Introduction

For decades, water eutrophication has been a major source of freshwater pollution in China (Wang et al. 2019) and has caused severe problems such as biodiversity loss and algal blooms (Yin et al. 2020). During the rainy season, agricultural systems release abundant nitrogen (N) and other nutrients into water bodies; therefore, organisms in these water bodies often experience high nutrient supplies (Yan et al. 2018). As the main primary producers of wetland ecosystems, plants (Zhang et al. 2016; Zhao et al. 2019), especially submerged macrophytes, are important participants in lake ecosystems and play an important role in nutrient removal (Qin et al. 2019) and other processes in constructed and natural water bodies. Submerged macrophytes can assimilate a variety of nutrients, such as organic and inorganic N (i.e., NO3−, NO2−, NH3, and NH4+) and phosphorus (P), which can result in chronic toxicity to Hydrilla verticillata at high concentrations (Wang et al. 2010). Yan et al. (2018) reported that P. malaianus, Vallisneria natans, and Hydrilla verticillata had strong nutrient removal effects, indicating that nutrients were easily assimilated by these aquatic plants. Therefore, submerged macrophytes are important for water self-purification systems and for maintaining ecological balance (Han et al. 2018).

Microbial communities are important parts of lake ecosystems (Yan et al. 2018) and play an important role in regulating the water quality of polluted lakes (Zhang et al. 2018). Microbial communities also play vital roles in biogeochemical cycling in the sediments of freshwater lakes (Liu and Yang 2020). Changes in microbial communities may reflect the status of the environment (e.g., water quality) (Liu et al. 2020). Excess N and P in wetlands can be removed through biological, physical, and chemical processes (Ballantine et al. 2014). Sediment N cycling is an important biological process for permanent N removal (Wu et al. 2021). Submerged macrophytes can provide oxygen and appropriate environmental conditions for epiphytic bacterial communities (Bustamante et al. 2011). Wu et al. (2021) reported the direct effects of submerged macrophytes on the bacterial community and their indirect effects through altering sediment C and concluded that a greater development of submerged macrophytes in lakes is associated with greater nitrogen removal from lake sediments. Furthermore, at night, respiration of submerged macrophytes may shift from aerobic to anaerobic because the conditions at night are favorable for anaerobic bacteria (Eriksson 1999). Changes in microbial communities can reflect the stability of effluent and sediment ecosystems. However, it is not clear how the cultivation of submerged plants for purifying water affects the bacterial community in rhizosphere sediments and nearby water.

Nansi Lake, which is located in Shandong Province, covers an area of 1266km2 and is the largest freshwater lake in northern China (34°27′–35°20′N, 116°34′–117°21′E) (Tian et al. 2013); as the main reservoir lake and biodiversity protection area in the east route of the South to North Water Diversion Project, it has an important impact on water quality (Zhang et al. 2021). P. lucens Linn. is an important submerged plant in Nansi Lake and has a good water purification effect. However, how the bacterial communities in water and rhizosphere sediments differ in areas cultivated with P. lucens Linn. (PL) from those in control areas without P. lucens cultivation (CK) remains unclear. In this study, we collected 24 water and sediment samples from PL and CK areas to study the effects of this plant on water quality, nutrients and the microbial community.

