the current flow for 21 days in summer (July 2014) and winter (January 2015) at 20 sampling sites. And then the collector was carefully removed from each site and the biofilm was scraped off with a sterile razor blade. Biofilms were washed carefully with sterilized deionized water to remove loosely attached cells from the surface. Afterwards, biofilms from the each site were transferred to sterile tubes (50 mL) respectively and stored at −20 °C until used in DNA extractions.
Figure 1

Distribution of sampling locations in the Qinhuai River in Nanjing.

Figure 1

Distribution of sampling locations in the Qinhuai River in Nanjing.

Monitoring of environmental parameters

River water physicochemical parameters, including temperature (T), pH, concentration of dissolved oxygen (DO), and conductivity (Cond) were measured in situ after sampling. A 1 L water sample was collected for laboratory analyses to acquire total nitrogen (TN) and ammonia nitrogen (NH3-N), which were measured in the laboratory within 24 h after fixing water samples with concentrated sulfuric acid. TN was measured by potassium persulfate oxidation–ultraviolet spectrophotometry and NH3-N was measured by Nessler's reagent spectrophotometry. Chemical oxygen demand (permanganate index, CODMn) was measured by titration analysis. The Winkler method was used for the analysis of biochemical dissolved oxygen (BOD5) by determining the difference between initial and 5-day oxygen concentrations in bottles assayed after incubation at 20 °C. Total phosphorus (TP) was determined by methods as explained in Kent et al. (2004) and Yannarell et al. (2003). The concentration of total organic carbon (TOC) was measured using standard methods described by Artigas et al. (2012). Arsenic (As), mercury (Hg) and selenium (Se) were analyzed using an atomic fluorescence spectrophotometer. Cadmium (Cd) was determined by inductively coupled plasma atomic emission spectrometry. Quality control measures included careful standardization, procedural blank measurements, and spiked and duplicate samples. The overall ionic charge balance of each sample was within ±5%.

DNA extraction and T-RFLP analysis

After melting at room temperature, each biofilm sample was transferred to preweighed sterile 35 mL centrifuge tubes and centrifuged at 14,000 × g for 8 min. Supernatants were decanted, and 5 g of the pellet was weighed out for DNA extraction. Genomic DNA was extracted from biofilm samples using E.Z.N.A® a soil DNA kit (Omega Bio-Tek Inc., USA) following the manufacturer's instructions. DNA quality was analyzed by agarose gel electrophoresis. The absorption ratios of DNA at 260/280 nm and 260/230 nm were assessed using an ND-2000 spectrophotometer (NanoDrop Inc., Wilmington, DE, USA). The bacterial community structure of each biofilm sample was detected using terminal restriction fragment length polymorphism (T-RFLP) method. Primer pairs 27F and 1492R were used to amplify 16S rRNA genes of bacteria from DNA extract. Primer 27F was labeled using 5-carboxyflurescein dye. Polymerase chain reaction (PCR) products were digested in duplicate with the restriction endonuclease (TaKaRa, Japan) using Hae III and Hinf I seperately and incubated at 37 °C for 3 h. The fluorescently labeled terminal restriction fragments (T-RFs) were run on an automated DNA sequencer (ABI Prism TM 3730). GeneMarker was used to automatically calculate the number, height and peak areas of T-RFs with markers ranging from 50 to 700 bp. By calculating the percentage of each T-RF peak area in all peaks within one sample, the relative abundance of each T-RF was standardized.

Clone library construction

Due to the potential discrepancy between in silico-determined T-RF length and actual T-RF length determined by sequencing (Wang et al. 2011), a 16S rDNA clone library was constructed to identify the phylogenetic taxa of each T-RF. The DNA extracts of the biofilm samples were mixed by equal volumes for the PCR amplification using the primer pairs 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-TACGGCTACCTTGTTACGACTT-3′). Each PCR mixture contained 2 μL of extracted DNA, 5 μL of 10× buffer, each 1 μL of 27F and 1492R primers, 1 μL of dNTP mix, 39.5 μL of ultra-pure water and 5 U of Ex Taq DNA polymerase in a total volume of 50 μL. Thermal cycler conditions were 95 °C for 5 min, followed by 38 cycles of 94 °C for 30 s, 55 °C for 40 s, 72 °C for 90 s, and 72 °C for 10 min. PCR amplification was used for the construction of clone libraries. A total of 104 partial 16S rRNA gene sequences were deposited in the GenBank, which were submitted to NCBI and given the accession numbers KR188878–KR188977. According to the T-RFLP profiles of cloned 16S rRNA genes, the origins of T-RFs were identified. The T-RFLP analysis of cloned 16S rRNA genes was conducted applying the same approach described above. Each peak was presumed to represent the phylogenetic affiliation determined by the cloned 16S rRNA gene sequences. The phylogenetic affiliation of the 16S rRNA gene sequences were identified by the Ribosomal Database Project (RDP) and BLASTN online. CLUSTAL W was used to align the representative sequences and the phylogenetic analysis was accomplished by the neighbor-joining method with software MEGA 6.0.

Statistical analyses

The variability of water quality of all water samples collected from the whole river basin was determined from hierarchical cluster analysis (CA) by software PAST 3.0, using the linkage distance, reported as Dlink/Dmax, which represents the quotient between the linkage distances for a particular case divided by the maximal linkage distance. CA of the microbial community structures of these samples was also performed to test the similarity between samples. The heat maps were generated using R software based on the T-RFLP profiles. The Shannon index, Simpson index and evenness index were calculated based on the T-RF counts. To explore the correlation between the abundance of 16S rRNA genes of different bacterial groups and a series of environmental parameters, redundancy analysis (RDA) with the Monte Carlo permutation testing (999 permutations) by software Canoco 4.5 (SCIENTIFIC software) was performed. A P-value <0.05 was considered significant when referring to the positive and negative correlation levels.

RESULTS AND DISCUSSION

Physicochemical characteristics of aquatic environment

Related physicochemical information about the water samples collected in July 2014 and January 2015 from Qinhuai River is presented in Table 1. The water pH was around 7.0 in both seasons. The temperature of water was similar among the sites in each season, with an average of 23.9 °C in summer and 6.9 °C in winter. High concentrations of TOC (35.40 mgL−1 in summer and 29.32 mgL−1 in winter on average), TN (5.6 mgL−1 and 11.27 mgL−1) and TP (1.04 mgL−1 and 1.92 mgL−1) were observed and indicated high nutrient levels in the Qinhuai River. The ion concentrations determined were all observed at μgL−1 levels and were placed in the following descending sequence: As > Cd > Se > Hg. The average concentrations of As, Cd, Se, Hg in studied sites were 1.1, 0.4, 0.2, and 0.03 μgL−1 in summer and 1.5, 0.6, 0.5, and 0.05 in winter, which showed a common upward trend from summer to winter. Comparing the parameter values in summer and winter, in general, lower water temperature and TOC and higher DO, CODMn, BOD5, TN, TP and heavy metal ion concentrations were observed in winter. It showed that in both seasons most water samples far exceeded class V of water quality standards of China according to Environmental Quality Standard for Surface Water (GB3838-2002). The national standard of water quality classification does not describe clearly the water pollution conditions in this river. Therefore, hierarchical agglomerative CA was used to group the water samples based on their water quality characteristics.

