Bacteria are important for ensuring and maintaining the environmental and ecological processes of river ecosystems [37]. Here, the bacterial communities in river sediments under different types of industrial pollution were analyzed by Illumina high-throughput sequencing technology. Overall, the sediment bacterial communities were dominated by Proteobacteria, Actinobacteria, Chloroflexi, Acidobacteria, Firmicutes and Bacteroidetes (Fig. 3 (b)). Wang et al. (2016) studied the community structure of aquatic bacteria in urban rivers and obtained similar results [38]. Su et al. (2018) investigated the bacterial communities in coastal sediments and found that Proteobacteria, Firmicutes, Chloroflexi, Acidobacteria, Bacteroidetes, Actinobacteria, Nitrospirae, Gemmatimonadetes and Planctomycetes predominated [39]. These results showed that these sediment samples shared the characteristic profile of high bacterial rank commonly found in other aquatic ecosystems.
Although there were similarly high taxonomic level characteristics between different sampling sites, the relative abundance was different among different sites (Fig. 3 (c)). According to the data from this study, Proteobacteria were the most abundant in SPS (food factory) and ZMS (lighting factory) and were mainly composed of Betaproteobacteria and Gammaproteobacteria (Fig. 4). Most Proteobacteria groups play very important roles in the decomposition of organic matter and circulation [40]. Further research found that the family Rhodocyclaceae (class Betaproteobacteria) not only dominates in SPS but also had significant differences in abundance at different sites (Fig. 4). This strain has extensive metabolic capabilities and can degrade multiple carbon sources, such as many aromatic compounds [41]. Therefore, some members of the family are active in the degradation of recalcitrant chemicals [42]. For example, the genus Dechloromonas is capable of degrading a variety of complex organic pollutants (Fig. 4) [43]. Therefore, the dominance of the family Rhodocyclaceae in SPS may indicate that the area was rich in organic wastewater compounds.
The most dominant phylum in FZS (textile mill) was Actinobacteria. Actinobacteria have been confirmed to play a pivotal role in the carbon cycle of freshwater ecosystems [44]. In fact, the data from this study showed that TOC was the lowest in FZS, and nutrition in this sample was relatively poor (Table S1). The results were consistent with previous studies, suggesting that Actinobacteria are indeed active in oligotrophic environments [45]. In addition, the family Sphingomonadaceae was significantly enriched in FZS, and there were significant differences among different sites (Fig. 4). The family Sphingomonadaceae is usually found in habitats contaminated by a high proportion of recalcitrant (poly) aromatic compounds of natural or anthropogenic origin [46-48]. Moreover, the genus Sphingomonas in Sphingomonadaceae can degrade various recalcitrant compounds (Fig. 4) [49]. Thus, the members of this genus can grow vigorously in polluted environments [50]. It is worth noting that the textile industry produces a large amount of wastewater, which contains a variety of chemical compounds, such as azo dyes, heavy metals, and surfactants [51]. Therefore, the relative advantages of the family Sphingomonadaceae in FZS indicated that they are well adapted to the environment in FZS and possibly utilize a wide variety of nutrients to resist or withstand environmental disturbances.
Firmicutes are well known for having many members that are able to degrade very recalcitrant organic compounds [43]. Few previous studies also found a dominance of Firmicutes in freshwater sediments [52]. However, the abundance of Firmicutes was highest in GGS (steel plant). The data from this study indicated that the content of TN, TP and TOC in GGS was the highest, especially the TOC, which was 2 to 3 times that of the other three groups of samples (Table. S1). Meanwhile, RDA showed that the abundance of Firmicutes was positively correlated with TN, TP, and TOC (Fig. 5). This is the same result as previous studies, suggesting that as copiotrophs or fast-growing organisms, Firmicutes can exist in carbon-rich environments that meet their high energy requirements and maintain their growth rates [53]. Moreover, Clostridiales are known as metal-coping bacteria and thrive in environments rich in metal contaminants such as GGS [54] (Fig. 4; Table S1). Therefore, the enrichment of Clostridiales may also be one of the reasons for the largest proportion of Firmicutes in GGS. These results indicated that the environmental conditions may select different bacterial species, which leads to different spatial distributions of bacterial populations.
