Introduction

Eutrophication was caused by high concentration of nutrients such as nitrogen and phosphorus. Phosphorus was usually the limiting factor of lake eutrophication1. With exogenous phosphorus properly controlled, phosphorus released from lake sediments became the main source of water eutrophication2. Studies on the release of endogeneous phosphorus from sediments have primarily focused on the impact of environmental physical and chemical factors on morphological transformation and release amount3,4,5,6. However, there were few studies on microbial taxa involved in phosphorus release. Phosphorus in sediments would be released into the overlying water through physical, chemical and biological interactions between sediment and overlying water7. Microorganisms played a crucial role in the phosphorus cycle at the sediment–water interface and the endogenous phosphorus release from sediments8. According to the method for phosphorus fraction in sediment developed under the frame of Standard Measurements and Testing (SMT) Program of the European Commission, the phosphorus forms in the sediments were classified into total phosphorus (TP), organic phosphorus (OP), inorganic phosphorus (IP), NaOH-P and HCl-P. Among them, OP was an important part of the phosphorus balance in the lake ecosystem, and also an important phosphorus source released from the sediments of eutrophic lakes. OP in sediments was mainly released through mineralization and decomposition of Alkaline phosphatase (ALP) secreted by microorganisms. ALP is responsible for dephosphorylation of organic phosphorus compounds by breaking the phosphodiester linkage to remove phosphate groups, and thereby forming orthophosphates. Algae utilized orthophosphates and proliferated massively, which resulted in water eutrophication. ALP activity and stability were mainly affected by pH, T, DO, PO43− and other environmental factors9.

ALP is an important enzyme that plays an important role in organic phosphorous hydrolysis10. Three genes mainly encode for ALP, namely, phoA, phoX, and phoD11. It is generally believed that the genetic diversity and abundance of the phoD gene are higher than those of the phoA and phoX genes in the soil and marine ecosystems12,13. An increasing number of researchers have been devoted to the study of the phoD gene, which can be used as a molecular marker. High-throughput sequencing is usually employed to study the phoD gene diversity, abundance and distribution in soil ecosystems14,15. On this basis, we can further discuss the influence of fertilization, soil pH, salinity, and other factors on phoD gene abundance and microbial community structure13,16,17. These results indicated that the expression of phoD gene was restricted by phosphorus content and salinity stress. Fertilization would increase the phoD gene abundance and change the phoD-harboring bacterial community structure. Relevant researches have also shown that the phoD gene exists abundantly in the ocean surface and plays an important role in the degradation of organic phosphorus in oceans12,18.

In the past, phoD gene in soil and marine ecosystems was reported in most studies, but phoD gene in the freshwater lake ecosystems was seldom studied. The studies of suspended particles phoD alkaline phosphatase gene diversity, effects of cyanobacterial growth and decline on the phoD-harboring bacterial community structure, and changes of phoD gene community in sediments in different seasons have been conducted in the shallow freshwater lake systems19,20,21. These results indicated that the diversity and abundance of phoD genes were high in the freshwater lake ecosystems. PH, T, DO, SRP, OP, TP and TN/TP were the main environmental factors affecting the diversity and abundance of phoD genes2. ALP played an important role in the process of eutrophication of water bodies.

No researches have been conducted on the diversity of phoD gene and its relationship with eutrophication in the sediments from sub-deep freshwater lake (between deep and shallow lake) systems. Investigating the diversity of the phoD gene in the sediments of eutrophic lakes and the microbial community structure based on the phoD gene can help identify the microbial communities that have an important influence on phosphorus release. Microbial mineralization of OP is an important component of phosphorus cycle in sub-deep freshwater lake ecosystems. The research on phoD gene is helpful to reveal microbial driving mechanism of OP mineralization in eutrophic sub-deep freshwater lake area. Besides, phoD-harboring microbial communities may make varying contributions to phosphate solubilization under different environmental conditions. For these reasons, it is highly necessary to study the intrinsic relationships between phoD-harboring microbial communities and environmental factors. Sancha Lake is a sub-deep freshwater lake featured by endogenous pollution and eutrophication22. The release of endogenous phosphorus from the sediments of Sancha Lake is the main cause of water eutrophication23. We have focused on the response of phoD-harboring microbial communities in the sediments of Sancha Lake. Here, we hypothesized a high diversity of the phoD gene encoding ALP in the sediments of Sancha Lake and that the phoD gene abundance and the phoD-harboring microbial communities vary seasonally and spatially as a response to environmental factors. In addition, phoD-harboring microbial communities make important contributions to soluble reactive phosphorus (SRP).