Materials and methods

Site description and sampling

The sampling site was located in Nansi Lake (34°37′N, 117°12′E and altitude 27 m) in Jining, Shandong Province, China (Fig. 1). This region has a temperate monsoon climate with an average annual temperature of 15 °C and a mean annual precipitation of 775 mm. With the development of industry and the increased application of pesticides along the lake area, the industrial and agricultural wastewater and domestic sewage discharged to Nansi Lake are increasing yearly. In 2002, water quality was inferior to class V, which corresponds to “polluted” and “dirty” (Kondrat’eva et al 2009), according to the “China surface water quality standard” (GB3838–2002). It thus has great impacts on agriculture, fisheries, and the domestic water supply. Near Wanzhuang Village (34°37′N, 117°12′E) in Nansi Lake, there is a large area (approximately 1 ha) where only P. lucens Linn. is cultivated (PL) that appears to be in a clear state and an adjacent area (approximately 500 m away) in which the water is muddy and lacks any aquatic plants (CK). P. lucens Linn. was morphologically identified by Dr. Fengyue Shu using taxonomic keys, and voucher specimens were deposited at Qufu Normal University (School of Life Sciences), China. The time of sampling was November 2020, when P. lucens Linn. was in the declining phase. We randomly established three plots (approximately 10 × 10 m) in each of the two areas and selected two sampling sites in each plot. For water samples, 3.0 L water (20 cm depth) was collected at each site and immediately transported to the laboratory at low temperature 0–4 °C. Then, the 1.5 L of each water samples were filtered through 0.22-μm membrane filters (Millipore, USA), and the filtrate was stored at − 80 °C until DNA extraction. The remaining 1.5 L water sample was used for physicochemical analysis (Chao et al. 2021). For sediment samples, we collected surface sediment (0–20 cm) by a Peterson dredger. The rhizosphere sediment was collected from the sediment adhering to the root crowns, where rooting was so dense that all sediment was determined to be under the influence of roots. Sediment was collected by removing a randomly selected plant and associated root crowns to a depth of 20 cm, lightly shaking the plant to remove sediment not associated with roots and then collecting soil attached to roots (Zhou and Fong 2021), All of the sediments from each site were screened, mixed, and packed in polyethylene bags and transported to the laboratory on ice. These samples were collected in the wild environment of Wanzhuang Village, and we obtained permission from the director of the village for our collection. In total, we collected 24 sediment samples [2 individuals × 3 plots × 2 sample sources (water/sediment) × 2 area types (PL/CK)]. The sediments were sieved through a 1.0-mm sieve and stored at –80 °C for further molecular analysis.

Fig. 1
figure 1

The green circles are the sampling sites of the P. lucens Linn., and the white circles are the sampling sites of CK

Full size image

Chemical characteristics

For chemical characterization, water samples were filtered through a 0.22-μm microporous filtering film, and sediment samples were air-dried at room temperature and sieved through a 1-mm screen. The pH was determined using a glass combination electrode (Li et al. 2013). The total nitrogen (TN) was determined according to potassium persulfate oxidation-UV spectrophotometry. KCl-extractable NO3− and NH4+ were determined by extraction with 2 M KCl, steam distillation, and titration (Mulvaney 1996). The organic matter (OM) and total potassium (TK) were determined by Nanjing Agricultural University. The total phosphorus (TP) was determined using the perchloric acid-sulfuric acid method (Hedley and Stewart 1982). The content of PO43− in water was analyzed by resin extraction following a protocol modified from Hedley and Stewart (1982).

Total community DNA extraction

Total DNA was extracted from 0.25 g of sediment or microporous filtering film using the Power Soil DNA Isolation Kit (MOBIO Laboratories Inc., Carlsbad, CA, USA) according to the manufacturer’s instructions. The DNA concentration and purity (A260/A280) of the extracts were estimated using a NanoDrop 2000/2000c spectrophotometer. High-quality DNA was stored at –80 °C for subsequent experiments.

Quantitative PCR (qPCR) analysis

The abundance of the bacterial 16S rRNA gene was quantified using a CFX96™ real-time PCR detection system (Bio-Rad, Hercules, CA). The reaction mixture (20 μL) contained FastFire qPCR PreMix (SYBR Green) (Vazyme, China), 10 nM of each primer, ROX Passive Reference Dye, and 1 μL of DNA. Bacterial assays used the primers 515FmodF and 806RmodR (Zhou and Fong 2021) and the following thermal program: 95 °C for 1 min followed by 40 cycles of 95 °C for 10 s and 60 °C for 30 s (Lauber et al. 2013). The standard for measuring the quantity of the 16S rRNA gene was developed from a clone with the correct insert. Plasmid DNA was prepared from the clone using a FastPure Plasmid Mini kit (Vazyme, Nanjing, China).The R2 of the standard curve was > 0.99. The qPCRs were run in quadruplicate with the DNA extracted from each sample.

Pyrosequencing and bioinformatics processing

The primers 515FmodF and 806RmodR (Zhou and Fong 2021) were used to amplify the V4 hypervariable region of the bacterial 16S rRNA gene. The PCR products were sequenced on the Illumina MiSeq PE 300 platform of Majorbio Pharm Technology Co., Ltd. (Shanghai, China). The obtained sequences were submitted to the NCBI Sequence Read Archive under the accession number PRJNA716102.