Table 1

Descriptive statistics of water quality variables in different positions of Qinhuai River in summer and winter

 Summer
Sample1234567891011121314151617181920Mean
pH 7.51 7.65 7.56 7.64 7.59 7.59 7.6 7.71 7.72 7.65 7.54 7.57 7.89 7.74 7.75 8.8 7.85 7.93 7.79 8.15 7.76 
T (°C) 22.4 22.2 22.2 22.4 21.9 21.5 26.6 26.1 26.3 25.8 25.7 25.8 24.5 23 23.1 23.3 23.1 23.7 23.3 23.6 23.9 
DO (mgL−13.32 3.96 3.12 3.01 3.54 2.09 1.57 3.17 2.98 2.9 2.33 1.78 5.59 5.13 3.53 4.03 6.21 6.72 5.43 7.59 3.9 
Cond (μscm−1412 419 412 455 411 484 474 477 466 506 540 535 426 496 545 553 567 558 555 613 495 
NH4+-N (mgL−11.12 1.87 5.42 4.37 2.67 2.18 3.86 3.23 3.18 3.54 2.84 4.13 2.79 2.85 2.77 0.74 2.84 1.69 1.57 0.85 2.73 
TN (mgL−11.93 1.98 7.05 5.55 7.29 6.91 8.27 8.33 7.89 8.34 5.72 8.76 8.16 7.4 5.79 1.87 1.94 5.3 1.91 1.47 5.6 
TP (mgL−10.38 0.36 1.52 0.87 0.98 1.98 1.55 1.98 1.35 1.24 1.32 1.35 1.88 1.24 0.55 0.37 0.4 0.66 0.38 0.28 1.04 
TN/TP 5.08 6.92 4.64 6.38 7.44 3.49 5.34 4.21 6.6 6.73 4.33 6.49 4.97 5.97 10.53 5.05 13.13 8.03 11.29 4.91 6.58 
CODMn (mgL−15.86 5.95 7.38 8.41 5.36 8.52 7.64 6.33 6.12 6.03 6.42 7.11 7.17 6.12 5.78 3.02 2.55 3.28 2.23 2.79 5.7 
BOD5 (mgL−12.31 3.55 1.38 3.84 2.15 5.39 2.78 2.41 3.88 3.14 3.86 4.32 4.59 3.24 3.01 1.29 1.18 1.85 1.04 1.18 2.82 
TOC (mgL−128.21 28.01 31.49 28.36 27.78 43.51 33.61 65.62 30.85 30.95 30.6 31.25 58.39 30.48 30.03 33.21 33.35 30.27 30.04 51.85 35.4 
As (μgL−10.1 0.4 1.1 0.6 0.7 0.3 1.3 2.1 0.6 3.7 1.4 4.2 0.8 0.3 2.2 0.3 1.2 0.8 0.3 0.2 1.1 
Hg (μgL−10.03 0.02 0.02 0.05 0.01 0.03 0.02 0.07 0.02 0.01 0.03 0.04 0.02 0.03 0.03 0.01 0.02 0.04 0.07 0.02 0.03 
Cd (μgL−10.2 0.3 0.4 0.3 0.3 0.5 0.4 0.8 0.3 1.2 0.4 1.5 0.5 0.3 0.3 0.2 0.1 0.2 0.4 0.3 0.4 
Cr (μgL−12.8 3.2 4.2 2.9 5.5 5.3 4.2 6.2 4.4 6.5 5.2 5.9 3.4 4.2 4.5 2.8 3.4 3.6 4.2 3.1 4.2 
Se (μgL−10.3 0.1 0.3 0.2 0.1 0.1 0.2 0.4 0.6 0.2 0.3 0.5 0.1 0.2 0.2 0.1 0.2 0.3 0.4 0.1 0.2 
Pb (μgL−13.1 2.2 2.9 2.5 2.6 2.4 3.5 3.3 2.7 2.5 3.1 3.8 3.4 2.6 2.9 1.9 1.5 2.4 2.3 1.9 2.7 
Cu (μgL−117 23 15 24 27 33 21 11 25 41 23 15 18 21 11 17.8 
Zn (μgL−122 27 31 46 38 64 88 42 27 54 43 53 92 64 34 51 22 18 24 42.5 
Fe (μgL−128 33 27 45 109 36 75 27 98 35 54 38 49 65 32 43 74 23 38 27 47.8 
Wq* V* V* V* V* V* V* V* V* V* V* V* V* V* V* IV – 
Clu* – 
 Winter
Sample1′2′3′4′5′6′7′8′9′10′11′12′13′14′15′16′17′18′19′20′Mean
pH 6.43 6.35 6.67 6.48 6.61 6.24 6.33 6.47 6.12 6.83 6.63 6.25 6.41 6.76 6.52 6.97 6.76 6.54 7.11 6.87 6.57 
T (°C) 6.4 6.2 6.7 6.9 7.1 7.5 7.8 7.9 7.8 7.3 7.1 7.4 7.8 6.8 6.9 5.7 6.2 6.3 6.1 5.9 6.9 
DO (mgL−18.36 7.55 6.08 5.98 5.82 4.16 3.32 5.43 5.29 6.46 4.23 3,02 6.21 8.34 7.12 10.33 11.29 11.38 12.23 11.87 7.4 
Cond (μscm−1628 693 674 681 642 613 594 722 705 789 678 745 653 684 704 741 812 711 832 816 705 
NH4+-N (mgL−11.97 2.23 6.96 6.14 5.21 7.17 6.98 7.04 5.89 6.82 5.91 7.88 8.35 6.14 5.39 1.77 3.31 4.02 1.94 1.73 5.14 
TN (mgL−14.12 3.79 11.13 8.98 9.75 15.12 14.36 18.66 15.87 13.64 8.43 20.31 25.88 13.78 12.52 5.98 4.53 6.39 3.39 3.12 10.9 
TP (mgL−10.89 1.21 2.79 2,13 1.79 2.88 2.57 2.48 2.76 2.42 2.13 2.64 2.95 2.38 0.98 1.03 1.12 1.74 0.93 0.62 1.91 
TN/TP 4.63 3.13 3.99 4.22 5.44 5.25 5.59 7.52 5.75 5.64 3.96 7.69 8.77 5.79 12.78 5.81 4.04 3.67 3.65 5.03 5.61 
CODMn (mgL−17.46 8.79 9.35 9.79 7.58 9.94 8.86 8.04 7.53 8.66 8.27 9.63 10.58 8.33 8.14 5.49 4.75 5.22 4.29 3.84 7.73 
BOD5 (mgL−14.17 5.13 6.03 6.32 3.26 4.69 3.48 4.12 3.72 5.08 4.83 5.32 6.83 4.75 4.19 3.42 2.59 2.87 1.97 1.78 4.23 
TOC (mgL−130.12 25.35 24.32 26.53 28.92 32.17 35.12 28.34 33.97 36.23 24.68 28.15 29.33 23.58 32.35 31.02 27.39 28.44 31.49 28.89 29.3 
As (μgL−10.2 0.3 0.8 1.1 1.2 0.5 1.1 2.4 0.9 4.2 2.3 5.1 0.7 0.5 3.2 2.1 0.9 1.1 0.7 0.5 1.5 
Hg (μgL−10.05 0.04 0.03 0.04 0.02 0.05 0.04 0.06 0.08 0.06 0.05 0.05 0.04 0.02 0.08 0.03 0.05 0.07 0.05 0.04 0.05 
Cd (μgL−10.4 0.4 0.6 0.2 0.5 0.6 0.9 0.5 0.8 1.3 0.7 1.8 0.4 0.8 0.7 0.5 0.2 0.3 0.6 0.2 0.6 
Cr (μgL−13.5 6.4 4.2 5.1 6.8 6.4 8.2 4.6 6.1 6.9 5.3 5.8 7.4 8.2 6.3 5.2 2.4 3.5 3.2 2.1 5.4 
Se (μgL−10.7 0.2 0.5 0.3 0.4 0.2 0.8 0.6 0.9 0.5 0.4 0.3 0.5 0.6 0.3 0.3 0.5 0.4 0.3 0.4 0.5 
Pb (μgL−12.3 2.8 3.3 3.1 1.9 3.7 4.1 3.2 3.5 3.3 2.8 3.8 3.2 3.4 2.5 2.7 2.3 3.1 2.2 2.1 2.9 
Cu (μgL−111 24 26 32 31 35 26 21 18 32 27 33 21 11 27 31 23 17 20 23.7 
Zn (μgL−142 55 38 25 58 62 55 92 68 57 52 28 77 53 44 51 33 41 18 29 48.9 
Fe (μgL−135 39 41 28 53 77 92 54 73 62 88 48 76 59 102 66 46 55 41 36 58.6 