We performed total DNA extraction and qPCR and high-throughput sequencing of the bacterial phoD gene in the sediments of Sancha Lake. On this basis, we analyzed the correlations between the diversity and spatial–temporal distribution of the phoD gene in sediments vs. dissolved oxygen (DO), temperature (T), pH, total organic carbon (TOC), total nitrogen (TN), and phosphorus.

Materials and method

Description of the sampling sites and sediment collection

Sancha Lake is situated in Sichuan Tianfu New Area (E 104°11′16″ ~ E 104°17′16ʺ, N 30°13′08″ ~ N 30°19′56″), in the eastern suburb of Chengdu and upstream of Jiangxi River, a tributary of Tuojiang River belonging to the Yangtze River System. The average water depth of Sancha Lake is 8.3 m and the maximum water depth 32.5 m. The study area has a subtropical humid monsoon climate, with an average temperature of 15.2–16.9 °C and an average precipitation of 786.5 mm. The catchment area above the dam site is 161.25 km2. The water source of Sancha Lake was mainly from the Min River, accounting for about 80% of the total water volume of the reservoir. The rest came from rainfall and two creeks. Spring was the period of irrigation and drainage, and autumn was the period of water diversion. Sancha Lake is not only an important water source area in Tianfu New Area, but also performs various functions, including maintaining biodiversity, storing water for agricultural irrigation, and regulating surface runoff and local climate. CODCr and BOD5 in the inflow water have been decreasing according to the water quality monitoring over the years. By contrast, TP in the lake water has been rising. Correspondingly, Chla also increases, while transparency has been decreasing over the years, resulting in an increasing severity of eutrophication24. Based on the features of sediment distribution and eutrophication status in Sancha Lake, 9 sampling points were selected, as shown in Fig. 125. The latitude and longitude of the sampling points were determined using a GPS unit. The sampling points L1 and L4 were located where concentrated area of fenced aquaculture used to be intensively carried out. The sampling points L2 and L3 were located in the tailwater area. The sampling point L9 was close to regions with intense human activities. The sampling points L5, L8 and L6 (dam) were closed to where cage culture used to be highly intense. The sampling point L7 was located in the main inflow water area. At the nine chosen sampling points as shown in Fig. 1, the surface sediments (0–10 cm) from the lake bottom were collected respectively with a Peterson grab and then sealed in polyethylene bags in April 15, 2017 (Spring) and November 3, 2017 (Autumn). Three samples were collected for each site and mixed together as the representative sample for this site. The samples were placed in polyethylene bags and sealed. They were frozen in ice and immediately taken back to the laboratory. One part of the sample was stored in the fridge at 4 °C for physicochemical analysis (within 24 h). The remaining part was stored in the fridge at − 80 °C for DNA extraction. Besides, at each sampling point, an airtight water sampler was used to collect water overlying the sediments for analyzing water environmental indicators.

Figure 1
figure 1

Sampling sites in the Sancha Lake.

Full size image

Determination of physical and chemical factors in the sediments and overlying water

The sequential P extraction tests in sediments were conducted by SMT26,27. Sequential P extraction in sediments by SMT Method includes TP, OP, NaOH-P and HCl-P in sediments. SMT method has three independent steps: 3.5 mol L−1 HCl extracts TP, 1 mol L−1 HCl extracts IP and OP, and 1 mol L−1 NaOH extracts NaOH-P and HCl-P. TP, OP, NaOH-P and HCl-P in sediments were determined using the ammonium molybdate spectrophotometric method28. TOC in sediments was determined according to the standard GB/T 19145-2003 released by the General Administration of Quality Supervision, Inspection and Quarantine of the People's Republic of China. TN in sediments was determined using alkaline potassium persulfate digestion UV spectrophotometric method25. The pH values of freshly collected sediments were determined using the glass electrode method after extraction with the ratio of sediments to distilled water being 1:5. The overlying water samples were passed through a 0.45 μm membrane filter. The filtrate was digested with potassium sulfate, and dissolved total phosphorus (DTP) was determined using ammonium molybdate spectrophotometric method. The overlying water samples were passed through a 0.45 μm membrane filter. The filtrate was measured for SRP using ammonium molybdate spectrophotometric method. Dissolved organic phosphorus (DOP) was DTP minus SRP29. T was measured with the HI991301 portable multiparameter meter; DO was measured with the HQ30D Portable Dissolved Oxygen Meter.