Paired-end reads were processed using Quantitative Insights into Microbial Ecology (QIIME) software, and presumptive chimeric sequences were screened and discarded using UCHIME (Zhou and Fong 2021). The original sequence data were separated, and the primers were removed (Martin 2011). According to the reading quality profile, the forward reading of the 16S rRNA gene was truncated to 240 bp, and the reverse reading was truncated to 160 bp (Schmidt et al. 2019). All reads were filtered and trimmed using the parameters maxEE = 2 and truncQ = 2. High-quality sequences were refined and resampled according to the lowest number of reads in the sample. To minimize the possibility of retaining OTUs due to sequencing errors, we deleted OTUs if (1) there were fewer than 5 sequences in less than 3 samples in each group or (2) the total number of sequences in all samples was less than 20 using the Silva database (132nd edition; http://www.arb-silva.de). All sequences matching “chloroplast” and “mitochondria” were removed from the dataset.

Statistical analysis

The concentrations of chemical characteristics among samples were determined using one-way analysis of variance (ANOVA), and paired comparisons of treatment means were achieved by Tukey’s procedure at P < 0.05 using SPSS BASE ver. 19.1 statistical software (SPSS, Chicago, IL, USA) (Ahn et al. 2012). Redundancy analysis (RDA) was performed using CANOCO 5.0 to identify the relationships between the bacterial communities and chemical characteristics (Zhang et al. 2015). A phylogenetic tree of the genera showed significant differences between PL and CK using the neighbor-joining method in MEGA v.6.0 and displayed using iTOL (Interactive Tree Of Life, https://itol.embl.de/) (Zhou and Fong 2021).

Results

Chemical characteristics of water and sediment

The chemical characteristics of the sediment and overlying water are shown in Table 1. There was no obvious difference in NH4+, TP, or PO43− in water among the sampling sites. The water concentrations of NO3− and NO2− were significantly lower in PL than in CK. There was no obvious difference in sediment NO3−, NH4+, OM, or TP between PL and CK. The sediment concentration of TK was lower in PL (7.367 g kg−1) than in CK (12.167 g kg−1).

Table 1 Chemical characteristics in water and sediment from Potamogeton lucens Linn. and none-plant areas
Full size table

Effects of P. lucens Linn. on the abundance of the 16S rRNA gene

The abundance of the 16S rRNA gene in sediment samples ranged from 1.1 × 108 to 8.2 × 108 copies/g sediment and in water samples ranged from 3.1 × 104 to 1 × 105 copies/mL water (Fig. 2). The abundance of the 16S rRNA gene in sediment samples was significantly lower (P < 0.05) in PL than in CK, whereas the opposite trend was observed in water samples.

Fig. 2
figure 2

Relative abundances of 16S rDNA of P. lucens Linn. (different lowercase letters indicates significant differences at P < 0.05)

Full size image

Bacterial alpha diversity

A total of 2,879 OTUs (24.97% of the total 11,532) were obtained from the 24 samples. There was a mean of 46,220 classifiable sequences per sample used in the subsequent analysis, with a mean read length of 253 bp. The Good’s coverage values were in the range of 0.96–0.99 at a 97% similarity cutoff, indicating that the current numbers of sequence reads were sufficient for capturing the bacterial diversity in the samples.

The alpha-diversity indices in the four groups are shown in Table 2. There were significant differences (P < 0.05) Simpson, Chao1, Shannoneven, and Simpsoneven index measures between PL and CK. The Simpson and Chao1 indices were higher, and Shannoneven and Simpsoneven indices were lower in CK water than in water with PL. Furthermore, Ace and Chao1 indices were lower in sediment with PL than in CK sediment. There was no significant difference in the Shannon or Ace index of the bacterial community in water or sediment between PL and CK (P > 0.05).

Table 2 Bacterial α-diversity in different groups
Full size table

Phylogenetic analysis of the bacterial community in water and sediment

A total of 702 bacterial and archaeal genera were identified in the 24 samples. At the genus level, strong clustering of bacterial (Fig. 3) communities according to sample source (water/sediment) and area type (PL/CK) was revealed. Samples from sediment clustered into one branch and were divided into group I and group II; samples from water clustered into another branch and were divided into group III and group IV (Fig. 3).

Fig. 3
figure 3

Hierarchical clustering analysis of bacterial community from different sampling site in Nansi Lake at genus level

Full size image

Variation in bacterial communities between water and sediment

A total of 30 genera differentially contributing to the bacterial community in sediment between PL and CK were identified via iTOL, including 17 genera which were significantly more abundant in PL than in CK and 13 genera which showed the opposite trend (Fig. 4a, Table S1).

