Bacterial community analysis of anoxic/aeration (A/O) system in a combined process for gibberellin wastewater treatment

Abstract

Gibberellin wastewater cannot be directly discharged without treatment due to its high concentrations of sulfate and organic compounds and strong acidity. Therefore, multi-stage anaerobic bioreactor + micro-aerobic+ anoxic/aeration (A/O) + biological contact oxidation combined processes are used to treat gibberellin wastewater. However, knowledge of the treatment effects of the A/O process and bacterial community structure in the aeration tank reactors of such systems is sparse. Therefore, this study was conducted to investigate the treatment effects and operation of the A/O process on gibberellin wastewater, as well as changes in the bacterial community structure of activated sludge in the aeration tank during treatment. Moreover, removal was examined based on evaluation of effluent after A/O treatment. Although influent chemical oxygen demand (COD), NH₃-N and total phosphorus (TP) fluctuated, effluent COD, NH₃-N and TP remained stable. Moreover, average COD, NH₃-N and TP removal efficiency were 68.41%, 93.67% and 45.82%, respectively, during the A/O process. At the phylum level, Proteobacteria was the dominant phylum in all samples, followed by Chloroflexi, Bacteroidetes and Actinobacteria. Proteobacteria played an important role in the removal of organic matter. Chloroflexi was found to be responsible for the degradation of carbohydrates and Bacteroidetes also had been found to be responsible for the degradation of complex organic matters. Actinobacteria are able to degrade a variety of environmental chemicals. Additionally, Anaerolineaceae_uncultured was the major genus in samples collected on May 25, 2015, while Novosphingobium and Nitrospira were dominant in most samples. Nitrosomonas are regarded as the dominant ammonia-oxidizing bacteria, while Nitrospira are the main nitrite-oxidizing bacteria. Bacterial community structure varied considerably with time, and a partial Mantel test showed a highly significant positive correlation between bacterial community structure and DO. The bacterial community structure was also positively correlated with temperature and SO₄^2-.

Introduction

Plant hormones are essential intrinsic regulators of plant development and growth [1]. Gibberellins not only promote internode elongation and cambial activity of stems [2], but also participate in the regulation of plant growth and development [3], including seed germination, stem elongation, flowering development, fruit growth, and root development. Gibberellin wastewater is generated by biological fermentation processes. The raw materials and methods for production of gibberellin result in gibberellin wastewater containing high concentrations of sulfate, organic compounds, and NH₃-N, as well as strong acidity. These features make wastewater difficult to treat; therefore, a highly efficient treatment is required before discharge to prevent pollution. Fig 1 shows the production process of gibberellin.

Fig 1. The production process of gibberellin.

High concentrations of sulfate have been found to have an inhibitory effect on microorganisms during biological treatment [4]. Nevertheless, Li. et al. [5] investigated the treatment of chemical synthesis-based pharmaceutical wastewater containing rich organic sulfur compounds and sulfate using an Upflow Anaerobic Sludge Blanket (UASB) reactor. They found that this method removed a considerable amount of COD and sulfate. The anaerobic/oxic (A/O) process is consists of a sequential anaerobic and aerobic stage for the biological phosphorus removal [6]. The A/O process has been widely used for sewage treatment in both urban and rural areas and is favored for its high efficiency and low energy consumption [7]. In this study, the concentrations of organic materials and sulfate were both high, and the composition of wastewater was very complex. The process for treatment of gibberellin production wastewater with a multi-stage anaerobic bioreactor combined with micro oxygen, A/O, and contact oxidation was established herein, while also applying chemical dosing to reduce the sulfate.

The specific purpose of this study was to investigate the impacts and operational conditions of the A/O process when applied for nutrient removal from gibberellin wastewater and to investigate changes in the bacterial community structure and diversity in activated sludge during treatment. The results demonstrated stability and effectiveness of the A/O process based on eight months of observation. Moreover, the bacterial community diversity in activated sludge was studied by Illumina MiSeq sequencing.

Materials and methods

Ethics statement

Permission for the samples was granted by the Jiangxi Ruifeng Biochemical Co. Ltd. (Jiangxi, China). The field studies also did not involve endangered or protected species.

Raw water

The wastewater was obtained from a biochemical company located in Jiangxi province, China (115.478584°E, 27.900485°N). This company can produce 110 tons of gibberellin per year, accounting for more than 39% of the world's total gibberellin production. The gibberellin wastewater consists of three parts: raffinate wastewater, high concentration wastewater and comprehensive wastewater. In addition, the gibberellin wastewater contained high concentrations of sulfate and organic compounds, as well as strong acidity. The quality of gibberellin wastewater was shown in Table 1.