Sediment ALP activity assay

The ALP activity was measured using Anupama's method30, specifically as follows: About 0.5 g of freshly collected sediments was weighed and placed in the disinfected reaction tube. Into the tube 10 mL of 0.5 mol L−1 Tris–HCl buffer with a pH value of 8.6 was added and mixed well by shaking, followed by ultrasonic treatment for 45 s. Into the mixture 1 mL of 10 mmol L−1 p-nitrophenyl Phosphate was added as substrate, mixed well and placed in the 37 °C water bath for 1 h. The reaction was terminated by adding 2 m of 1 mmol L−1 NaOH solution. The solution was then subjected to refrigerated centrifugation at 5000 rpm at 10 min, and the supernatant was collected. The absorbance was measured using a spectrophotometer at 410 nm and converted to activity intensity per gram (dry weight) of sediments. A standard curve was plotted for varying contents of p-nitrophenol (PNP). Each sample had three replicates. ALP activity was expressed as μmol PNP produced by each gram (dry weight) of sediments per hour.

DNA extraction, PCR and Illumina Miseq sequencing

Total DNA was extracted from the samples according to the instruction provided with the E.Z.N.A.® soil kit (Omega Bio-tek, Norcross, GA, US). DNA concentration and purity were determined using a Nanodrop 2000 spectrophotometer. The DNA quality was evaluated by 1% agarose gel electrophoresis.

The phoD gene abundance was quantified using the TIB8600 fluorescence qPCR analyzer (Triplex International Biosciences (China) Co., Ltd.). The primers used to amplify the phoD gene were ALPS-F730 (5′-CAGTGGGACGACCACGAGG T-3′) and ALPS-R1101 (5′-GAGGCCGATCGGCATGTCG-3′)17,31,32. Ten-fold dilution was performed for all DNA samples. About 10 ng of DNA was used as template for each qPCR reaction. Each sample had three replicates, and a negative control was set up. The qPCR reaction system (20 μL) consisted of the following: 9 μL 2 × SYBR real-time PCR premixture (BioTeke Corporation, Beijing, China), 0.5 μL upstream and downstream primers (10 µM), and 10 μL of diluted DNA template. The concentration of different mixed reagents in the qPCR reaction system is initial. qPCR consisted of the following steps: Predenaturation at 95 °C for 5 min, 40 cycles (denaturation at 95 °C for 15 s, annealing at 60 °C /55 °C for 30 s, extension at 72 °C for 30 s); final extension at 72 °C for 1 min. The plasmid standard harboring the target gene was constructed with the pMD18-T Vector (TaKaRa, Japan) and sequenced for verification. Standard curves were generated using tenfold serial dilutions of the plasmids. The range of phoD gene copy numbers is 6.87 × 102 ~ 6.87 × 108 from the different standard curves. The number of phoD gene copies was calculated by measuring the concentration of the plasmid and the number of base pairs. Amplification efficiencies ranged from 94 to 98%, and R2 value of 0.9932 for phoD gene.

The communities of phoD-harboring bacteria were assessed using the Illumina Miseq 300 bp paired-end sequencing platform. The variable barcodes were designed and respectively linked to the 5' end of primer pairs (ALPS-F730/ALPS-S1101) to allow distinguishing of the sequences of each sample. DNA thus obtained was used as template for phoD gene amplification on ABI Applied Biosystem 2720 Thermal Cycler (Applied Biosystems, Foster City, CA, USA). Amplification of each sample was done in triplicate, with 20 μL reactions under the same conditions as the qPCR experiment. The PCR amplification system consisted of the following (20 µL): 2 μL Ex Taq buffer (10 ×, 0.2 μL Ex Taq (5 U·μL−1), 1.6 μL dNTP Mix (2.5 mM), 1.0 μL Forward primer (5 µM), 1.0 μL Forward primer (5 µM), 0.5 μL template DNA, 13.7 μL ddH2O. The concentration of different mixed reagents in the PCR amplification system is initial. The PCR product was recovered using the 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA). After assessing quantity and quality of the purified PCR products using a QuantiFluor TM-ST (Promega, USA), the three purified PCR products per sample were pooled as one PCR amplicon. Finally, the PCR amplicons were pooled in an equimolar con-centration for Illumina Miseq sequencing at Personal Biotechnology Co., Ltd. (Shanghai, China). All sequences were deposited in the NCBI Sequence Read Archive (SRA) database (BioProject number PRJNA623151).

Bioinformatics of Illumina Miseq sequencing

Quality control of raw sequences was implemented by Trimmomatic. The sequences were merged using the FLASH software (version 2.7, http://ccb.jhu.edu/software/FLASH/, Center for Bioinformatics and Computational Biology, Iowa City, IA, USA). The sequences were clustered into OTUs at the 80% similarity threshold using the UPARSE software (version 7.1, http://drive5.com/uparse/, Edgar, R.C., Tiburon, CA, USA)30. Simple sequences and chimeras were excluded during the clustering process. OTUs with abundance values lower than 0.005% of the total reads of all samples were removed33.