Wq V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* – 
Clu – 
 Summer
Sample1234567891011121314151617181920Mean
pH 7.51 7.65 7.56 7.64 7.59 7.59 7.6 7.71 7.72 7.65 7.54 7.57 7.89 7.74 7.75 8.8 7.85 7.93 7.79 8.15 7.76 
T (°C) 22.4 22.2 22.2 22.4 21.9 21.5 26.6 26.1 26.3 25.8 25.7 25.8 24.5 23 23.1 23.3 23.1 23.7 23.3 23.6 23.9 
DO (mgL−13.32 3.96 3.12 3.01 3.54 2.09 1.57 3.17 2.98 2.9 2.33 1.78 5.59 5.13 3.53 4.03 6.21 6.72 5.43 7.59 3.9 
Cond (μscm−1412 419 412 455 411 484 474 477 466 506 540 535 426 496 545 553 567 558 555 613 495 
NH4+-N (mgL−11.12 1.87 5.42 4.37 2.67 2.18 3.86 3.23 3.18 3.54 2.84 4.13 2.79 2.85 2.77 0.74 2.84 1.69 1.57 0.85 2.73 
TN (mgL−11.93 1.98 7.05 5.55 7.29 6.91 8.27 8.33 7.89 8.34 5.72 8.76 8.16 7.4 5.79 1.87 1.94 5.3 1.91 1.47 5.6 
TP (mgL−10.38 0.36 1.52 0.87 0.98 1.98 1.55 1.98 1.35 1.24 1.32 1.35 1.88 1.24 0.55 0.37 0.4 0.66 0.38 0.28 1.04 
TN/TP 5.08 6.92 4.64 6.38 7.44 3.49 5.34 4.21 6.6 6.73 4.33 6.49 4.97 5.97 10.53 5.05 13.13 8.03 11.29 4.91 6.58 
CODMn (mgL−15.86 5.95 7.38 8.41 5.36 8.52 7.64 6.33 6.12 6.03 6.42 7.11 7.17 6.12 5.78 3.02 2.55 3.28 2.23 2.79 5.7 
BOD5 (mgL−12.31 3.55 1.38 3.84 2.15 5.39 2.78 2.41 3.88 3.14 3.86 4.32 4.59 3.24 3.01 1.29 1.18 1.85 1.04 1.18 2.82 
TOC (mgL−128.21 28.01 31.49 28.36 27.78 43.51 33.61 65.62 30.85 30.95 30.6 31.25 58.39 30.48 30.03 33.21 33.35 30.27 30.04 51.85 35.4 
As (μgL−10.1 0.4 1.1 0.6 0.7 0.3 1.3 2.1 0.6 3.7 1.4 4.2 0.8 0.3 2.2 0.3 1.2 0.8 0.3 0.2 1.1 
Hg (μgL−10.03 0.02 0.02 0.05 0.01 0.03 0.02 0.07 0.02 0.01 0.03 0.04 0.02 0.03 0.03 0.01 0.02 0.04 0.07 0.02 0.03 
Cd (μgL−10.2 0.3 0.4 0.3 0.3 0.5 0.4 0.8 0.3 1.2 0.4 1.5 0.5 0.3 0.3 0.2 0.1 0.2 0.4 0.3 0.4 
Cr (μgL−12.8 3.2 4.2 2.9 5.5 5.3 4.2 6.2 4.4 6.5 5.2 5.9 3.4 4.2 4.5 2.8 3.4 3.6 4.2 3.1 4.2 
Se (μgL−10.3 0.1 0.3 0.2 0.1 0.1 0.2 0.4 0.6 0.2 0.3 0.5 0.1 0.2 0.2 0.1 0.2 0.3 0.4 0.1 0.2 
Pb (μgL−13.1 2.2 2.9 2.5 2.6 2.4 3.5 3.3 2.7 2.5 3.1 3.8 3.4 2.6 2.9 1.9 1.5 2.4 2.3 1.9 2.7 
Cu (μgL−117 23 15 24 27 33 21 11 25 41 23 15 18 21 11 17.8 
Zn (μgL−122 27 31 46 38 64 88 42 27 54 43 53 92 64 34 51 22 18 24 42.5 
Fe (μgL−128 33 27 45 109 36 75 27 98 35 54 38 49 65 32 43 74 23 38 27 47.8 
Wq* V* V* V* V* V* V* V* V* V* V* V* V* V* V* IV – 
Clu* – 
 Winter
Sample1′2′3′4′5′6′7′8′9′10′11′12′13′14′15′16′17′18′19′20′Mean
pH 6.43 6.35 6.67 6.48 6.61 6.24 6.33 6.47 6.12 6.83 6.63 6.25 6.41 6.76 6.52 6.97 6.76 6.54 7.11 6.87 6.57 
T (°C) 6.4 6.2 6.7 6.9 7.1 7.5 7.8 7.9 7.8 7.3 7.1 7.4 7.8 6.8 6.9 5.7 6.2 6.3 6.1 5.9 6.9 
DO (mgL−18.36 7.55 6.08 5.98 5.82 4.16 3.32 5.43 5.29 6.46 4.23 3,02 6.21 8.34 7.12 10.33 11.29 11.38 12.23 11.87 7.4 
Cond (μscm−1628 693 674 681 642 613 594 722 705 789 678 745 653 684 704 741 812 711 832 816 705 
NH4+-N (mgL−11.97 2.23 6.96 6.14 5.21 7.17 6.98 7.04 5.89 6.82 5.91 7.88 8.35 6.14 5.39 1.77 3.31 4.02 1.94 1.73 5.14 
TN (mgL−14.12 3.79 11.13 8.98 9.75 15.12 14.36 18.66 15.87 13.64 8.43 20.31 25.88 13.78 12.52 5.98 4.53 6.39 3.39 3.12 10.9 
TP (mgL−10.89 1.21 2.79 2,13 1.79 2.88 2.57 2.48 2.76 2.42 2.13 2.64 2.95 2.38 0.98 1.03 1.12 1.74 0.93 0.62 1.91 
TN/TP 4.63 3.13 3.99 4.22 5.44 5.25 5.59 7.52 5.75 5.64 3.96 7.69 8.77 5.79 12.78 5.81 4.04 3.67 3.65 5.03 5.61 
CODMn (mgL−17.46 8.79 9.35 9.79 7.58 9.94 8.86 8.04 7.53 8.66 8.27 9.63 10.58 8.33 8.14 5.49 4.75 5.22 4.29 3.84 7.73 
BOD5 (mgL−14.17 5.13 6.03 6.32 3.26 4.69 3.48 4.12 3.72 5.08 4.83 5.32 6.83 4.75 4.19 3.42 2.59 2.87 1.97 1.78 4.23 
TOC (mgL−130.12 25.35 24.32 26.53 28.92 32.17 35.12 28.34 33.97 36.23 24.68 28.15 29.33 23.58 32.35 31.02 27.39 28.44 31.49 28.89 29.3 
As (μgL−10.2 0.3 0.8 1.1 1.2 0.5 1.1 2.4 0.9 4.2 2.3 5.1 0.7 0.5 3.2 2.1 0.9 1.1 0.7 0.5 1.5 
Hg (μgL−10.05 0.04 0.03 0.04 0.02 0.05 0.04 0.06 0.08 0.06 0.05 0.05 0.04 0.02 0.08 0.03 0.05 0.07 0.05 0.04 0.05 
Cd (μgL−10.4 0.4 0.6 0.2 0.5 0.6 0.9 0.5 0.8 1.3 0.7 1.8 0.4 0.8 0.7 0.5 0.2 0.3 0.6 0.2 0.6 
Cr (μgL−13.5 6.4 4.2 5.1 6.8 6.4 8.2 4.6 6.1 6.9 5.3 5.8 7.4 8.2 6.3 5.2 2.4 3.5 3.2 2.1 5.4 
Se (μgL−10.7 0.2 0.5 0.3 0.4 0.2 0.8 0.6 0.9 0.5 0.4 0.3 0.5 0.6 0.3 0.3 0.5 0.4 0.3 0.4 0.5 
Pb (μgL−12.3 2.8 3.3 3.1 1.9 3.7 4.1 3.2 3.5 3.3 2.8 3.8 3.2 3.4 2.5 2.7 2.3 3.1 2.2 2.1 2.9 
Cu (μgL−111 24 26 32 31 35 26 21 18 32 27 33 21 11 27 31 23 17 20 23.7 
Zn (μgL−142 55 38 25 58 62 55 92 68 57 52 28 77 53 44 51 33 41 18 29 48.9 
Fe (μgL−135 39 41 28 53 77 92 54 73 62 88 48 76 59 102 66 46 55 41 36 58.6 
Wq V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* V* – 
Clu – 

