Search tips
Search criteria 


Logo of scirepAboutEditorial BoardFor AuthorsScientific Reports
Sci Rep. 2017; 7: 10927.
Published online 2017 September 7. doi:  10.1038/s41598-017-11448-x
PMCID: PMC5589776

Spatiotemporal heterogeneity of core functional bacteria and their synergetic and competitive interactions in denitrifying sulfur conversion-assisted enhanced biological phosphorus removal


Denitrifying sulfur conversion-assisted enhanced biological phosphorus removal (DS-EBPR) has recently been developed for simultaneously removing nitrogen and phosphorus from saline sewage with minimal sludge production. This novel process could potentially enable sustainable wastewater treatment. Yet, the core functional bacteria and their roles are unknown. Here, we used high-throughput 16S rRNA gene sequencing coupled with principal coordinates analysis and ANOVA with Tukey’s test to unravel the spatiotemporal heterogeneity of functional bacteria and their synergetic and competitive interactions. We did not find any obvious spatial heterogeneity within the bacterial population in different size-fractionated sludge samples, but the main functional bacteria varied significantly with operation time. Thauera was enriched (9.26~13.63%) as become the core functional genus in the DS-EBPR reactors and links denitrifying phosphorus removal to sulfide oxidation. The other two functional genera were sulfate-reducing Desulfobacter (4.31~12.85%) and nitrate-reducing and sulfide-oxidizing Thiobacillus (4.79~9.92%). These bacteria cooperated in the DS-EBPR process: Desulfobacter reduced sulfate to sulfide for utilization by Thiobacillus, while Thauera and Thiobacillus competed for nitrate and sulfide as well as Thauera and Desulfobacter competed for acetate. This study is the first to unravel the interactions among core functional bacteria in DS-EBPR, thus improving our understanding of how this removal process works.


For over 40 years, a process known as enhanced biological phosphorus removal (EBPR) has been successfully used to remove and recover phosphorus (P) in full-scale wastewater treatment plants. EBPR does not require chemical dosing for P precipitation and therefore produces less sludge than do conventional processes13. The functional bacteria in the EBPR system are referred to as polyphosphate-accumulating organisms (PAOs) and include Candidatus Accumulibacter phosphatis and Tetrasphaera-related bacteria48. PAOs compete extensively with glycogen-accumulating organisms (GAOs) in the EBPR system and are sensitive to both warm water temperatures (> 20°C)9 and high sulfate concentrations10, 11.

Wastewater contains sulfur (S) as a result of saltwater intrusion12, the application of sulfate coagulants in drinking water treatment13, brackish water supply systems14 and/or direct seawater toilet flushing15. In Hong Kong, seawater has been used for toilet flushing since 1958 with the installation of a dual water supply system to conserve fresh water resources and reduce energy consumption16. As a result, the sewage has a high ratio of sulfate-sulfur to chemical oxygen demand (COD) of ~0.4 mg S/mg COD17. This motivated a group of scientists and engineers to develop a completely new S-assisted biological wastewater treatment process called the sulfate reduction, autotrophic denitrification and nitrification integrated (SANI) process in the last decade for removing organics and nitrogen (N) with minimal sludge production1821. To alleviate the eutrophication of receiving waters due to the presence of P in sewage, five years ago the same group extended the SANI process by introducing biological S-assisted denitrifying luxury P uptake so that besides organics and N, P could also be removed. The extended process is called denitrifying S cycle-assisted enhanced biological phosphorus removal (DS-EBPR)11, 22. In a tightly sealed sequencing batch reactor (SBR), polyphosphate (poly-P) is used to form poly-β-hydroxyalkanoates (PHAs) and convert polysulfide and/or elemental sulfur (poly-S) in anaerobic conditions. Nitrate is added into the SBR as soon as the acetate is exhausted, then the PHAs and poly-S is utilized for biological luxury P uptake in a denitrifying process. The process also can work even at temperatures as high as 30 °C22.

As DS-EBPR was only recently developed, few studies have been conducted and the ones available are mainly focused on process development, performance verification and evaluation. Some of these studies have touched on the microbial community governing this process and have even speculated about its underlying mechanism11, 17, 22, 23. For instance, one study using 454-pyrosequencing detected none of the conventional PAOs (i.e. Candidatus Accumulibacter and Tetrasphaera-related bacteria) in the microbial community of a DS-EBPR sludge sample cultivated under 30 ± 1°C; instead it found that the community was dominated by non-P-removing genera (83%) with unclassified genera making up the rest (17%)22. Another study using 16S rRNA gene pyrosequencing revealed that sulfate-reducing bacteria (SRB) and sulfide-oxidizing bacteria (SOB) were enriched at the genus level and highlighted the critical role of external sulfide addition in effectively restoring the DS-EBPR system from failure17. However, none of these studies has offered insights into the key functional bacteria governing the DS-EBPR process in carbon (C), N, P and S turnovers or the interactions between S-related and P-related bacteria in this complex microbial system, not to mention an in-depth understanding of bioprocess metabolism.

The overarching goal of this study is thus to unravel the spatiotemporal heterogeneity within the core functional bacterial population in different size-fractionated sludge samples during the long-term operation of DS-EBPR reactors and shed light on the synergetic and competitive interactions among these bacteria. To this end, three lab-scale DS-EBPR reactors were set up and operated for different periods of time to characterize the microbial community in the sludge samples. The characterization was performed on the Illumina MiSeq PE300 platform using high-throughput 16S rRNA gene sequencing. The temporal variation and spatial heterogeneity of the microbial community were investigated via principal coordinates analysis (PCoA) and ANOVA with Tukey’s test. The traits of functional bacteria and possible process metabolism and interactions during C, N, P and S conversions were revealed by combining the results of reactor operation and molecular bioinformation obtained from the DS-EBPR system.