RDP classifier (http://rdp.cme.msu.edu/) was used for microbial species classification and annotation for the most abundant sequence within each OTU. phoD gene sequences were aligned with the NCBI database (http://www.ncbi.nlm.nih.gov/), with the threshold set to 70%. Statistics were done and plots were drawn using the R-package based on the taxonomic information obtained for each OTU. Representative sequences for the top 15 most abundant OTUs were chosen. Next, the phylogenetic tree was plotted based on the representative sequences for each OTU using the maximum likelihood method and FastTree software (version 2.1.3 http://www.microbesonline.org/fasttree/)34.

Based on the OTU clustering results, QIIME software (version 2.0, http://qiime.org/, Rob Knight Lab, Boulder, CO, USA) was implemented to estimate the abundance-based coverage estimators (ACE), Chao1 index, Simpson index, and Shannon–Wiener index for each sample. We further examined the similarity of the structure of phoD-harboring bacterial communities across the samples. Principal component analysis (PCA) was conducted for microbial community structure data using QIIME35. The samples were ranked, and the diversity and spatial and temporal heterogeneity in the composition of phoD-harboring bacterial communities in sediment samples were observed. The phoD gene sequences were deposited in the NCBI Sequence Read Archive (SRA) database under accession number PRJNA623151.

Statistical analysis

Variance inflation factor (VIF) was used to screen for known environmental factors that had little interaction with each other29. Adonis test and Metastats were implemented using the Vegan package in R to assess the significance of intragroup and intergroup differences36. Pearson's correlation coefficients were calculated in SPSS (version 20.0, IBM, Armonk, NY, USA) to estimate the intensity of correlations between phoD gene abundance, microbial biodiversity, and environmental factors. Vegan package in R was implemented to conduct redundancy analysis (RDA) between the spatial distribution features of phoD-harboring bacterial communities in sediments and environmental factors37. The significance of differences between the samples was estimated by analysis of variance (ANOVA). The significance level was set to p = 0.05, and the level of extreme significance was p = 0.01.

Result and analysis

Physical and chemical properties of the sediments and overlying water

The results of physicochemical factor determination in sediments and overlying water in spring and autumn are shown in Table 1. Since the Sancha Lake is a stratified reservoir, the temperature in the water overlying the sediments ranged from 11.6 ~ 17.0 °C, and did not vary significantly in spring and autumn. The value of DO ranged from 4.20 to 10.87 mg L−1, higher in spring than in autumn for all locations. Besides, DO was higher at the sampling point L7 in the inflow water area and at the sampling points L2 and L3 in the shallow water. The sampling point L6 in the deep water section of the dam was the lowest. At this sampling point, the value of SRP ranged from 0.006 to 0.086 mg L−1, higher in spring than in autumn and was higher in the inflow water area and in the arm of lake. By contrast, SRP was lower in the lake center. The value of DOP ranged from 0.001 to 0.008 mg L−1, lower in spring than in autumn and was higher in the tail and arm of lake than in other parts of the lake. The pH value of sediments ranged from 6.13 to 8.32, and did not vary significantly at different sampling points. Nevertheless, it was much higher in autumn than in spring. The TP value in sediments ranged from 0.27 to 2.46 mg g−1, higher in autumn than in spring, and it varied significantly across the sampling points in the lake area. TP was the highest where cage culture used to be intensive (L5, L6 and L8). It was the second highest where concentrated area of fenced aquaculture used to be intensive (L1 and L4), followed by the value in the region where human activities used to be intensive (L9). TP was the lowest in the inflow water area (L7) and the tailwater area (L2 and L3). These results might be explained by the practice of dropping in fish food for cage culture and the associated human activities. As indicated by TP, the water was moderately polluted according to US EPA's standard. The OP value of sediments ranged from 0.13 to 0.77 mg g−1. The TOC value of sediments ranged from 13.8 to 28.9 mg g−1. The seasonal variation and spatial distribution features of OP and TOC were consistent with those of TP. NaOH-P, HCl-P and TN did not vary significantly across the sampling points. However, these indicators were slightly higher in autumn than in spring. The ALP activity of sediments ranged from 0.22 to 2.42 μmol g−1 h−1, significantly higher in autumn than in spring. It was also higher in the tailwater area and the arm of lake where cage culture used to be intensive, but lower in the inflow area and deep water area in the dam.