1 and 1′, Sancha estuary; 2 and 2′, Dashengguan Bridge; 3 and 3′, Xiaguan Bridge; 4 and 4′, Tiexin Bridge; 5 and 5′, National Defense Garden; 6 and 6′, Zhenhuai Bridge; 7 and 7′, Qingliangmen Bridge; 8 and 8′, Hanzhongmen Bridge; 9 and 9′, Sanshan Bridge; 10 and 10′, Fengtai Bridge; 11 and 11′, Yuhua Bridge; 12 and 12′, Wuding Bridge; 13 and 13′, Wenjin Bridge; 14 and 14′, Desheng Bridge; 15 and 15′, Yunliang River; 16 and 16′, Wetland Park; 17 and 17′, Shangyuan Bridge; 18 and 18′, Xiaolongwan Bridge; 19 and 19′, Caihong Bridge; 20 and 20′, Fangshan Bridge.

Wq: water quality based on Environmental Quality Standard for Surface Water (GB3838-2002).

IV: DO; 3–5 mg L−1; NH3-N, 1–1.5 mg L−1; TN, 1–1.5 mg L−1; TP, 0.2–0.3 mg L−1; CODMn, 6–10 mg L−1; V: DO; 2–3 mg L−1; NH3-N, 1.5–2.0 mg L−1; TN; 1.5–2.0 mg L−1, TP; 0.3–0.4 mg L−1; CODMn, 10–15 mg L−1; V*: DO, <2 mg L−1; NH3-N, >2.0 mg L−1; TN, >2.0 mg L−1; TP, >0.4 mg L−1; CODMn, >15 mg L−1.

Clu: clusters were classified in this study by hierarchical cluster analysis based on the water quality characteristics.

As shown in Figure 2, CA rendered a dendrogram, grouping a total of 40 water samples of the basin into three statistically significant clusters at (Dlink/Dmax) × 100 < 60. The cluster number was also decided by practicality of the results as there is ample information available on the study samples. Cluster A (Sample 1, 2, 16, 17, 19, 20), Cluster B (Sample 3–5, 10, 11, 14, 15, 18, 1′, 2′, 16′, 19′, 20′) and Cluster C (Sample 6–9, 12, 13, 3′–15′, 17′, 18′) correspond to relatively slight pollution (SP) samples, moderate pollution (MP) samples, and high pollution (HP) samples, respectively. This clustering was confirmed by the Environmental Quality Standard for Surface Water (GB3838-2002), reflecting their pollution levels. A few previous reports also indicated that the hierarchical agglomerative CA technique is a useful tool to render reliable classification of surface waters in a whole region and has successfully been applied to water quality programs (Kim et al. 2005; Shrestha & Kazama 2007).
Figure 2

Dendrogram showing clustering of samples according to the water quality characteristics of the Qinhuai River according to Ward's method (Ward 1963).

Figure 2

Dendrogram showing clustering of samples according to the water quality characteristics of the Qinhuai River according to Ward's method (Ward 1963).