Fig. 4
figure 4

Bacterial genera statistically different between PLS and CKS (a) and PLW and CKW (b). Colored circles represent the relative abundance of each genus. Taxonomic dendrogram shows the inferred evolutionary relationship of the enriched microbiota of each sample. Total relative abundances of all genera and significant effects across organic and conventional managements are listed in Table S1 and Table S2. PLW, water samples in Potamogeton lucens Linn.; CKW, water samples in CK; PLS, sediment samples in Potamogeton lucens Linn.; CKS, sediment samples in CK

Full size image

A total of 29 genera differentially contributing to the bacterial community in water between PL and CK were identified via iTOL, including 26 genera which were significantly more abundant in PL than in CK (e.g., Limnohabitans, Algoriphagus, Dinghuibacter) and 3 genera (e.g., Omnitrophus, Terrimonas) which showed the opposite trend (Fig. 4b, Table S2).

Environmental factors influencing bacterial communities’ structure

The RDA results are shown in Fig. 5. The first two axes together explained 84.07% (Fig. 5a) and 84.62% (Fig. 5b) of the variance in bacterial community structure in the sediment and water samples, respectively. The distances between the PL and CK samples were large, indicating that bacterial community structure significantly differed between CK and PL in both sediment (Fig. 5a) and water (Fig. 5b). In the water samples, the diversity of the bacterial community was positively correlated with the TP (P = 0.033) content in PL; such a pattern was not observed in CK (Fig. 5b).

Fig. 5
figure 5

RDA of bacterial communities and environmental factor for individual samples. Environmental factor include TP (total phosphorus), NH4+ (concentration of NH4+), NO3− (concentration of NO3−), TN (total nitrogen), OM (organic matter), and TK (total potassium). a Sediment samples and b water samples. PLW, water samples in Potamogeton lucens Linn.; CKW, water samples in CK; PLS, sediment samples in Potamogeton lucens Linn.; CKS, sediment samples in CK

Full size image

The bacterial community in PL was affected by TK and NO3− (Fig. 5a). In CK sediment samples, OM and TK were primarily distributed in the same group of taxa in the bacterial communities (Fig. 5a).

Discussion

P. lucens Linn. decreased the concentrations of NO3 − and NO2 − in water and TK in sediment

In this study, P. lucens Linn. removed nutrients from the water, significantly decreasing the concentrations of NO3− and NO2− (P < 0.05) (Table 1). This founding is consistent with the previous finding that P. lucens Linn. can remove N from water (Huo et al. 2010). However, P. lucens Linn. did not substantially reduce N in sediment, which may have resulted from the effect of root exudation on bacterial communities, and P. lucens Linn. was in a declining phase (Yin et al. 2020; He et al. 2020). Furthermore, P. lucens Linn. did not reduce the concentration of TP in water or sediment, which may have been due to the fact that P. lucens Linn. was in a declining phase (He et al. 2020), and their yellow leaves may release more P into the water when we collected samples. Zhang et al. (2019) found that rising temperature significantly increased the growth of P. lucens Linn; we collected our samples in late autumn, so the performance of these plants was reduced. Jin et al. (2017) confirmed that the synergistic purification effect of P. maackianus and four other macrophytes was much greater than the individual uptake effects in water purification. Most likely, the slightly weaker water purification capacity observed in our study was due to the presence of only the single submerged macrophyte species at our sampling sites. Liu and Chen (2018) similarly demonstrated that single plant types show poorer purification effects than several submerged macrophytes in lake systems.

Significant correlation between 16S rRNA gene copy number and P. lucens Linn.

The significant correlation between 16S rRNA gene copy number and PL also indicated that P. lucens Linn. could have a substantial effect on the population sizes of bacteria. The concentration of NO3− in the CK water samples was decreased by 58 times compared with that in the PL water samples. These results may demonstrate that P. lucens Linn. plays an important role in N removal (i.e., NO3− and NO2−) in wetlands (Chang et al. 2006) and that N removal stimulated the growth of some bacterial taxa and subsequently increased the number of 16S rRNA gene copies (Yan et al. 2018). In sediment, the number of 16S rRNA gene copies was obviously different between PL and CK. This result is consistent with the finding that the number of bacteria was negatively correlated with the total organic acid concentration secreted from the roots of P. maackianus, a congener of P. lucens Linn. (Yin et al. 2020).