Table 1 Physicochemical properties of the sediments and overlying water in spring and autumn.
Full size table

Genetic diversity and phylogenetic tree based on the phoD gene

Total DNA was extracted from the microorganisms in sediment samples, and phoD gene sequences were acquired by PCR amplification. A total of 881,717 valid sequences were obtained by high-throughput sequencing, with an average length of 362 bp. The number of sequences varied between 6364 and 99,463 at different sampling sites, the average number of sequences being 48,984. The rarefaction curve of all sediment samples tended to be flat, indicating that the sequencing data volume was reasonable at each site (Supplementary Fig. S1). The samples were clustered into 477 OTUs at the 80% similarity threshold30. For spring and autumn, the number of shared OTUs was 96; the number of OTUs unique to these two seasons was 238 and 143, respectively. The number of OTUs and the number of valid sequences were larger in autumn than in spring. The number of OTUs for each sample varied between 24 and 152. The distributions of the number of OTUs and the number of valid sequences across the samples are shown in Table 2.

Table 2 The richness and diversity of phoD genes in sediments of the Sancha Lake.
Full size table

Using the statistics of OTUs, we calculated Shannon's index and Simpson's index to represent genetic diversity based on the phoD gene and ACE index and Chaol index to represent abundance, as shown in Table 2. The genetic diversity of bacterial communities in the sediments of Sancha Lake was high based on the phoD gene. The Shannon’s index varied between 1.95 and 5.96, and the Simpson’s index between 0.4837 and 0.9685. As for abundance, the Chaol index varied between 57.11 and 160.33, and the ACE index between 58.69 and 167.38. The Shannon's index and the Simpson's index were slightly higher in autumn than in spring (P > 0.05). The Chaol index and the ACE index were significantly higher in autumn than in spring (P < 0.05). These results indicated that the phoD-harboring bacterial communities in autumn and spring had varying levels of diversity and abundance. In autumn, Shannon's index and Simpson's index were the lowest at the center of the lake and increased in the arm of the lake. The genetic diversity of bacterial communities increased significantly in the tail of the lake.

A phylogenetic tree was drawn based on representative phoD gene sequences for the top 15 OTUs with the highest relative abundance. This phylogenetic tree was composed of three branches (Fig. 2). These sequences were primarily aligned with genera Pseudomonas and Cupriavidus belonging to the phylum Proteobacteria, genus Streptomyces belonging to the phylum Actinobacteria, and genus Paludisphaer belonging to the phylum Planctomycetes. Genera Pseudomonas and Streptomyces were the most observed in all treatments and accounted for 37.97% of all sequences.

Figure 2
figure 2

Phylogenetic tree of PhoD gene sequences (15 OTUs) from sediments of the Sancha Lake.

Full size image

phoD-harboring bacterial community composition and structure

In 18 sediment samples, 41 genera belonging to 31 families, 23 orders, 12 classes, and 9 phyla with definite taxonomic information were identified in the phoD-harboring bacterial communities. On the phylum level, the distributions and relative abundances of more abundant phoD-harboring bacterial communities in the sediments of Sancha Lake in spring and autumn are shown in Fig. 3. Phyla with higher relative abundances (with an average abundance > 1%) were Proteobacteria (with an average abundance of 61.9% and a relative abundance of 39.4–98.6%), Actinobacteria (24.9%, 0.3–60.1%), and Planctomycetes (2.9%, 0.1–13.0%). Among the above, Proteobacteria and Actinobacteria were integral components of phoD-harboring bacterial communities. The relative abundance of Proteobacteria in spring was significantly higher than that in autumn (P < 0.05). Besides, the relative abundances of Actinobacteria and Planctomycetes in spring were significantly lower than that in autumn (P < 0.05).

Figure 3
figure 3

Relative abundance and composition of phoD-harboring bacterial phyla detected in the sediments of Sancha Lake in the spring and autumn. SPS Spring Location, AUS Autumn Location.

Full size image

On the genus level, the distributions and relative abundances of more abundant phoD-harboring bacterial communities in the sediments of Sancha Lake in spring and autumn are shown in Fig. 4. Genera with higher relative abundances (average abundance > 1%) were Pseudomonas (30.5%, 1.4–77.4%), Streptomyces (23.1%, 0.3–60.1%), Phaeobacter (11.8%, 0.0–58.8%), Cupriavidus (4.3%, 0.2–38.8%), Lysobacter (4.1%, 0.0–63.2%), Paludisphaera (2.8%, 0.1–10.8%), Bradyrhizobium (2.5%, 0.6–4.7%), Phenylobacterium (2.2%, 0.0–29.2%), Pleomorphomonas (1.6%, 0.0–9.2%), Collimonas (1.5%, 0.0–8.0%), and Actinoplanes (1.3%, 0.0–16.4%). Genera Pseudomonas, Streptomyces, Paludisphaera, Cupriavidus, and Bradyrhizobium were shared genera at different sampling points in autumn and spring. The genus Pseudomonas had the highest relative abundance, accounting for 49.3% of the phylum Proteobacteria. Bacteria belonging to this genus were important components of the phoD-harboring bacterial communities in the sediments of Sancha Lake. The metastats test showed that the relative abundances of genera Pseudomonas, Streptomyces and Paludisphaera were significantly higher in autumn than in spring (P < 0.05). By contrast, the relative abundances of the genera Cupriavidus and Phaeobacter were significantly higher in spring than in autumn (P < 0.05).