As a typical urban river, Qinhuai River was characterized by low velocities and different levels of pollution pressures. It was obviously observed that all the samples in Cluster A were collected in summer, while most of the samples in Cluster C were collected in winter. Actually, the Qinhuai River basin suffered more serious pollution loads in winter with lower velocities and smaller water flux, thus presenting higher average concentrations of TN, TP, NH3-N, CODMn, BOD5 and heavy metal ions. It also showed that most sites in the midsection of Qinhuai River belonged to Cluster C, which were of the worst water quality with the highest nutrient level. It is probable that the existing Wudingmen sluice interrupts and alters the flow of water from upstream to downstream, as well as possibly presenting an impassable obstacle to pollutants.

Bacterial diversity and bacterial community structure of biofilms along Qinhuai River

In this study, the T-RFLP analysis was used to determine the bacterial diversity within biofilms in Qinhuai River. A variety of techniques have been used for the assessment of microbial community depending on the scale and objective of the studies, such as T-RFLP and next-generation sequencing (NGS) (Nancy et al. 2011). T-RFLP data can provide the most abundant microbial T-RFs but cannot identify species or even genera precisely, while NGS is able to identify rarer members of microbial communities (Pilloni et al. 2012). However, in the present study, the T-RFLP method can meet our need to serve as a relatively simple, fast and cost-effective molecular tool to figure out the typical characteristics of the bacterial community structure in urban river biofilms. A detailed overview of T-RF numbers and diversity indices generated from T-RFLP profiles in all the biofilm samples is presented in Table 2. The numbers of T-RFs were estimated to range from 39 to 108 in summer and 28 to 70 in winter. Generally diversity indices among samples in both seasons were also variable. Relatively high numbers of T-RFs and diversity were detected, especially in summer. As previously described, environmental seasonality, as well as the spatial variation, has the potential to lead changes in the bacterial community (Ortmann & Ortell 2014). However, in our study, generally not significant variation existed in the values of diversity indices in each season between samples (Table 2). In summer, an exceptionally high number of T-RFs (108) and diversity were detected in Sample 20, while in winter, the values of the Simpson and Shannon diversity indices were higher for Sample 16′. However, no obvious differences were found in the bacterial diversity between all samples.

Table 2

Overview of T-RFs and diversity index generated from T-RFLP profiles in different biofilm samples

 Summer
 Winter
SampleT-RFCoverageSimpsonShannonEvennessSampleT-RFCoverageSimpsonShannonEvenness
71 0.99 0.84 2.64 0.20 1′ 57 0.98 0.86 2.77 0.31 
84 0.98 0.82 2.63 0.22 2′ 52 0.97 0.82 2.83 0.15 
75 0.99 0.83 2.63 0.19 3′ 63 0.99 0.84 2.66 0.24 
52 0.99 0.82 2.74 0.24 4′ 49 0.94 0.82 2.64 0.29 
71 0.97 0.82 2.37 0.16 5′ 43 0.97 0.85 3.11 0.41 
49 0.98 0.87 2.77 0.22 6′ 38 0.92 0.80 2.89 0.26 
82 0.96 0.82 2.85 0.18 7′ 47 0.97 0.79 3.35 0.20 
56 0.92 0.85 3.05 0.22 8′ 49 0.97 0.84 2.37 0.28 
83 0.99 0.80 2.47 0.15 9′ 51 0.95 0.82 2.75 0.18 
10 47 0.99 0.77 2.39 0.26 10′ 45 0.99 0.78 2.39 0.37 
11 63 0.98 0.79 2.26 0.16 11′ 54 0.99 0.90 3.04 0.32 
12 57 0.99 0.88 2.99 0.24 12′ 34 0.98 0.88 2.88 0.28 
13 48 0.98 0.82 2.59 0.22 13′ 28 0.97 0.84 2.65 0.22 
14 39 0.99 0.82 2.65 0.42 14′ 41 0.99 0.82 1.89 0.19 
15 46 0.99 0.82 2.37 0.15 15′ 38 0.95 0.84 2.34 0.27 
16 69 0.98 0.84 2.64 0.29 16′ 54 0.96 0.90 3.58 0.38 
17 102 0.94 0.82 2.63 0.20 17′ 66 0.98 0.88 2.73 0.47 
18 88 0.99 0.82 2.63 0.25 18′ 57 0.96 0.82 2.36 0.28 
19 99 0.99 0.82 2.74 0.37 19′ 61 0.99 0.83 3.27 0.53 
20 108 0.96 0.90 2.97 0.56 20′ 70 0.98 0.85 2.57 0.25 
Mean 67 0.98 0.83 2.65 0.25 Mean 50 0.97 0.84 2.75 0.29 
 Summer
 Winter
SampleT-RFCoverageSimpsonShannonEvennessSampleT-RFCoverageSimpsonShannonEvenness
71 0.99 0.84 2.64 0.20 1′ 57 0.98 0.86 2.77 0.31 
84 0.98 0.82 2.63 0.22 2′ 52 0.97 0.82 2.83 0.15 
75 0.99 0.83 2.63 0.19 3′ 63 0.99 0.84 2.66 0.24 
52 0.99 0.82 2.74 0.24 4′ 49 0.94 0.82 2.64 0.29 
71 0.97 0.82 2.37 0.16 5′ 43 0.97 0.85 3.11 0.41 
49 0.98 0.87 2.77 0.22 6′ 38 0.92 0.80 2.89 0.26 
82 0.96 0.82 2.85 0.18 7′ 47 0.97 0.79 3.35 0.20 
56 0.92 0.85 3.05 0.22 8′ 49 0.97 0.84 2.37 0.28 
83 0.99 0.80 2.47 0.15 9′ 51 0.95 0.82 2.75 0.18 
10 47 0.99 0.77 2.39 0.26 10′ 45 0.99 0.78 2.39 0.37 
11 63 0.98 0.79 2.26 0.16 11′ 54 0.99 0.90 3.04 0.32 
12 57 0.99 0.88 2.99 0.24 12′ 34 0.98 0.88 2.88 0.28 
13 48 0.98 0.82 2.59 0.22 13′ 28 0.97 0.84 2.65 0.22 
14 39 0.99 0.82 2.65 0.42 14′ 41 0.99 0.82 1.89 0.19 
15 46 0.99 0.82 2.37 0.15 15′ 38 0.95 0.84 2.34 0.27 
16 69 0.98 0.84 2.64 0.29 16′ 54 0.96 0.90 3.58 0.38 
17 102 0.94 0.82 2.63 0.20 17′ 66 0.98 0.88 2.73 0.47 
18 88 0.99 0.82 2.63 0.25 18′ 57 0.96 0.82 2.36 0.28 
19 99 0.99 0.82 2.74 0.37 19′ 61 0.99 0.83 3.27 0.53 
20 108 0.96 0.90 2.97 0.56 20′ 70 0.98 0.85 2.57 0.25 
Mean 67 0.98 0.83 2.65 0.25 Mean 50 0.97 0.84 2.75 0.29 

In order to compare T-RFs, the relative 16S rRNA gene frequencies of individual T-RFs within samples were calculated and displayed in Figure 3. Among the total T-RFs, dominant T-RFs have been found to be 51 bp, 68–69 bp, 194–199 bp, 216–219 bp, 241–242 bp, 253–254 bp, 272 bp, 291–293 bp, and 401 bp, which represented major bacterial groups in bacterial communities, accounting for 62.3% abundance of bacterial community of each sample in summer and 67.9% in winter. The T-RF with a fragment size of 68 bp was detected in all samples, showing a relative abundance of 6.4% and 7.6% on average in summer and winter respectively as the dominant T-RF.
Figure 3

Heat maps of the relative abundance of T-RFs of bacterial communities among different samples revealed by T-RFLP profiles.