Table 1. Quality of gibberellin wastewater.

Wastewater		COD (mg/L)	SO42- (mg/L)	NH3-N (mg/L)	pH
Raffinate wastewater		40000	30000	170	2–3
High concentration wastewater		4800	1000	100	4–5
Comprehensive wastewater	Acid and alkali production wastewater	3000	100	——	5–6
	Low concentration wastewater	350	——	——	6.5–7.5
	Domestic sewage	250	——	35	7–7.5

Start-up and operation of A/O system

The A/O system could remove most of the NH₃-N and a portion of the organics. The volume of A/O system was 2200m³ and the hydraulic retention time was 26.52h. The inoculum of in the A/O system is activated sludge and it was obtained from a previous company which produces gibberellin. The quantity of the sludge inoculation was 30 t. In the preliminary stage of the setup, the A/O system took in water once a day in the first day and the COD was kept at about 300 mg/L. The influent was stopped over the next two days and the A/O system stopped taking in water and kept on aeration. After the third day, the A/O system took in water twice a day. After operation for 15 days, the reactor was operated with continuous influent at half of the design flow. Additionally, the COD of the influent was strengthened during this time. After 25 days, there were large amounts of Vorticella sp. and rotifers in the reactor. The removal of COD and NH₃-N was relatively high and the quality of the effluent was good. At this time, the influent COD was equal to the design loading of COD. After 30 days, the sludge became brown owing to flocculation, indicating successful reactor setup and good sludge activity. After setup, the reactor was used to treat the wastewater. During the setup period, the aerobic tank must remain aerated, and the dissolved oxygen should be maintained at 3 mg/L, while the pH was kept at about 6.5 to 8. Additionally, the pH of the anaerobic tank was maintained at 6 to 8 and the DO was less than 0.8 mg/L. Finally, KH₂PO₄ was added to the wastewater to maintain a COD:N:P ratio of 100:5:1. Fig 2 shows a schematic diagram of the treatment system.

Fig 2. Process flow of wastewater treatment.

Microbial diversity analysis

Sample collection

A total of 12 samples were collected from the aeration tank. Samples were collected per half a month. Each sample was dispensed into a 1.5 mL sterile Eppendorf tube and centrifuged at 14,000×g for 10 min. The supernatant was decanted, and the pellets were stored at -20°C prior to analysis. For all samples, the sample numbers begin with the sample date.

DNA extraction and PCR amplification

Microbial DNA was extracted from samples using an E.Z.N.A.^® activated sludge DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer’s protocols. The V4-V5 region of the bacterial 16S ribosomal RNA gene was amplified by PCR (95°C for 2 min, followed by 25 cycles at 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s and then final extension at 72°C for 5 min) using primers 515F 5′-barcode- GTGCCAGCMGCCGCGG-3′ and 907R 5′-CCGTCAATTCMTTTRAGTTT-3′, which contained an eight-base sequence barcode unique to each sample. PCR reactions were performed in triplicate 20 μL mixtures containing 4 μL of 5 × FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase, and 10 ng of template DNA.

Illumina MiSeq sequencing

Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to the manufacturer’s instructions, then quantified using QuantiFluor™-ST (Promega, Madison, WI, USA). Purified amplicons were pooled in equimolar amounts and paired-end sequenced (2 × 250) on an Illumina MiSeq platform according to the standard protocols. The raw reads were deposited into the NCBI Sequence Read Archive (SRA) database (Accession Number: SRP077353).

Processing of sequencing data

Raw fastq files were demultiplexed, then quality-filtered using QIIME (version 1.17) with the following criteria: (i) 300 bp reads were truncated at any site receiving an average quality score <20 over a 50 bp sliding window, discarding truncated reads that were shorter than 50 bp. (ii) Exact barcode matching, any sequences with two nucleotide mismatches upon primer matching or reads containing ambiguous characters were removed. (iii) Only sequences that overlapped by more than 10 bp were assembled according to their overlap sequence. Reads that could not be assembled were discarded.

Operational units (OTUs) were clustered with a 97% similarity cutoff using UPARSE (version 7.1 http://drive5.com/uparse/) and chimeric sequences were identified and removed using UCHIME. The taxonomy of each 16S rRNA gene sequence was analyzed by the RDP Classifier (http://rdp.cme.msu.edu/) against the silva (SSU115) 16S rRNA database using a confidence threshold of 70% [8].

Statistical analysis

The Shannon–Wiener index was used to assess bacterial diversity. Principal component analysis (PCA) was conducted to examine variations among bacterial communities of these thirty-six samples. Canonical correspondence analysis (CCA) was used to illustrate relationship among operational conditions and bacterial community. Statistical analyses were performed using a VEGAN package in R (v.2.15.1; http://www.r-project.org/) and CCA was performed by Canoco for Windows 4.5.