Results and Discussion

Reactor performance

The operational performance, typical cycle measurements and performance of the three lab-scale reactors on the sampling dates are summarized in Figs 1 and and22 and Table 1. The P release, P uptake and P removal (calculated as Puptake − Prelease) were measured to evaluate the EBPR performance and the results are shown in Fig. 1b,c and d respectively. The cycle time was significantly reduced from 24 and 14.5 h/cycle for reactor R0 (days 1~116 and days 117~200) to 9.5 h/cycle for R1 (days 201~400) and 7 h/cycle for R2 (days 201~400). The average P removal was determined to be 1.9 ± 0.3 mg P/(L·cycle) for R0 during days 1~116, 5.1 ± 0.2 mg P/(L·cycle) for R0 during days 117~200, 5.6 ± 1.0 mg P/(L·cycle) for R1 (days 201~400) and 4.5 ± 1.8 mg P/(L·cycle) for R2 (days 201~400). Therefore, R0 (after day 117), R1 and R2 were all able to remove P efficiently. The experimental stages and the initial concentrations of acetate, sulfate, phosphate and nitrate in the DS-EBPR SBRs are summarized in Supplementary Table S1. The representative cyclic performance of R0, R1 and R2 is shown in Fig. 2 (C and P conversions, as well as S and N conversions). The results demonstrate a clear association between S conversion and biological denitrifying P removal as reported in our previous study11, 23. Based on the experimental data in Table 1 such as the acetate uptake rate, the nitrate consumption rate, the sulfate reduction rate and the sulfide oxidation rate, the Spearman correlation was analyzed to investigate the significance of the operating conditions and reactor performance, and the results are shown in Supplementary Table S2. Phosphate removal was significantly and positively correlated with sulfide oxidation, initial acetate and nitrate (correlation coefficients > 0.7, P < 0.01). This might have been because the main P-removing bacteria could utilize acetate, nitrate and sulfide. However, phosphate removal was significantly and negatively correlated with initial sulfate (correlation coefficient > 0.7, P < 0.01). It was possible that the main P-removing bacteria could not utilize sulfate while other bacteria could and competed with the P-removing bacteria for the C source. These indicate that the microbes play a positive role in this complex ecosystem and cooperate to remove P. Hence, it was necessary to analyze the core functional microorganisms and their interactions.

Figure 1
Experimental design and operational performance of the DS-EBPR process. (a) Experimental design and sampling. (bd) P release, P uptake and P removal in each operation cycle of R0, R1 and R2.
Figure 2
Typical cycle of the DS-EBPR process. (ac) Typical cycle of R0, R1 and R2 on days 200, 400 and 400.
Table 1
Operating conditions and performance of DS-EBPR SBRs (measured average value ± standard deviation, n = 3).

Composition and spatiotemporal heterogeneity of the microbial community in DS-EBPR reactors

An inoculum sludge sample S and sludge samples S0 (taken from R0 on day 200), S1 (taken from R1 on day 400) and S2 (taken from R2 on day 400) were analyzed on the Illumina MiSeq platform through the high-throughput sequencing of 16S rRNA to characterize the microbial community and determine its spatiotemporal heterogeneity in the DS-EBPR reactors. The operational taxonomic units (OTUs), as well as microbial diversity estimators such as ACE, Chao1, Shannon diversity index and Simpson index, are summarized in Supplementary Table S3. The data show that the microbial diversity was slightly reduced after a period (400 days) of S-assisted denitrifying P removal. The relative abundance of the dominant microbial taxa is shown in Fig. 3. Nine major bacterial phyla and archaea were detected in the three sludge samples, in which Proteobacteria were by far the most abundant phylum accounting for over half of the microbial community. This phylum also made up 24.4% of the microbial community in the inoculum sludge sample. Proteobacteria were also significantly enriched in our previous study22 and another study on the conventional EBPR process24. Betaproteobacteria accounted for 25.9~33.4% of the microbial community in the DS-EBPR reactors, much more so than other Proteobacteria and similar to the situation in the biological nutrient removal activated sludge system where Betaproteobacteria accounted for about 33% of the microbial community25. Chloroflexi and Bacteroidetes became more abundant the longer the reactors were operated. Both were enriched to a greater extent in the bigger biomass aggregates (S2 with average particle size of 109.5 ± 0.5 μm) than in the smaller ones (S1 with average particle size of 63.2 ± 0.3 μm) after sludge size fractionation, confirming their important role in biomass aggregation as reported in a study on the anaerobic sludge granulation of SRB for organics removal26. By comparing the sequences detected in the sludge samples (S0, S1 and S2) with BLAST against NCBI’s non-redundant database, we found that the three most abundant genera were Thauera phenylacetica (OTU1), Desulfobacter postgatei (OTU7, OTU578, OTU8276 and OTU10330) and Thiobacillus thioparus (OTU4) as shown in Supplementary Table S6. The phylogenetic trees in Supplementary Fig. S3 show the core functional lineages in the sludge samples and the common lineages.

Figure 3
Relative abundance (%, n = 3) of the dominant microbial taxa in the inoculum (S) and the sludge samples (S0, S1 and S2).

In order to investigate the temporal variation and spatial heterogeneity of the microbial community in the DS-EBPR reactors, Euclidean-based PCoA27 was conducted with 12 16S rRNA gene amplicons and 11 spatial dimensions. The first two principal coordinates (PC1 and PC2) explained 62.52% of the variance in microbial assemblages, in which PC1 alone explained 48.58%. Supplementary Fig. S2 shows three distinct clusters of data points: (1) those for the inoculum sludge sample are in yellow (S); (2) those for the sludge sample collected on day 200 are in blue (S0 from R0); and (3) those for the two sludge samples collected on day 400 are in red (S1 from R1) and green (S2 from R2). In general, the closer the clusters are together, the higher the similarity among microbial communities in the four sludge samples. This is also confirmed in Supplementary Table S4 with β-diversity (Bray-Curtis distance) metrics28. According to Supplementary Fig. S2 and Supplementary Table S4, the composition of the microbial community was primarily affected by the operating time, and no spatial heterogeneity existed between two of the size-fractionated sludge samples (S1 and S2). Similar conclusions could be drawn from the heatmap in Fig. 4 and the results of OTU abundance analysis in Supplementary Table S5. The relative abundances of Thauera (OTU1) and Desulfuromonas (OTU12) increased noticeably with operating time. On the contrary, the relative abundances of Desulfobacter (OTU7, OTU578, OTU8276 and OTU10330) and Thiobacillus (OTU4 and OTU11521) after 400 days of operation were much lower than those after just 200 days of operation.