P. lucens Linn. increased bacterial alpha-diversity in water but decreased it in sediment

Our results showed that P. lucens Linn. can improve bacterial evenness (Shannoneven, Simpsoneven) and decrease bacterial richness (Chao1 index) in water (Table 2). The lower Chao1 index in PL water was likely due to the fact that sampling was conducted during the declining period (November) of this plant (He et al. 2020). A previous study indicated that P. maackianus can release organic acids (Yin et al. 2020). The amount of organic acids has been indicated to be negatively correlated with the diversity of DNA-based bacterial communities (Weisskopf et al. 2008). In contrast, higher organic acid root exudation from some plants (e.g., soybean) has been shown to increase the diversity of the microbial community (Yang et al. 2012). This inconsistency may be due to the different types and amounts of organic acids produced by different plants. In this study, P. lucens Linn. may have produced similar organic acids, leading to a locally weakly acidic environment in the rhizosphere sediment unconducive to the survival of some acid-sensitive bacteria, thus reducing the alpha-diversity of bacterial community. However, this hypothesis needs to be tested in future experiments by in situ GC–MS and related metagenomics techniques to investigate the composition of root exudates.

P. lucens Linn. changed the bacterial community composition in water and sediment

The differences between PL and CK in microbial community structure in water may be explained by the lower concentrations of NO3− and NO2− in PL water than in CK water (Table 1). These results are consistent with previous studies (He et al. 2007; Jorquera et al. 2014). In PL water, there was a higher abundance of Limnohabitans, which are aerobic anoxygenic phototrophs that can supplement their mostly heterotrophic metabolism with harvested light energy (Kasalický et al. 2018). This result may suggest that P. lucens Linn. can purify and improve the light transmittance of the surrounding water, leading to an increase in the number of such microorganisms. Han et al. (2019) reported that Chloroflexi and Bacteroidota played a dominant role when P. malaianus was in the decline period, and we found that the abundances of the phyla Chloroflexi and Bacteroidota (0.04%) were higher in PL water than in CK water (Fig. 4b). These results suggest that P. lucens Linn. may provide suitable conditions for the growth and reproduction of microorganisms in Chloroflexi and Bacteroidota. Desulfatiglans and Ignavibacterium have been shown to contribute to methane oxidation by nitrite and sulfate reduction (Jochum et al. 2018), and Sulfuritalea plays an important role in the degradation of aromatic pollutants (Sperfeld et al. 2019). The higher proportions of these three taxa in PL sediment than in CK sediment indicated that P. lucens significantly promoted carbon metabolism in its rhizosphere sediments. We found higher proportions of Cyanobacteria and Firmicutes, which are abundant under heavy metal stress (Huang et al. 2020), in CK sediment than in PL sediment.

Responses of the bacterial community to environmental conditions in water and sediment

The RDA results showed that N and P were the most important factors related to bacterial community structure. Hu et al. (2020) reported that the concentrations of NH4+ and NO3− were two important factors affecting the abundances of anammox bacteria and denitrifying bacteria, and these two microbial groups compete in many ecological environments. In the sediment samples of this study, we found that these two ions substantially affected the bacterial community, but whether these functional microorganisms are affected at the DNA and RNA levels requires experimental investigation. Yin et al. (2020) showed that the composition of the bacterial community is likely related to variation in NH4 + content and thus the rhizosphere states of aquatic plants. However, in this study, the diversity of the bacterial community in water was negatively correlated with the concentration of NH4+-N.

In this study, the sediment in the unplanted CK area had significantly higher concentrations of TP than PL sediment, and in PL sediment, the diversity of the bacterial community was positively correlated with the concentration of TP (Fig. 5a). These results were similar to those of previous studies (Chen et al. 2014; Dai et al. 2019). In Guanting Reservoir, China, TP concentration was shown to directly affect the number of phosphate-dissolving and/or phosphate-decomposing bacteria in sediment (Li et al. 2005). Furthermore, the TP concentration was found to shape the variation in bacterial community composition within two different drainage areas (Lindström and Bergström 2005). Therefore, the removal of TP from the rhizosphere by P. lucens Linn. strongly affected microorganisms involved in phosphorus metabolism and thus the entire bacterial community.

Conclusion

In this study, we showed that P. lucens Linn. has strong effects on the chemical characteristics and bacterial communities in water and rhizosphere sediment in Nansi Lake, China. P. lucens Linn. can alter the environment by affecting the quality of water, which affects the composition of the bacterial community. The results of this study clarify the effect of P. lucens Linn. in lake ecosystems, especially in structuring the composition of the bacterial community. An optimal purification effect for sewage treatment may be achieved with two or more plant types. However, the relationship between submerged macrophytes and the bacterial community require further study.