Figure 3

Heat maps of the relative abundance of T-RFs of bacterial communities among different samples revealed by T-RFLP profiles.

Figure 4 shows the phylogenetic affiliations of dominant microbes among entire biofilm samples. Clone libraries from bacteria genes were constructed for the assessment of bacterial diversity based on molecular biology. Examination of the database revealed that 15 phyla have been recovered from Qinhuai river biofilms in both seasons. The bacterial community was dominated by Alpha-, Beta-, Gamma- and Deltaproteobacteria, Firmicutes, Actinobacteria, Cyanobacteria, Acidobacteria, Bacteroidetes, Chlorobi, Deinococcus-Thermus, Glaucophyta, Nitrospira and Spirochaetes with relative abundances of 12.4%, 19.5%, 7.9%, 1.2%, 16.7%, 13.9%, 9.3%, 3.2%, 1.5%, 2.9%, 3.3%, 2.3%, 0.9%, and 1.7%, respectively. In the present study, sequences resembling the Betaproteobacteria, Firmicutes and Actinobacteria take up the top three most abundant taxa detected. This is basically consistent with previous study, which showed that Proteobacteria (Alpha, Beta and Gamma), Actinobacteria, Cyanobacteria and Bacteroidetes are the frequently detected bacteria in river biofilms.
Figure 4

Phylogenetic analysis of the representative partial 16S rRNA gene sequences from bacterial clone libraries. The sequences were aligned by Clustal W. The tree was constructed using the neighbor-joining method with 1,000 bootstrap replicates. Labels without GenBank accession numbers indicate the sequences obtained from clone libraries, and the others show their GenBank relatives. The colorful clades and leaves represent different phylogenetic affiliation at the phylum level. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wst.2016.224.

Figure 4

Phylogenetic analysis of the representative partial 16S rRNA gene sequences from bacterial clone libraries. The sequences were aligned by Clustal W. The tree was constructed using the neighbor-joining method with 1,000 bootstrap replicates. Labels without GenBank accession numbers indicate the sequences obtained from clone libraries, and the others show their GenBank relatives. The colorful clades and leaves represent different phylogenetic affiliation at the phylum level. Please refer to the online version of this paper to see this figure in color: http://dx.doi.org/10.2166/wst.2016.224.

Relative abundance of different phylogenetic groups of biofilms at phylum level in summer and winter is presented in Figure 5. It is observed that Proteobacteria were the most frequently encountered in summer and winter, comprising 38.4% and 35.9% of total sequences on average respectively, followed by Firmicutes (18.1% in summer and 15.1% in winter) and Actinobacteria (16.2% and 11.5%). Among different classes of the phylum Proteobacteria, Betaproteobacteria constituted the most dominant group (20.1% and 18.2%), followed by Alphaproteobacteria (12.2% and 11.9%), Gammaproteobacteria (6.2% and 9.0%) and Deltaproteobacteria (1.1% and 1.4%). The remaining nine recovered phyla make up 27.3% and 37.5% of the total sequences collected in our database in summer and winter respectively. At the family level, Comamonadaceae, Oxalobacteraceae, and Alcaligenaceae were the most abundant, which accounted for an average of 10.5% to 13.3% of sequences in total. Among them, Comamonadaceae sequences were the most abundant in summer and winter representing 21.4% and 23.6% of Betaproteobacteria-related sequences on average. These ratios were within the range of values reported previously in the investigation of biofilms (Araya et al. 2003a, 2003b; Manz et al. 1999).
Figure 5

Relative abundance of different phylogenetic groups at phylum level present in Qinhuai River of different sites in summer (a) and winter (b).

Figure 5

Relative abundance of different phylogenetic groups at phylum level present in Qinhuai River of different sites in summer (a) and winter (b).

Seasonal fluctuation in the abundance of bacteria was observed as shown in Figures 3 and 5. Higher fractions of Actinobacteria, Betaproteobacteria and Firmicutes were detected in biofilms collected in summer as compared with those in winter. In addition to these seasonally dependent patterns, differences in the bacterial community composition were observed between samples 2 and 19. The bacterial community was dominated by Firmicutes (42.27% of total) in sample 2, while the Actinobacteria took the highest proportion (43.78%) in sample 19. This may be explained primarily by their spatial location: Site 2 lies where the Qinhuai River empties into the Yangtze River and thus has large amounts of domestic sewage, whereas Site 19 is in the upstream of Qinhuai River and close to the water source and thus probably has better water quality.

Relationship between the bacteria community composition and the pollution levels

As shown in Figure 5, significant changes were observed in the proportion of Actinobacteria, Firmicutes, and Proteobacteria among biofilm samples in different clusters. As the dominant classes of Proteobacteria, Alpha-, Beta- and Gammaproteobacteria showed an increase with the increasing water pollution, whereas their abundance would slightly change when their living environment was seriously polluted. The proportion of Alpha- and Betaproteobacteria would slowly decrease when their living environment was seriously polluted, while Gammaproteobacteria showed the opposite trend. In this case, a few bacteria may therefore have the potential to be used as indicators of pollution levels.

Betaproteobacteria was the most dominant group both in summer and in winter. It has been described as the dominant group within biofilms and contributes to the process of ammonia oxidization and pollutant degradation (Brümmer et al. 2000). In different clusters, the biofilm communities differed mainly in regard to the relative abundance of their predominant Betaproteobacteria. In addition, it is observed that the percentage of Betaproteobacteria was larger in Cluster B (22.45%) and C (23.91%) than Cluster A (12.92%) in summer. Approximately 24.56% of the biofilm clones were detected in Cluster B in winter, which is higher than that of Cluster B (16.27%) in summer. A relationship between the abundance of Betaproteobacteria and the degree of pollution was observed whereby the abundance of Betaproteobacteria increased with the increasing pollution level followed by a small fluctuation in samples of HP areas. Betaproteobacteria has been reported to be probably related to the pollution levels in riverine basins. Although the dominant Betaproteobacteria described here may be tolerant to the pollution present, the survival ability would decline in HP areas with excessive nitrogen and phosphorus (Brümmer et al. 2003).