Figure 4
Heatmap of genera (relative abundance ≥ 1% in at least one sample). The color bar indicates the range of contribution of a genus to a sample, based on the color key (log10 scale) at the top. The family and genus names of the OTUs are shown ...

Key microorganisms involved in the metabolism of C, N, P and S

The relative abundance of the representive functional bacteria was at least ≥ 1% in each of S0, S1 and S2, implying that at least three bacteria (Thauera, Thiobacillus and Desulfobacter) significantly contributed to denitrifying S-assisted biological P removal. A total of nine functional traits for the three most abundant genera, i.e. Thauera, Thiobacillus and Desulfobacter, were coded from the closest related bacteria based on the phylogenetic functional traits of prokaryotes29. As indicated in Table 2, Thauera phenylacetica B4P (forming 9.24% and 12.66% of the microbial community in S0 and S1, respectively) was responsible not only for acetate uptake and PHA and glycogen formation but also for completing the glycolysis and tricarboxylic acid (TCA) cycle for acetate conversion30. However, Desulfobacter postgatei (4.89%) and Desulfuromonas (2.38%) were also detected in DS-EBPR reactor R1 after 200 days of operation. These bacteria utilize acetate as electron donors in anaerobic conditions for sulfate reduction. However, they were less abundant than Thauera phenylacetica, which was highly enriched in the DS-EBPR reactors after long-term operation. This was likely because Thauera can oxidize PHAs stored in cells through the TCA cycle to provide a proton-motive force for adenosine triphosphate (ATP) formation thereby accelerating the growth of cells10, 31.

Table 2
Functional trait analysis for three core functional genera regarding the metabolism of C, N, P and S in DS-EBPR reactors.

For the N cycle, Table 2 shows that Thauera phenylacetica and Thiobacillus thioparus can trigger denitrification. Thauera and Thiobacillus coexist in autotrophic and heterotrophic denitrification processes, but their competition has never been documented because their relative abundances have never been analyzed statistically32. In our study, the relative abundance of Thiobacillus was reduced (from 9.92% to 4.79%) once the relative abundance of Thauera had increased from 9.98% to 13.63%, implying that the two competed for nitrate as electron acceptors when conditions in the DS-EBPR systems were anoxic.

For the P cycle, we speculate that Thauera is the core genus of bacteria responsible for P removal, based on the results of this study. Thauera has been reported to be not only a common denitrifier33, 34 but also a denitrifying P-removing agent35. In addition, we also found that Thauera, Thiobacillus and Desulfobacter could trigger the formation of poly-P (Table 2). However, Thiobacillus and Desulfobacter could not induce the formation of PHAs, which are needed for the formation of poly-P in anoxic conditions. PHAs are some of the most important intracellular polymers in P-removing bacteria. However, none of the conventional PAOs was detected in our DS-EBPR reactors. The main reason could be the strong presence of sulfide (approximately 124~162 mg S/L in R0, R1 and R2) at the end of the anaerobic phase in the DS-EBPR system. It has been reported that sulfide has a negative effect on normal PAO activity, and the effect is likely related to the concentration of undissociated H2S. Fifty percent inhibition of the maximum acetate uptake rate of PAOs has been observed at around 57 mg S/L36. Moreover, Thauera is the key enabler of the DS-EBPR process. Studies have shown that the energy production efficiency of PAOs with nitrate in the EBPR process could be up to 40% lower than that with oxygen37, leading to a 20–30% lower cell yield and achieving minimal biological sludge production5. The addition of nitrate in the DS-EBPR process could also reduce sludge production, besides reduction through S-cycle conversion19.

For the S cycle, Desulfobacter postgatei, which is known to be a sulfate-reducing bacterium38, was the main species that reduced sulfate to sulfide for utilization as electron donors in anaerobic conditions. The relative abundance of Desulfobacter fell sharply from 12.85 to 4.89%. In spite of this, it was still the most abundant among all of the S-reducing bacteria present. The anaerobic S-reducing bacterium Desulfuromonas 39 was also present in this system (0.57, 1.15, and 2.38% in R2, R0, and R1, respectively). However, its abundance increased from day 200 to day 400. Hence the two bacteria may be competitors, with Desulfobacter being less effective at utilizing C than Desulfuromonas. This result has been reported elsewhere40. Moreover, Thiobacillus can autotrophically oxidize inorganic S compounds through aerobic respiration or denitrification under anoxic conditions32, 41, 42. Thiobacillus thioparus is known as a nitrate-reducing sulfide-oxidizing bacterium43 and its traits are confirmed in Table 2. It is interesting to note that although the relative abundance of Thiobacillus decreased, the sulfide-oxidizing rates of R0, R1 and R2 were still approximately 3.0 ± 0.1, 4.6 ± 0.1 and 4.3 ± 0.1 mg S/(g VSS·h) respectively. The stable performance implies that Thiobacillus is not the only bacterium capable of oxidizing sulfide in the system. Thauera sp. may also oxidize sulfide44, 45. In addition, based on the genome, we also speculate that Thauera sp. can use poly-S as electron donors for autotrophic denitrification.