Figure 4
figure 4

Relative abundance and composition of phoD-harboring bacterial genera detected in the sediments of Sancha Lake in the spring and autumn. SPS Spring Location, AUS Autumn Location.

Full size image

The PCA analysis results of the phoD-harboring bacterial community structure on the genus level are shown in Fig. 5. PC1 on the x-axis (52.2%) represented the first coordinate axis, along which the samples were differentiated from each other to the maximal degree. PC1 explained 52.2% of all differences in the samples. PC2 on the Y-axis (17.37%) explained 17.37% of all differences in the samples. These two principal components defined a plane, on which over 69.57% of all differences in the samples were shown. The phoD-harboring bacterial communities at 9 sampling points in autumn were scatteredly distributed in the first and fourth quadrants. The phoD-harboring bacterial communities at 9 sampling points in autumn were concentratedly distributed in the second and third quadrants. On the first coordinate axis with the maximal explained variance ratio, the sampling points represented by circles were effectively differentiated from those by squares. In other words, the intersample differences were primarily attributed to the seasonal factor (spring and autumn) (P < 0.01). Besides, samples from different sites in the same season could be well differentiated along the second coordinate axis. These results indicated that the sampling site had an important impact on the phoD-harboring bacterial community structure. The adonis test showed that PCA could differentiate the phoD-harboring bacterial communities in autumn and spring with extreme significance (R2 = 0.581, P = 0.001).

Figure 5
figure 5

Principal component analysis of phoD-harboring bacterial genera detected in the sediments of Sancha Lake in the spring and autumn. SPS Spring Location, AUS Autumn Location.

Full size image

Quantification of phoD gene abundance

The results of fluorescence quantitative PCR are shown in Fig. 6. The copy number of the phoD gene varied between 5.039 × 105 and 2.275 × 107 copies/g (dry sediment) across the sampling points in autumn and spring, the average copy number being 5.065 × 106 copies/g (dry sediment). Significant spatial heterogeneity and seasonal changes were observed in phoD gene abundance in sediments of Sancha Lake (0.01 or P < 0.05). The phoD gene abundances at different sampling points were significantly higher in autumn than in spring (P < 0.05). In spring, the phoD gene abundance was higher in the tail of lake and where cage culture used to be intensive. By contrast, the phoD gene abundances were lower where concentrated area of fenced aquaculture used to be intensive and near the regions with intense human activities. It was the lowest in the inflow water area and deep water area of the dam. The sampling points were ranked in a similar order in terms of phoD gene abundance in autumn and spring. However, the phoD gene abundance fluctuated less significantly in autumn than in spring.

Figure 6
figure 6

Abundance of phoD gene in the sediments of Sancha Lake in the spring and autumn. At the same sampling time, Bars with different letters (a, b, c, d, e, f, g, h, i) at each location point are significantly different at P < 0.05 according to one-way ANOVA. g is dry sediment.

Full size image

Correlation between the distribution of phoD-harboring bacterial communities in sediments of Sancha Lake and environmental factors

The correlation results between environmental factors vs. diversity and abundance indices and abundance of the bacterial communities based on the phoD gene are shown in Table 3. The phoD gene abundance had significant or highly significant positive correlations with the pH value, OP and ALP (P < 0.05 or P < 0.01) but significant or highly significant negative correlations with SRP and DO (P < 0.05 or P < 0.01). The number of OTUs clustered based on the phoD gene had significant or highly significant positive correlations with T, DOP and TOC (P < 0.05 or P < 0.01), but highly significant negative correlations with OP (P < 0.01). The Shannon's index based on the phoD gene had significant positive correlations with the pH value and DOP (P < 0.05), but significant negative correlations with DO, TP and OP (P < 0.05). The Chaol index based on the phoD gene had a significant positive correlation with ALP (P < 0.05). Overall, the diversity and abundance indices and abundance of the bacterial communities based on the phoD gene were closely related to the physicochemical factors.

Table 3 Statistical analysis of diversities and abundances of phoD gene with physicochemical parameters.
Full size table

The results of a linear regression analysis showed that the phoD gene copy number had an highly significant positive correlation with the ALP activity (R2 = 0.705, y = 0.008X + 0.528, p = 0.000). The phoD gene abundance had a significant negative correlation with SRP (R2 = 0.510, y = − 1644.065X + 93.722, p = 0.004). The ALP activity had a significant negative correlation with SRP (R2 = 0.631, y = − 16.788X + 1.366, p = 0.001).