Multiple studies have shown that environmental factors, such as temperature [38], nutrients [55], pH [56], water turbidity [57] and sediment particle size [58], often affect the composition and structure of bacterial communities. On the one hand, RDA showed that TN, TP and TOC were significantly related to the composition of bacterial communities in different types of industrial polluted sediments (P <0.01) (Fig. 5). TN, TP, and TOC are important factors for structuring bacterial communities in river sediments, which is consistent with other research results. In reality, microorganisms may prefer to use bioavailable forms of phosphorus, nitrogen, and carbon (e.g., PO43-, NO2-, NO3-, NH4+, etc.) and the forms may be more closely related to bacterial community composition [45, 59, 60]. Therefore, the relationship between microbial communities and detailed environmental factors needs to be studied in greater depth. On the other hand, the heavy metals Cu, Zn, and Cd were significantly correlated with the distribution of bacterial communities in different types of industrially polluted sediments (P <0.01) (Fig. 5). It has been documented that high concentrations of heavy metals can significantly reduce bacterial biomass in sediments [61]. This was consistent with our observations that the species richness and diversity in the GGS samples, which had the highest heavy metal concentrations, were the lowest (Fig. 2 (a, b); Table S1).
In summary, the differences in bacterial community composition in different types of industrially polluted sediments reflect the tolerance of OTUs to specific environments. The above abiotic factors, such as TN and TP, may directly change the composition of bacterial communities by affecting the growth of certain bacteria in the sediment. That is, changes in the physicochemical properties of river sediments caused by the input of different types of industrial wastewater drive the formation of different bacterial communities. Meanwhile, this also means that microbial communities in urban river sediments have potentially evolved phylogenetic versatilities under the long-term effects of industrial pollution.
The interrelationships among different microbial communities play a pivotal role in maintaining the structure, function and stability of microbial ecosystems [14]. In the network analysis of this study, most nodes belonged to three dominant phyla: Proteobacteria, Actinobacteria and Chloroflexi (Fig. 6 (a)). The keystone genera Gaiella, Denitratisoma, Anaeromyxobacter, Candidatus_Microthrix, and unclassified_p__Chloroflexi were the top five with the highest number of connections. This suggested that the other genera respond more strongly to the metabolites produced by these five genera [62]. Compared with other taxa in the network, keystone taxa play an important role in maintaining the network structure [22]. It is speculated that the disappearance of keystone taxa may lead to disintegration of the network [44]. Keystone taxa are by the definition the taxa essential for ensuring and maintaining stability, so their existence is by definition important for the stability of ecosystem structure and function. Additionally, the average relative abundances of the genera Gaiella, Denitratisoma, Anaeromyxobacter, Candidatus_Microthrix, and unclassified_p__Chloroflexi were all low (0.24% ~ 0.65%), suggesting the significance of rare genera in bacterial communities. Currently, rare genera are being increasingly recognized as crucial components of communities in biochemical processes and community assemblies [63]. Although the abundance of such genera may not have been high, more attention should be paid to them as key nodes in the microbial community [14].
In addition, the genera Denitratisoma, Anaeromyxobacter, and Candidatus_Microthrix were considered central species due to their high degree (> 60) and low betweenness centrality values (<200). The genus Denitratisoma contains denitrifying bacteria that contribute to the removal of nitrogen [64]. The genus Anaeromyxobacter is metal-reducing bacteria, and members of the bacteria can affect the mobility of metal contaminants [65, 66]. Moreover, previous studies have shown that the genus Candidatus_Microthrix helps in the removal of total nitrogen [67]. Therefore, these keystone taxa may play a pivotal role in ecological function processes.