It should be mentioned that the microbial community in conventional EBPR is mainly dominated by conventional PAOs and supplanted by GAOs and the other major groups (e.g. Alphaproteobacterial, Planctomycete, and Flexibacter-CytophagaBacteroides46 or Xanthomonadales, Flavobacteriales, and Rhizobiales4). In contrast, none of the conventional PAOs were enriched in DS-EBPR. In fact, only S-related bacteria—Thauera, Desulfobacter and Thiobacillus—were enriched, which is not surprising as S cycle biotransformations are prevalent in the system.

We found a significantly positive correlation between phosphate removal and sulfide oxidation (correlation coefficient > 0.7, P < 0.01) (Supplementary Table S2). Thus Thauera is likely to be the key functional bacterium for denitrifying P removal and to be associated with sulfide oxidation in the DS-EBPR system. Moreover, the purple and green S bacteria (i.e. OTU2 and OTU11) might play a role in the conversion of poly-S. These bacteria can oxidize sulfide to poly-S using CO2 as electron acceptors47. The poly-S could then be oxidized to sulfate48. However, in this study we could not confirm the exact pathway of poly-S formation. Metagenomic and metatranscriptomic studies are needed to obtain more solid evidence and insights.

Competitive and cooperative interactions among core functional microorganisms in the DS-EBPR system

The Spearman correlation between operating conditions and reactor performance was calculated (see Supplementary Table S2). Phosphate removal was significantly and positively correlated with initial acetate and nitrate levels (correlation coefficients > 0.7, P < 0.01) because the two were utilized by Thauera. However, it was significantly and negatively correlated with the initial sulfate level (correlation coefficient > 0.7, P < 0.01). SRB might compete with Thauera for C. They might grow more strongly when more sulfate is present, which would explain the negative correlation between phosphate removal and the initial sulfate level. However, the lack of sulfate initially was not necessarily beneficial for phosphate removal. The net P removal decreases when sulfate reduction weakens22. In addition, this study shows that the presence of SRB benefits the reduction of sulfate and the generation of sulfide, the latter of which is then utilized by SOB in DS-EBPR. Meanwhile, phosphate removal was significantly and positively correlated with sulfide oxidation (correlation coefficient > 0.7, P < 0.01). The Spearman correlation further confirms that Thauera was the denitrifying polyphosphate-accumulating organism (DPAO) and was likely associated with sulfide oxidation. Thus cooperation and competition were both observed in the DS-EBPR system.

Based on this and our previous studies11, 17, 22, 23, we propose for the first time that competitive and cooperative metabolic interactions (Fig. 5) occur within the microbial community in the DS-EBPR system. Thauera is likely a major contributor to P removal. Equation (1) describes the generation and degradation of poly-P23. Other PAO-related bacteria may have been present but went undetected because their concentrations were too low and the functional traits of their genomes are unknown. Thauera and SRB (mainly Desulfobacter) compete with each other for acetate. Desulfobacter oxidizes acetate and reduces sulfate to sulfide under anaerobic conditions, as described in equation (2)49. Meanwhile, SOB utilize sulfide as electron donors in the DS-EBPR system. Studies have demonstrated syntrophic growth between SRB and SOB, such that a positive-feedback S cycle is established between the reductive and oxidative processes of these two kinds of bacteria50. Moreover, Thiobacilus can also stimulate further sulfide oxidation with the help of nitrate42, as described in equation (3)51. Furthermore, Thiobacillus may compete with Thauera for nitrate as electron acceptors and sulfide as electron donors in the DS-EBPR system. Competition existed among bacteria, but on the whole the core functional bacteria related to C, N, P and S interacted syntrophically in the DS-EBPR system. This work demonstrates conceptual metabolic interactions. The results help improve our understanding of the DS-EBPR system, allowing us to incorporate insights derived from laboratory experiments into a framework that can potentially be used to assess risk and design optimal schemes for treating wastewater.

ATP + H3PO4 ↔ HPO3(poly − P) + H2O

Acetate + SO42− → HS + 2 HCO3

5 HS + 8 NO3 + 3 H+ → 5 SO42− + 4 N2 ↑  + 4 H2O

Figure 5
Possible competition and cooperation among bacteria in DS-EBPR. S-DPAO stands for denitrifying polyphosphate-accumulating organism and is associated with sulfide oxidation; SRB represents sulfate-reducing bacteria; NR-SOB means nitrate-reducing and sulfide-oxidizing ...


Reactor setup, seeding inoculum and operation

This study used three separate tightly-sealed SBRs (R0, R1 and R2) for denitrifying S-assisted P removal as shown in Supplementary Table S1 of supplementary information (SI). These reactors were operated under similar alternating anaerobic/anoxic conditions according to the following operating steps: (1) the SBRs were fed with synthetic sewage for 20 min with the influent volumes shown in Table 1 (sewage characteristics are shown in SI); (2) P release, i.e. reaction phase R, occurred in anaerobic conditions for different periods of time as shown in Table 1; (3) the sewage was spiked with 100 mL of nitrate solution with an average concentration of 40 mg N/L for R0 and R1 and 31 mg N/L for R2; (4) P uptake, i.e. reaction phase U, occurred in anoxic conditions for different periods of time as shown Table 1; and (5) settling and decanting lasted 80 and 20 min, respectively.