The results of the RDA analysis on the genus level of phoD-harboring bacteria are shown in Fig. 7. Dynamic changes in the phoD-harboring bacterial communities in sediments of Sancha Lake were influenced by physical and chemical factors in the water bodies and sediments. In autumn and spring, DO, DOP and SRP in overlying water and pH value, TP, OP, ALP and TOC of sediments were environmental factors that had an important impact on the phoD-harboring bacterial community structure. The two axes offered an explanatory power of 15.19% and 10.73%, respectively. All samples in autumn fell within the first and fourth quadrants in a concentrated manner; the samples in spring fell within the second and third quadrants in a scattered manner. Thus, the samples in spring and autumn were effectively differentiated from each other (P < 0.05).

Figure 7
figure 7

The RDA analysis of phoD-harboring bacterial genera and physico-chemical factors in the sediments of Sancha Lake in the spring and autumn. SPS Spring Location, AUS Autumn Location.

Full size image

Discussion

phoD-harboring bacterial diversity and community structure

Shifts in the phoD-harboring bacteria community were reported recently in many soils and marine. Little is known, however, about phoD-harboring bacterial diversity and community structure in eutrophicated sub-deep freshwater lakes. We reported a high diversity of the phoD gene in the sediments of Sancha Lake, an eutrophic sub-deep freshwater lake. On the phylum level, Proteobacteria, Actinobacteria and Planctomycetes were integral components of the phoD-harboring bacterial communities in the sediments of Sancha Lake. On the genus level, Pseudomonas, Streptomyces, Cupriavidus, and Collimonas were dominant bacterial genera in phoD-harboring bacterial communities.

Genes encoding functional enzymes have been extensively used as molecular markers to study the diversity, structure and functional activity of microbial communities in ecological science38. Bacteria are important mineralizers of organic phosphorus39, secreting ALP to mineralize organic phosphorus into orthophosphates that are involved in chemical cycling of phosphorus and maintaining the eutrophication status of lakes40. So far, some studies have been conducted using the phoD gene as a molecular marker to characterize the genetic diversity and structure of ALP-secreting bacteria in freshwater ecosystems. Zhao et al.41 employed RFLP to analyze the diversity of bacterial phoD gene in the water bodies of the Pearl River. They reported a high diversity of bacterial phoD gene in the Pearl River. Limnohabitans, Plesiocystis and Pirellula are important bacterial genera in the phoD-harboring bacterial community. ALP encoded by the phoD gene may play an important role in eutrophication of the Pearl River. Valdespino-Castillo et al.42 combined high-throughput sequencing with clone library to detect the composition of the phoD-harboring bacterial community in shallow Lake Alchichica, Mexico. It was found that Actinobacteria, Alphaproteobacteria, Betaproteobacteria and Gammaproteobacteria were dominant bacterial classes. Zhang et al.19 applied fluorescence quantitative PCR and high-throughput sequencing to study the spatial and temporal distribution features of bacterial phoD gene in suspended particles in Taihu Lake. It was found that the dominant phoD-harboring phylum in all samples was Actinobacteria, followed by Proteobacteria, Cyanobacteria and Gemmatimonadetes. Sun21 analyzed the composition and genetic diversity of the phoD-harboring bacterial community in the sediments of Lake Chaohu. A high genetic diversity of the phoD-harboring bacterial community in sediments was noted, with Actinobacteria, Proteobacteria and Planctomycetes being dominant phyla and Pseudonocardia and Friedmanniella being dominant genera. The dominant bacterial phyla and genera in the phoD-harboring bacterial community in sediments of Sancha Lake were different from those in the Pearl Riverand and in Lake Alchichica, Mexico41,42, but similar to those in the sediment of Lake Chaohu and in suspended particles in Taihu Lake19,21.

Relationship between phoD-harboring bacterial diversity, abundance and community structure vs. environmental factors

Our study indicated significant correlations between the phoD-harboring bacterial community in sediments of Sancha Lake and environmental factors. The phoD-harboring bacterial community structure was altered as a response to changes in DO, pH value, TOC and phosphorus content, all of which were environmental factors. The above finding agreed with other published studies. The TN/TP ratio was reported to be a primary environmental factor influencing the phoD-harboring bacterial community structure and gene copy number in Lake Alchichica, Mexico42. DO, TN and TP were primary environmental factors influencing phoD gene abundance of suspended particles in shallow Lake Taihu19. The structure and genetic diversity of the phoD-harboring bacterial community structure in the sediments of Lake Chaohu were influenced by pH value, DO, T, and TP content21. Although the environmental factors influencing the phoD-harboring bacterial community vary across the habitats, the phosphorus content seems to have a universal impact on the phoD-harboring bacterial community in freshwater ecosystems.