Coincidentally, the distribution characteristics of Alphaproteobacteria were similar to those of Betaproteobacteria. Members of this class also showed an increase with the increasing pollution levels but slightly declined in HP samples, suggesting that the unknown mechanisms may keep the abundance of the Alphaproteobacteria low in HP areas. The most likely mechanism may relate to the competition ability of Alphaproteobacteria for organic and inorganic substrates. Also, Gammaproteobacteria is another group that is widely distributed in lakes all over the globe, but always not the dominant community in most aquatic environments (Newton et al. 2011). In our study, this group is also an important section in the bacterial community structures that were related to the pollution levels. An obvious trend in the abundance of Gammaproteobacteria was observed in that it seemed to be more abundant in the more polluted clusters in both seasons. The percentage of Gammaproteobacteria in Cluster A was 3.8% on average, while it was enhanced in Cluster B and Cluster C with a percentage of 6.2% and 8.4% respectively. The Gammaproteobacteria grew significantly faster than the average river bacterioplankton and exhibited even faster growth rates when nitrogen and phosphorus were added to enclosures (Gasol et al. 2002). Thus, the increased abundance of Gammaproteobacteria in Cluster B and C may be explained. Regarding the Firmicutes phylum, no obvious indication of occurrence regularity was found. In summer the abundance of Firmicutes seemed to be the lowest in Cluster B, whereas it showed no regularity in winter among different samples. Previous studies also showed that no seasonal or along-river variations were found for Firmicutes in river systems (Tirodimos et al. 2010), indicating that the mechanism of the effects of environmental conditions on the Firmicutes may be a complex process. The Actinobacteria phyla were found to occupy lower percentages in Cluster C than in Cluster A and B. It suggested that the occurrence of Actinobacteria may be association with less eutrophic conditions.

The three most abundant portions of analyzed sequences at the family level fell into the Comamonadaceae, Oxalobacteraceae and Alcaligenaceae, which all belong to the Betaproteobacteria class. In addition, the bacteria communities differed mainly in regard to the relative abundance of Comamonadaceae in different clusters. As an important member of Betaproteobacteria, Comamonadaceae-related microorganisms have been reported to be the most abundant typical freshwater taxa in biofilms (Liu et al. 2012). In our study, Cluster B contained the most abundance of Comamonadaceae in both seasons, almost 1.5–2 times higher than Cluster A and C. While for Oxalobacteraceae and Alcaligenaceae of Betaproteobacteria, as another dominant groups, Cluster B also appeared to be more favorable for their survival than the other clusters. However, with the increase of the pollution degree, the abundance of Burkholderiaceae (Betaproteobacteria) seemed to show a decline. However, comparison of the summed abundance of the sequences of the major families (Comamonadaceae, Oxalobacteraceae, Alcaligenaceae, Burkholderiaceae) belonging to Betaproteobacteria in different clusters revealed that the trend climbed up and then slightly fluctuated with the increasing pollution level, consistent with the trend of the abundance of Betaproteobacteria. However, the abundance of Pseudomonadaceae, belonging to Gammaproteobacteria, showed a decline with the increasing pollution degree, which was opposite to the trend in Gammaproteobacteria. It may indicate that different families belonging to the same class may show different response to the same environmental condition.

Factors determining the composition of functional bacterial communities

In the present study, we have shown that when a shallow river shifts between different pollution levels, the bacterial community structure also changes considerably. Figure 6(a) reveals the relationship between the distribution of the bacterial community in biofilm samples and environmental parameters of river water by RDA. In the RDA model, pH, DO, Cond, COD, NH3-N, TN and the TN/TP ratio were the environmental variables that statistically better explained the variations in the distribution of the bacterial groups among samples. The eigenvalues of the first two RDA axes were 0.218 and 0.153, respectively. The two axes accounted for 49.8% of the observed variation in the species data. High species–environment correlation coefficients (0.906 and 0.925 for the first two axes, respectively) indicated that the species compositions were strongly related to the measured environmental variables. Environmental parameters including pH, DO, Cond and the TN/TP ratio were positively correlated with the first species axis. TN/TP was found to be the most significantly correlated factor (r = 0.4986, P < 0.05). Among the other factors in the first species axis, temperature showed the strongest negative correlation (r = −0.6437, P < 0.05). Generally lower correlation coefficients were found in the second species axis. TN was the most remarkable one (r = 0.5316, P < 0.05), followed by NH3-N (r = 0.3851, P < 0.05). Conversely, DO was the most negatively related to the second RDA axis (r = −0.2757, P < 0.05). The RDA analysis showed that the distribution of bacterial assemblages was mainly driven by changes in temperature, the TN/TP ratio and concentrations of NH3-N and TN. Other factors such as pH and concentrations of DO, Cond, and CODMn might also contribute to the observed difference in the species distribution.
Figure 6

Results of redundancy analysis biplots. (a) Bacterial community structure (OUT assignment distance cutoff 0.1) in relation to the environmental parameters. Codes 1 to 20 indicate summer sites and codes 1′ to 20′ indicate winter sites. (b) Dominant bacterial community (ALP: Alphaproteobacteria; Act: Actinobacteria; Bet: Betaproteobacteria; Fir: Firmicutes; Gam: Gammaproteobacteria) in relation to the environmental parameters. The directions and lengths of the arrows indicate importance and correlation to the respective axes.

Figure 6

Results of redundancy analysis biplots. (a) Bacterial community structure (OUT assignment distance cutoff 0.1) in relation to the environmental parameters. Codes 1 to 20 indicate summer sites and codes 1′ to 20′ indicate winter sites. (b) Dominant bacterial community (ALP: Alphaproteobacteria; Act: Actinobacteria; Bet: Betaproteobacteria; Fir: Firmicutes; Gam: Gammaproteobacteria) in relation to the environmental parameters. The directions and lengths of the arrows indicate importance and correlation to the respective axes.

With regard to the five dominant bacteria groups in Figure 6(b), Alpha-, Beta- and Gammaproteobacteria were closely associated with higher temperature, TN, and NH3-N, and lower ratio of TN/TP. In contrast, Actinobacteria occurrence was correlated with higher ratio of TN/TP, DO and lower NH3-N, and Firmicutes to lower temperature. In Figure 6(b), Actinobacteria appeared to be the most positively related to the concentration of DO and the most positively related to the concentration of NH3-N, whereas Gammaproteobacteria showed clear opposite correlations for these variables. It was suggested that the relative abundance of these phyla was correlated with the same parameters either positively or negatively. To conclude, although the influencing parameters were complex both in summer and in winter, the distributions of biofilm bacterial communities were significantly related to temperature, the TN/TP ratio and concentrations of NH3-N and TN.

Temperature was generally considered as the dominant parameter controlling the bacterial community structure in aquatic environments (Winter et al. 2007; Liu et al. 2013). In our study, Alpha- and Betaproteobacteria both showed obvious negative correlation with temperature. In Nanjing, obviously seasonal diversity of temperature between summer and winter was observed, thus resulting in the abundance difference in biofilm bacteria community between seasons.