Reactor R0 with a working volume of 10 L was initially inoculated with concentrated flocculent activated sludge (having an average particle size of 55.8 ± 0.4 μm) obtained from the secondary sedimentation tank of the Sha Tin Sewage Treatment Works, the largest saline sewage treatment plant in Hong Kong. This reactor was operated according to the five steps described above with mechanical stirring at 200 rpm at 22 ± 2 °C for 200 days to cultivate functional microorganisms and achieve denitrifying S-assisted P removal. In order to investigate the spatial heterogeneity of functional microorganisms in the DS-EBPR reactors, 10 L of activated sludge were removed from R0 on day 200 and immediately filtered with a mesh screen having a pore size of 57 μm for size fractionation following the protocol used in our previous study52. We thus obtained 0.2 L of concentrated sludge with an average particle size exceeding 57 μm which we then transferred to a smaller reactor R2 (with a working volume of 1.4 L). The initial sludge concentration was 3.8 ± 0.2 mg/L mixed liquor volatile suspended solids (MLVSS). Gentle stirring was applied at 150 rpm at 22 ± 2 °C to promote biomass aggregation. Sludge with an average particle size of less than 57 μm after size fractionation was transferred to another reactor R1 (with a working volume of 10 L). In this case the initial sludge concentration was 2.9 ± 0.2 mg/L MLVSS and stirring was applied at 200 rpm at 22 ± 2 °C to prevent biomass aggregation. Figure 1a illustrates the setup. Images of sludge samples S1 and S2 were taken using a field emission scanning electron microscope and are shown in Supplementary Fig. S1.

Sludge sampling, DNA extraction, polymerase chain reaction (PCR) amplification, and Illumina MiSeq sequencing

In order to investigate the spatial and temporal changes in the bacterial population in the DS-EBPR reactors, an inoculum sludge sample S was taken on day 0 and three activated sludge samples were taken on day 200 (S0 from reactor R0 after cultivation for 200 days) and day 400 (S1 from R1 and S2 from R2 after size fractionation and cultivation for another 200 days) (as shown in Fig. 1a). The DNA in each sludge sample was extracted immediately after sludge sampling via the following steps28: (1) 2 mL of thoroughly mixed sludge sample were centrifuged at 4 °C for 10 min at 10,000 g in a high-speed centrifuge (HERMLE Labortechnik GmbH, Germany); (2) after pouring out the suspension, 200 mg of sludge were transferred to a 2 mL DNA LoBind tube (Eppendorf, Germany) for DNA extraction; and (3) the community DNA of microbes was extracted using the FastDNA Spin Kit (Qbiogene, Carlsbad, CA) according to the manufacturer’s instructions.

The universal primer set F515/R806 was used to amplify the V4 hypervariable region of the 16S rRNA gene. This region covered nearly all bacterial and archaeal taxa53. The DNA samples were amplified in 50 μL of reaction mixtures. Amplification was performed with the GeneAmp PCR System 9700 (ABI, US) according to the following steps: denaturing at 94 °C for 5 min, followed by 31 cycles of denaturing at 94 °C for 30 s, annealing at 52 °C for 30 s, extension at 72 °C for 45 s, and final extension at 72 °C for 10 min. Replicate amplicons were pooled for purification with the QIAquick Gel Extraction Kit (Qiagen, Chatsworth, CA). The integrity of the DNA and PCR amplification was verified using agarose gel electrophoresis, and concentration and purity were determined using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, US). Adaptor-appended fragments were sequenced on the Illumina MiSeq PE300 platform. The sequences were deposited in the European Molecular Biology Laboratory (EMBL) sequence read archive with accession number ERR1778916.

Bioinformatic analysis and data availability

The raw paired-end reads with a Phred quality score of less than 20 at the 3’ end were trimmed using a custom Perl script. Contigs were then assembled using FLASH54 for each read pair. Chimeric contigs were identified and removed using USEARCH (version 8.1.1861). Then the quality sequences were processed using the pipelines of Mothur55 (version 1.29) and QIIME56. Briefly, we clustered all of the contigs at the 97% similarity level to obtain OTUs for all samples. Taxonomic information was determined using the Ribosomal Database Project (RDP) Classifier57 (version 2.6) with a minimum confidence of 80%. The relative abundance of each individual OTU was computed as the ratio of the sequence number of that OTU to the total number of sequences assigned to that sample. To avoid bias caused by sequencing depth, we sampled the microbial community by taking the smallest sample size of 10,000 sequences to calculate the metrics of α-diversity (mainly the ACE, Chao1, Shannon and Simpson indexes) and β-diversity (e.g., Bray-Curtis, weighted unifrac and unweighted unifrac). Neighbor-joining-based phylogenetic trees of core OTUs and their closely related species were generated using MEGA58 (version 5.2). The Euclidean-based PCoA was used to show the different assemblages between the DS-EBPR sludge samples and the seeding sludge. Independent sample t-tests based on Bray-Curtis distances (P < 0.05), Spearman correlation (two-tailed) and one-way ANOVA followed by Tukey and Tamhane’s T2 post-hoc test (P < 0.05) were performed in SPSS 18.0. A heatmap was generated using the pheatmap package in the statistical program R. Sequences were compared using BLAST against NCBI’s non-redundant database (Supplementary Table S6). Bacterial functional traits are detailed in SI.

Chemical analysis

Mixed liquor suspended solids, MLVSS and sulfide were determined according to Standard Methods59. Key anions in the bulk liquid, including acetate, nitrate, nitrite, phosphate, thiosulfate, sulfite and sulfate, were determined with an ion chromatograph equipped with an IC-AS23 analytical column (DIONEX ICS-900, US). The concentration of poly-P (in mg P/L) was calculated from the changes in the bulk liquid phosphate concentration22. Glycogen and poly-S were detected following previous methods11. PHAs were determined with a high-performance liquid chromatograph (HPLC, Dionex Ultimate 3000, US) following the protocol described in SI. The sludge size was measured using the Mastersizer 3000 laser diffraction floc size analyzer (Malvern Instruments Ltd., UK). A scanning electron microscope was used to take pictures of sludge samples prepared according to procedures detailed in Kloep et al.60.

Data availability

The datasets generated in the current study are available from the corresponding author on request.