Response of functional microbes in sediments to available phosphorus in eutrophicated water bodies

The phosphorus content in sediments is closely related to that in overlying water. This is because the overlying water comes into contact with bottom sediments. There are usually material and energy exchanges between the overlying water and bottom sediments due to biological actions43. In East Lake, phosphorus is released from sediments due to microbial activities under a low phosphorus concentration in overlying water. The amount of phosphorus released from the sediments was negatively correlated with the phosphorus concentration in the lake water44. In Lake Tai, Wang et al.45 observed close connections between the changes in phosphorus concentration in overlying water and ALP activity in sediments. Besides, ALP was closely related to the amount of phosphorus-dissolving bacteria in sediments. We calculated Pearson's correlation coefficients and performed ANOVA in the present study. It was found that the bacterial phoD gene abundance and ALP activity in the sediments of Sancha Lake were significantly negatively correlated with SRP in overlying water (P < 0.05). However, the phoD gene abundance had an highly significant positive correlation with ALP activity (P < 0.01). Therefore, phoD-harboring bacteria were the main producers of ALP, which made important contributions to SRP in overlying water.

Algae proliferative massively in eutrophic water bodies and utilize a large amount of phosphorus during their growth, disrupting the balance of phosphorus release from sediments. As a result, the amount of phosphorus released from sediments to water increases. Organic phosphorus is the form of phosphorus first released from sediments in the presence of algae and has a higher bioavailability46. In the present study, the phoD gene abundance and ALP activity of the bacterial community in sediments of Sancha Lake were significantly higher in autumn than in spring (P < 0.05). Besides, they were closely related to organic phosphorus content in sediments. We inferred that the vigorous growth of aquatic plants in Sancha Lake in autumn significantly increased the diversity of the phoD gene in sediments. As the demand of aquatic plants for available phosphorus increased, the phoD-harboring bacterial community began to show stronger ability to decompose organic phosphorus, resulting in a greater amount of phosphorus released to overlying water. This explanation was supported by higher relative abundances of Pseudomonas and Streptomyces in sediments in autumn, as observed in the present study. Pseudomonas and Streptomyces are among the most important organic phosphorus-degrading bacteria in sediments47. The growth of algae and other plankton leads to the "pumping" of phosphorus in sediment48. With the continuous reproduction of algae in the Sancha Lake in autumn, the demand for phosphorus gradually increases, and the phosphorus consumption in the overlying water can no longer be supplemented by the independent diffusion alone. ALP mineralizes and hydrolyzes OP in sediments, releasing SRP, which is needed for Alga growth. Therefore, in Sancha Lake, the rapid decline of SRP in overlying waters due to the increasing algaeal photosynthesis in the early growth of algae induces the ALP, promotes the growth of Pseudomonas and Streptomyces and alleviates the phosphorus restriction.

Potential limitations of the methodological approach used

The inherent methodological limitation of phoD primers does to be acknowledged although it has not limited the importance of this study. The species of phoD-harboring bacteria recovered with primer sets ALPS-F730/ALPS-S1101 in this study seem to be restricted to a few bacterial phyla. Nonetheless, our result showed that a significant positive correlation between phoD gene abundance and ALP activity (R2 = 0.715, p < 0.01), indicating that the identified phoD species represented the majority of phoD populations in our sediment. Newly designed primers based on metagenome databases will probably help to detect larger phoD gene diversity in future studies49. Moreover, future research could consider the phoD expression level, which is a direct indication of sediment ALP activity. Finally, SRP release from fractions has biological, physical and chemical effects, and can be carried out under anaerobic and oxidizing conditions.

We did not investigate differences between the "biotic" and "abiotic" regulation of SRP release from sediments with oxic or anoxic bottom waters. We believes that these aspects need to be further studied in future.

Conclusions

We observed a high diversity of the phoD gene in sediments of Sancha Lake, which implies an abundance of phosphorus-dissolving bacterial resources.

The phoD-harboring bacterial community structure in sediments of Sancha Lake showed a significant difference in spring and autumn, but no apparent spatial heterogeneity. Significant spatial heterogeneity and seasonal changes were observed in phoD gene abundance.

pH value, DO, TOC, ALP and phosphorus content were important environmental factors influencing genetic diversity and structure of the phoD-harboring bacterial community in the sediments of Sancha Lake. The spatial and seasonal changes in the phoD abundance and phoD-harboring bacterial community structure represent a dynamic response to SRP in overlying water. The phoD-harboring bacterial community may play an important role in the transfer and conversion of phosphorus in sediments and overlying water in Sancha Lake, supplementing phosphorus to overlying water.