Relations between specific bacteria groups and TN and NH3-N in freshwater environment were discovered in previous study. In regions with high nitrate concentrations, Beta- and Gammaproteobacteria in the stream biofilms seemed to be the most abundant groups, while in low-nitrate environment, Alphaproteobacteria was the more abundant group (Gao et al. 2005). Consistently, Beta- and Gammaproteobacteria were obviously positively related to the TN concentration in our study. However, an inconsistent result was observed in that Alphaproteobacteria was also positively correlated with the TN concentration.

Opinions are divided when it comes to the role of TN/TP ratio in the microbial diversity. Microcystis growth was related to the TN/TP ratio, with the maximum value at an optimum TN/TP ratio and the minimum values when the ratio was 0 or too high, which may be explained by the nutrient elution tests under controlled TN/TP ratios (Amano et al. 2008). Several ranges of TN/TP ratios have been suggested for use in the classification of nutrient limitation within a lake or reservoir. However, the relationship between the TN/TP ratio and bacteria community in biofilms has never been mentioned before. Our study showed that Firmicutes appeared to be the most positively related to the TN/TP ratio, whereas Beta- and Gammaproteobacteria showed negative correlations to this variable. It may be a consequence of the combined action of other environmental parameters in this study.

CONCLUSIONS

The abundance and diversity of relative bacterial communities of biofilms in the Qinhuai River revealed a significant difference between regions under different pollution levels. Betaproteobacteria dominated the community, followed by Firmicutes and Actinobacteria. Significant changes were observed in the proportion of Actinobacteria, Firmicutes, and Proteobacteria among samples in different pollution levels. The abundance of Alpha- and Betaproteobacteria commonly showed an increase from MP areas to SP areas, whereas those of Alpha-, Beta- and Gammaproteobacteria did not change notably in HP areas compared with MP areas. The distribution of bacterial assemblages was mainly driven by changes in temperature, the TN/TP ratio and concentrations of NH3-N and TN. Alpha-, Beta- and Gammaproteobacteria were closely associated with higher temperature, lower TN/TP ratio and concentrations of TN and NH3-N, while the Actinobacteria occurrence was correlated with higher TN/TP ratio, DO and lower NH3-N, and Firmicutes to lower temperature. Nevertheless, further studies should focus on the beneficial bacteria community that contributes to the biological self-purification of polluted urban rivers, thus to promote effective strategies of water pollution control and ecological restoration.

ACKNOWLEDGEMENTS

The study was financially supported by the National Natural Science Foundation of China (No. 51322901 and No. 51479066), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (51421006), the Fundamental Research Funds for the Central Universities (2016B10614), the Priority Academic Program Development of Jiangsu Higher Education Institutions, and Top-notch Academic Programs Project of Jiangsu Higher Education Institutions (TAPP).

REFERENCES

REFERENCES
Allan
J. D.
Castillo
M. M.
2007
Stream Ecology: Structure and Function of Running Waters
.
Springer Science & Business Media
,
Dordrecht, The Netherlands
.
Amano
Y.
Machida
M.
Tatsumoto
H.
George
D.
Berk
S.
Taki
K.
2008
Prediction of Microcystis blooms based on TN: TP ratio and lake origin
.
Scientific World J.
8
,
558
572
.
Anderson-Glenna
M. J.
Bakkestuen
V.
Clipson
N. J.
2008
Spatial and temporal variability in epilithic biofilm bacterial communities along an upland river gradient
.
FEMS Microbiol. Ecol.
64
(
3
),
407
418
.
Artigas
J.
Soley
S.
Perez-Baliero
M. C.
Romanı
A. M.
Ruiz-Gonzalez
C.
Sabater
S.
2012
Phosphorus use by planktonic communities in a large regulated Mediterranean river
.
Sci. Total Environ.
426
,
180
187
.
Davenport
A.
Gurnell
A.
Armitage
P.
2001
Classifying urban rivers
.
Water Sci. Technol.
43
(
9
),
147
155
.
Fechner
L. C.
Versace
F.
Gourlay-Francé
C.
Tusseau-Vuillemin
M. H.
2012
Adaptation of copper community tolerance levels after biofilm transplantation in an urban river
.
Aquat. Toxicol.
106
,
32
41
.
Gao
X.
Olapade
O. A.
Leff
L. G.
2005
Comparison of benthic bacterial community composition in nine streams
.
Appl. Environ. Microbiol.
40
,
51
60
.
GB3838-2002
2002
Environmental Quality Standard for Surface Water
.
Environmental Science Press
,
Beijing, China
.
Kent
A. D.
Jones
S. E.
Yannarell
A. C.
Graham
J. M.
Lauster
G. H.
Kratz
T. K.
Triplett
E. W.
2004
Annual patterns in bacterioplankton community variability in a humic lake
.
Microbiol. Ecol.
48
(
4
),
550
560
.
Köck-Schulmeyer
M.
Ginebreda
A.
Postigo
C.
López-Serna
R.
Pérez
S.
Brix
R.
Liorca
M.
de Alda
M. L.
Petrović
M.
Munné
A.
Tirapu
L.
Barceló
D.
2011
Wastewater reuse in Mediterranean semi-arid areas: the impact of discharges of tertiary treated sewage on the load of polar micro pollutants in the Llobregat river (NE Spain)
.
Chemosphere
82
,
670
678
.
Krause
E.
Wichels
A.
Giménez
L.
Lunau
M.
Schilhabel
M. B.
Gerdts
G.
2012
Small changes in pH have direct effects on marine bacterial community composition: a microcosm approach
.
PLoS One
7
(
10
),
755
771
.
Newton
R. J.
Jones
S. E.
Eiler
A.
McMahon
K. D.
Bertilsson
S.
2011
A guide to the natural history of freshwater lake bacteria
.
Microbiol. Mol. Biol. Rev.
75
(
1
),
14
49
.
Wang
X.
Wen
X.
Yan
H.
Ding
K.
Zhao
F.
Hu
M.
2011
Bacterial community dynamics in a functionally stable pilot-scale wastewater treatment plant
.
Bioresour. Technol.
102
(
3
),
2352
2357
.
Ward
J. H.
Jr
1963
Hierarchical grouping to optimize an objective function
.
J. Am. Stat. Assoc.
58
,
236
244
.
Winter
C.
Hein
T.
Kavka
G.
Mach
R. L.
Farnleitner
A. H.
2007
Longitudinal changes in the bacterial community composition of the Danube River: a whole-river approach
.
Appl. Environ. Microbiol.
73
(
2
),
421
431
.
Yannarell
A. C.
Kent
A. D.
Lauster
G. H.
Kratz
T. K.
Triplett
E. W.
2003
Temporal patterns in bacterial communities in three temperate lakes of different trophic status
.
Microbiol. Ecol.
46
(
4
),
391
405
.