Electronic supplementary material


This study was supported by the National Natural Science Foundation of China (grant nos 51308558 and 51638005), the Natural Science Foundation of Guangdong Province (grant no. 2015A030313149) and the Hong Kong Innovation and Technology Commission (grant no. ITC-CNERC14EG03). The authors also acknowledge support from Tip-top Scientific and Technical Innovative Youth Talents of Guangdong special support program (No. 2016TQ03Z336), the Fundamental Research Funds for the Central Universities of China (grant no. 13lgpy62), the Research Fund Program of Guangdong Provincial Key Laboratory of Environmental Pollution Control and Remediation Technology (grant no. 2016K0002), Guangdong Provincial Science and Technology Planning Project (grant no. 2016A050503041) and the Hong Kong Research Grants Council (grant no. 16213515).

Author Contributions

Author Contributions

Y.Z., D.W. and H.L. proposed the project, analysed the data, and wrote the manuscript; Y.Z. and M.Y. conducted most of the experiments; Z.H. conducted the high-throughput 16S rRNA gene analysis; J.G. and G.C. checked the manuscript before submission.


Competing Interests

The authors declare that they have no competing interests.


Electronic supplementary material

Supplementary information accompanies this paper at doi:10.1038/s41598-017-11448-x

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Di Wu, kh.tsu@iduwec.

Hui Lu, nc.ude.usys.liam@3iuhvl.