Due to the modularity, the entire network was mainly divided into six modules (Fig. 6 (b)). Modularity may reflect habitat heterogeneity and divergent selection regimes [68]. Meanwhile, the habitat preference of microorganisms may also help determine their co-occurrence patterns [19]. Therefore, it can be reasonably found that microorganisms in different types of industrial polluted sediments tend to form distinct modules. The main taxa in module I may be bacteria involved in the biogeochemical C- and N-cycles. For instance, the genus Nocardioides can utilize multiple organic compounds as carbon source [69]. The genus Nitrospira consists of chemically autotrophic nitrite-oxidizing bacteria [60]. The genus Streptomyces has been shown to participate in the nitrogen cycle [70]. Other bacteria in module I included Microbacteriaceae, Rhodospirillaceae, Bradyrhizobiaceae and Streptomycetaceae, which are also involved in C and N cycling [71]. Apparently, these bacteria had the highest abundance in FZS, indicating that carbon and nitrogen cycling associated with microorganisms occurs frequently in this area. Sulfate-reducing bacteria (SRB) in module II, including the genera Clostridium_sensu_stricto_1, Defluviicoccus, and Desulfobulbus, were significantly enriched in GGS [72]. Interestingly, SRB plays a major role in repairing the environment polluted by heavy metals (Fe, Cu, Pb, Zn, etc.) [72]. Therefore, heavy metals in GGS may be one of the important factors driving these bacteria. In conclusion, the non-random assembly pattern of these bacteria indicates that the complexity of bacterial community structure and functional processes in river sediments under different types of industrial pollution seems to be dominated by environmental filtering and function-driven.
According to PICRUSt analysis, the overall functional profiles of bacterial communities in different river sediment samples were similar. Carbohydrate metabolism and amino acid metabolism were the dominant metabolic genes in the bacterial community (Fig. 7(a)). As core resources metabolism pathways, they were potential drivers of microbial community structure and function of microbial communities in rivers [36]. It is worth noting that the xenobiotics biodegradation and metabolism that we focus on had more advantage in FZS (Fig. 7(a)). The possible reasons were FZS accepted more xenobiotic compounds from industrial wastewater, and these compounds were used by bacteria as the sources of carbon, nitrogen, or energy. Some previous reports indicated that there was a correlation between the abundance of xenobiotic degradation genes and the rate of xenobiotic biodegradation [73, 74]. Therefore, biodegradation genes could be proposed as indicators of the presence of xenobiotics and their metabolites [45, 75]. Furthermore, FZS surpassed the other three groups in the functional profiles related to the degradation of 15 chemical pollutants (Fig. 7(b)). For instance, benzoate degradation and aminobenzoate degradation. Meanwhile, the enrichment of many organic pollutant-degrading bacteria was found in FZS, such as genera Sphingomonas, Mycobacterium, Novosphingobium, and Bacillus (Fig. 4) [76-78]. The predicted higher enrichment of chemical pollutant degradation pathways means that the pollution was more severe in FZS, while such functional response of bacterial community might accelerate the bioremediation of contaminated zone. In general, the PICRUSt algorithm determined the predicted functions of the microbial community, providing a general overview of the functional potential within the community. However, rarefaction of pooled DNA samples fails to capture the full extent of diversity present within the system, which is likely reflected in the predicted functional profile [79]. Furthermore, this method is influenced by phylogenetic differences between environmental samples and sequenced genomes [28]. Therefore, we propose that further studies are needed in this system that use metagenomic sequencing and marker gene studies, to fully assess gene categories.
In this study, because the pollutants at each sampling point have been discharged into the water through pipes or channels for a long time, the sampling conditions were restricted. Therefore, we only paid attention to the impact of long-term discharge of industrial wastewater on the bacterial community of river sediments, and we did not cover the research before wastewater discharge. This may cause some limitations to our research. However, our findings represent an important step in understanding the impact of long-term industrial pollution on bacterial communities in urban river sediments, and thus contributes to the increasing knowledge of microbial ecology in the urban river sediments.