1. Kong Y, Xia Y, Nielsen JL, Nielsen PH. Structure and function of themicrobial community in a full-scale enhanced biological phosphorus removal plant. Microbiology. 2007;153:4061–4073. doi: 10.1099/mic.0.2007/007245-0. [PubMed] [Cross Ref]
2. Nielsen PH, et al. A conceptual ecosystem model of microbial communities in enhanced biological phosphorus removal plants. Water Res. 2010;44(17):5070–5088. doi: 10.1016/j.watres.2010.07.036. [PubMed] [Cross Ref]
3. Lanham AB, et al. Metabolic versatility in full-scale wastewater treatment plants performing enhanced biological phosphorus removal. Water Res. 2013;47(19):7032–7041. doi: 10.1016/j.watres.2013.08.042. [PubMed] [Cross Ref]
4. Garcia Martin H, et al. Metagenomic analysis of two enhanced biological phosphorus removal (EBPR) sludge communities. Nat Biotechnol. 2006;24(10):1263–1269. doi: 10.1038/nbt1247. [PubMed] [Cross Ref]
5. Oehmen A, et al. Advances in enhanced biological phosphorus removal: from micro to macro scale. Water Res. 2007;41(11):2271–2300. doi: 10.1016/j.watres.2007.02.030. [PubMed] [Cross Ref]
6. Kristiansen R, et al. A metabolic model for members of the genus Tetrasphaera involved in enhanced biological phosphorus removal. ISME J. 2013;7(3):543–554. doi: 10.1038/ismej.2012.136. [PMC free article] [PubMed] [Cross Ref]
7. Mao Y, Yu K, Xia Y, Chao Y, Zhang T. Genome reconstruction and gene expression of “Candidatus Accumulibacter phosphatis” Clade IB performing biological phosphorus removal. Environ Sci Technol. 2014;48(17):10363–10371. doi: 10.1021/es502642b. [PubMed] [Cross Ref]
8. Oyserman BO, Noguera DR, Del Rio TG, Tringe SG, McMahon KD. Metatranscriptomic insights on gene expression and regulatory controls in Candidatus Accumulibacter phosphatis. ISME J. 2016;10(4):810–822. doi: 10.1038/ismej.2015.155. [PMC free article] [PubMed] [Cross Ref]
9. Bassin JP, Kleerebezem R, Dezotti M, van Loosdrecht MC. Simultaneous nitrogen and phosphate removal in aerobic granular sludge reactors operated at different temperatures. Water Res. 2012;46(12):3805–3816. doi: 10.1016/j.watres.2012.04.015. [PubMed] [Cross Ref]
10. Baetens D, Weemaes M, Hosten L, Vos PD, Vanrolleghem A. P. Enhanced Biological Phosphorus Removal Competition and symbiosis between SRBs and PAOs on lactateacetate feed. Journal of Experimental Botany. 2001;57(14):3813–3824.
11. Wu D, et al. A new biological phosphorus removal process in association with sulfur cycle. Water Res. 2013;47(9):3057–3069. doi: 10.1016/j.watres.2013.03.009. [PubMed] [Cross Ref]
12. Abou-Elela SI, Kamel MM, Fawzy ME. Biological treatment of saline wastewater using a salt-tolerant microorganism. Desalination. 2010;250(1):1–5. doi: 10.1016/j.desal.2009.03.022. [Cross Ref]
13. Rauch W, Kleidorfer M. Replace contamination, not the pipes. Science. 2014;345(6198):734–735. doi: 10.1126/science.1257988. [PubMed] [Cross Ref]
14. Bear, J., Cheng, A. H., Sorek, S., Ouazar, D. & Herrera, I. In Seawater Intrusion in Coastal Aquifers-Concepts, Methods and Practices. (Kluwer Academic Publishers, 1999).
15. Hong K Water Supplies Department (HKWSD) Annually Report 2014/2015: Exploit New Water Resources and Foster Water Conservation Cultureurl. (2015).
16. Leung RW, et al. Integration of seawater and grey water reuse to maximize alternative water resource for coastal areas: the case of the Hong Kong International Airport. Water Sci Technol. 2012;65(3):410–417. doi: 10.2166/wst.2012.768. [PubMed] [Cross Ref]
17. Guo G, et al. Functional bacteria and process metabolism of the Denitrifying Sulfur conversion-associated Enhanced Biological Phosphorus Removal (DS-EBPR) system: An investigation by operating the system from deterioration to restoration. Water Res. 2016;95:289–299. doi: 10.1016/j.watres.2016.03.013. [PubMed] [Cross Ref]
18. Wang J, et al. A novel sulfate reduction, autotrophic denitrification, nitrification integrated (SANI) process for saline wastewater treatment. Water Res. 2009;43(9):2363–2372. doi: 10.1016/j.watres.2009.02.037. [PubMed] [Cross Ref]
19. Lu H, et al. Steady-state model-based evaluation of sulfate reduction, autotrophic denitrification and nitrification integrated (SANI) process. Water Res. 2009;43(14):3613–3621. doi: 10.1016/j.watres.2009.05.013. [PubMed] [Cross Ref]
20. Lu H, et al. SANI® process realizes sustainable saline sewage treatment: steady state model-based evaluation of the pilot-scale trial of the process. Water Res. 2012;46(2):475–490. doi: 10.1016/j.watres.2011.11.031. [PubMed] [Cross Ref]
21. Lu H, et al. The demonstration of a novel sulfur cycle-based wastewater treatment process: sulfate reduction, autotrophic denitrification, and nitrification integrated (SANI®) biological nitrogen removal process. Biotechnol Bioeng. 2012;109(11):2778–2789. doi: 10.1002/bit.24540. [PubMed] [Cross Ref]
22. Wu D, et al. Simultaneous nitrogen and phosphorus removal in the sulfur cycle-associated Enhanced Biological Phosphorus Removal (EBPR) process. Water Res. 2014;49:251–264. doi: 10.1016/j.watres.2013.11.029. [PubMed] [Cross Ref]
23. Yu M, et al. Effects of carbon-to-sulfur (C/S) ratio and nitrate (N) dosage on Denitrifying Sulfur cycle-associated Enhanced Biological Phosphorus Removal (DS-EBPR) Sci Rep. 2016;6:23221. doi: 10.1038/srep23221. [PMC free article] [PubMed] [Cross Ref]
24. Crocetti GR, et al. Identification of Polyphosphate-Accumulating Organisms and Design of 16S rRNA-Directed Probes for Their Detection and Quantitation. Applied and Environmental Microbiology. 2000;66(3):1175–1182. doi: 10.1128/AEM.66.3.1175-1182.2000. [PMC free article] [PubMed] [Cross Ref]
25. Wilen BM, Onuki M, Hermansson M, Lumley D, Mino T. Microbial community structure in activated sludge floc analysed by fluorescence in situ hybridization and its relation to floc stability. Water Res. 2008;42(8-9):2300–2308. doi: 10.1016/j.watres.2007.12.013. [PubMed] [Cross Ref]
26. Hao T, et al. Characterization of sulfate-reducing granular sludge in the SANI® process. Water Res. 2013;47(19):7042–7052. doi: 10.1016/j.watres.2013.07.052. [PubMed] [Cross Ref]
27. Legendre, P. & Birks, H. J. Tracking environmental change using lake sediments in From classical to canonical ordination. (eds Birks, H. J. et al.) 201–248 (Springer, 2012).
28. Liu J, et al. Ecological effects of combined pollution associated with e-waste recycling on the composition and diversity of soil microbial communities. Environ Sci Technol. 2015;49(11):6438–6447. doi: 10.1021/es5049804. [PubMed] [Cross Ref]
29. Goberna M, Navarro-Cano JA, Valiente-Banuet A, Garcia C, Verdu M. Abiotic stress tolerance and competition-related traits underlie phylogenetic clustering in soil bacterial communities. Ecol Lett. 2014;17(10):1191–1201. doi: 10.1111/ele.12341. [PubMed] [Cross Ref]
30. Mao Y, Xia Y, Wang Z, Zhang T. Reconstructing a Thauera genome from a hydrogenotrophic-denitrifying consortium using metagenomic sequence data. Appl Microbiol Biotechnol. 2014;98(15):6885–6895. doi: 10.1007/s00253-014-5756-x. [PubMed] [Cross Ref]
31. He S, et al. Metatranscriptomic array analysis of ‘Candidatus Accumulibacter Phosphatis’-enriched enhanced biological phosphorus removal sludge. Environ Microbiol. 2010;12(5):1205–1217. doi: 10.1111/j.1462-2920.2010.02163.x. [PubMed] [Cross Ref]
32. Xu G, et al. Evaluation of simultaneous autotrophic and heterotrophic denitrification processes and bacterial community structure analysis. Applied Microbiology and Biotechnology. 2015;99(15):6527–6536. doi: 10.1007/s00253-015-6532-2. [PubMed] [Cross Ref]
33. Thomsen TR, Kong Y, Nielsen PH. Ecophysiology of abundant denitrifying bacteria in activated sludge. FEMS Microbiol Ecol. 2007;60(3):370–382. doi: 10.1111/j.1574-6941.2007.00309.x. [PubMed] [Cross Ref]
34. Peng X, Guo F, Ju F, Zhang T. Shifts in the microbial community, nitrifiers and denitrifiers in the biofilm in a full-scale rotating biological contactor. Environ Sci Technol. 2014;48(14):8044–8052. doi: 10.1021/es5017087. [PubMed] [Cross Ref]
35. Tsuneda S, Miyauchi R, Ohno T, Hirata A. Characterization of denitrifying polyphosphate-accumulating organisms in activated sludge based on nitrite reductase gene. J Biosci Bioeng. 2005;99(4):403–407. doi: 10.1263/jbb.99.403. [PubMed] [Cross Ref]
36. Saad SA, Welles L, Lopez-Vazquez CM, Loosdrecht MCM. v. & Brdjanovic, D. Sulfide Effects on the Anaerobic Kinetics of Phosphorus-Accumulating Organisms. Chemical Engineering Journal. 2017;326:68–77. doi: 10.1016/j.cej.2017.05.074. [Cross Ref]
37. Kuba T, Murnleitner E, van Loosdrecht MC, Heijnen JJ. Metabolic Model for Biological Phosphorus Removal by Denitrifying Organisms. Biotechnol Bioeng. 1996;52:685–695. doi: 10.1002/(SICI)1097-0290(19961220)52:6<685::AID-BIT6>3.3.CO;2-M. [PubMed] [Cross Ref]
38. Ingvorsen K, Zehnder AJ, Jørgensen BB. Kinetics of sulfate and acetate uptake by Desulfobacter postgatei. Appl Environ Microbiol. 1984;47(2):403–408. [PMC free article] [PubMed]
39. Finster K, Bak F. & Pfennig, N. Desulfuromonas acetexigens sp. nov., a dissimilatory sulfur-reducing eubacterium from anoxic freshwater sediments. Arch Microbiol. 1994;161(4):328–332. doi: 10.1007/BF00303588. [Cross Ref]
40. Purdy KJ, Nedwell DB, Embley TM. Analysis of the Sulfate-Reducing Bacterial and Methanogenic Archaeal Populations in Contrasting Antarctic Sediments. Applied and Environmental Microbiology. 2003;69(6):3181–3191. doi: 10.1128/AEM.69.6.3181-3191.2003. [PMC free article] [PubMed] [Cross Ref]
41. Chen LX, et al. Shifts in microbial community composition and function in the acidification of a lead/zinc mine tailings. Environ Microbiol. 2013;15(9):2431–2444. doi: 10.1111/1462-2920.12114. [PubMed] [Cross Ref]
42. Hubert C, Voordouw G. Oil field souring control by nitrate-reducing Sulfurospirillum spp. that outcompete sulfate-reducing bacteria for organic electron donors. Appl Environ Microbiol. 2007;73(8):2644–2652. doi: 10.1128/AEM.02332-06. [PMC free article] [PubMed] [Cross Ref]
43. Vlasceanu L, Popa R, Kinkle B. Characterization of Thiobacillus thioparus LV43 and Its Distribution in a Chemoautotrophically Based Groundwater Ecosystem. Appl Environ Microbiol. 1997;63(8):3123–3127. [PMC free article] [PubMed]
44. Cytryn E, et al. Identification of bacteria potentially responsible for oxic and anoxic sulfide oxidation in biofilters of a recirculating mariculture system. Appl Environ Microbiol. 2005;71(10):6134–6141. doi: 10.1128/AEM.71.10.6134-6141.2005. [PMC free article] [PubMed] [Cross Ref]
45. Auguet O, Pijuan M, Guasch-Balcells H, Borrego CM, Gutierrez O. Implications of Downstream Nitrate Dosage in anaerobic sewers to control sulfide and methane emissions. Water Res. 2015;68:522–532. doi: 10.1016/j.watres.2014.09.034. [PubMed] [Cross Ref]
46. Bond PL, Hugenholtz P, Keller J, Blackall LL. Bacterial community structures of phosphate-removing and non-phosphate-removing activated sludges from sequencing batch reactors. Appl Environ Microbiol. 1995;61(5):1910–1916. [PMC free article] [PubMed]
47. Casamayor EO. Carbon dioxide fixation in the dark by photosynthetic bacteria in sulfide-rich stratified lakes with oxic–anoxic interfaces. Limnol. Oceanogr. 2008;53(4):1193–1203. doi: 10.4319/lo.2008.53.4.1193. [Cross Ref]
48. Fike DA, Bradley AS, Rose CV. Rethinking the Ancient Sulfur Cycle. Annual Review of Earth and Planetary Sciences. 2015;43(1):593–622. doi: 10.1146/annurev-earth-060313-054802. [Cross Ref]
49. Lens PNL, Visser A, Janssen AJH, Pol LWH, Lettinga G. Biotechnological Treatment of Sulfate-Rich Wastewaters. Critical Reviews in Environmental Science and Technology. 1998;28(1):41–88. doi: 10.1080/10643389891254160. [Cross Ref]
50. van den Ende FP, Meier J, van Gemerden H. Syntrophic growth of sulfate-reducing bacteria and colorless sulfur bacteria during oxygen limitation. FEMS Microbiol Ecol. 1997;23(1):65–80. doi: 10.1016/S0168-6496(97)00014-7. [Cross Ref]
51. Robertson LA, Kuenen JG. The Genus. Thiobacillus. Prokaryotes. 2006;5:812–827.
52. Yang W, et al. Sulfide-driven autotrophic denitrification significantly reduces N2O emissions. Water Res. 2016;90:176–184. doi: 10.1016/j.watres.2015.12.032. [PubMed] [Cross Ref]
53. Bates ST, et al. Examining the global distribution of dominant archaeal populations in soil. ISME J. 2011;5(5):908–917. doi: 10.1038/ismej.2010.171. [PMC free article] [PubMed] [Cross Ref]
54. Magoc T, Salzberg SL. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics. 2011;27(21):2957–2963. doi: 10.1093/bioinformatics/btr507. [PMC free article] [PubMed] [Cross Ref]
55. Schloss PD, et al. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol. 2009;75(23):7537–7541. doi: 10.1128/AEM.01541-09. [PMC free article] [PubMed] [Cross Ref]
56. Caporaso JG, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7(5):335–336. doi: 10.1038/nmeth.f.303. [PMC free article] [PubMed] [Cross Ref]
57. Wang Q, Garrity GM, Tiedje JM, Cole JR. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol. 2007;73(16):5261–5267. doi: 10.1128/AEM.00062-07. [PMC free article] [PubMed] [Cross Ref]
58. Hall BG. Building phylogenetic trees from molecular data with MEGA. Mol Biol Evol. 2013;30(5):1229–1235. doi: 10.1093/molbev/mst012. [PubMed] [Cross Ref]
59. Apha, A. & Wef. In Standard methods for examination of water and wastewater (ed. Franson, M.) Ch. 2, 54–60, Ch. 4, 162–173 (APHA, 1998).
60. Kloep F, Röske I, Neu TR. Performance and microbial structure of a nitrifying fluidized-bed reactor. Water Res. 2000;34(1):311–319. doi: 10.1016/S0043-1354(99)00123-2. [Cross Ref]

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group