Results and discussion

Effect of combined A/O process on nutrient removal from gibberellin wastewater

As shown in Fig 3, the influent COD concentrations fluctuated between 346 and 772 mg/L, with the lowest concentration (346 mg/L) being observed at 29 days and 121 days. Moreover, the highest influent COD concentration of 772 mg/L was observed at 53 days. The effluent COD concentrations fluctuated between 50.29 and 76.45. Taken together, these results indicate that removal efficiency changed with influent COD concentration. Moreover, the average COD removal efficiency was 68.41%.

Fig 3. Removal of COD by A/O processes.

As shown in Fig 4, the influent NH₃-N concentrations fluctuated between 17.87 and 72.50 mg/L, with 17.87 being observed on days 33 and 81 and 72.50 being observed at 149 days. However, the effluent NH₃-N concentration remained relatively stable, and the NH₃-N removal efficiency was around 93.67%.

Fig 4. Removal of NH₃-N by A/O processes.

As shown in Fig 5, the concentration of effluent SO₄^2- was higher than that of influent SO₄^2-, possibly because some of the SO₄^2- had transformed into H₂S, and some of the H₂S had transformed into SO₄^2- in the A/O system.

Fig 5. Removal of SO₄^2- by A/O processes.

As shown in Fig 6, the influent TP concentrations fluctuated between 0.747 and 1.158 mg/L, while the effluent TP concentrations fluctuated between 0.428 and 0.604 mg/L. Moreover, the average TP removal efficiency was 45.82%.

Fig 6. Removal of TP by A/O processes.

These results show that A/O processes had good effects on gibberellin wastewater. Moreover, this method could remove most of the COD, NH₃-N and TP with removal efficiencies of 68.41%, 93.67% and 45.82%, respectively.

Illumina MiSeq sequencing results and microbial community structure

As shown in Table 2, there were 18,563 effective reads for 12 activated sludge samples. On average, the samples had 955.9 OTUs. The diversity of the bacterial community was determined based on the Shannon–Wiener index, which ranged from 4.81 to 5.43 (average = 5.18) [9]. These values are typical for diverse microbial populations without a few strongly dominant taxa [10].

Table 2. Diversity indices from 12 samples.

Samples	Sequences	OTUs	Shannon-Wiener
20150525	18563	824	4.81
20150605	18563	833	5.07
20150625	18563	813	4.93
20150729	18563	884	5.00
20150813	18563	1006	5.42
20150828	18563	954	5.28
20150909	18563	890	5.06
20150925	18563	968	5.31
20151015	18563	974	5.17
20151117	18563	1118	5.43
20151120	18563	1093	5.33
20151205	18563	1114	5.4

As show in Fig 7, Proteobacteria was the dominant phylum in all samples, accounting for 27.19%–61.33% of the total effective bacterial sequences. These findings are concordant with those of our previous study, which showed that Proteobacteria was the dominant community member [11–15]. Evidence by Miura et al. [16] has shown that Proteobacteria was considered to be the most dominant phylum in MBR that played an important role in the removal of organic matter. The abundance of Proteobacteria obviously increased with operation of the aeration tank. Other dominant phyla included Chloroflexi (4.09%–34.25%), Bacteroidetes (1.56%–21.35%) and Actinobacteria (1.90%–19.37%). Chloroflexi has been detected as the predominant bacteria in anaerobic bioreactors and found to be responsible for the degradation of carbohydrates [17]. Chloroflexi also showed a great decrease with increased running time. Similar results were observed in previous studies that showed Bacteroidetes was the dominant community [18–21]. The anaerobes Bacteroidetes play an important role in the fermentation system to break down macromolecules such as protein, starch, cellulose and fiber [22]. Bacteroidetes also had been found to be responsible for the degradation of complex organic matters [23]. The relative abundance of Bacteroidetes decreased greatly at first, then increased to a relatively high abundance in sample collected on August 13, 2015, then finally stabilized. Actinobacteria play an important ecological role in recycling substances in natural ecosystems and are able to degrade a variety of environmental chemicals [24–26]. Actinobacteria was also the main phyla in a study conducted by Chen et al. [27]. The abundances of the dominant phyla were also different between samples. For example the abundance of Chloroflexi in sample collected on May 25 was significantly higher than that of other samples. Planctomycetes was the dominant route for ammonia removal in UASB reactor [28]. The relative abundance of anammox bacteria Planctomycetes increased with operation of the aeration tank and became one of the major phyla. Huang et al. [29] found that Planctomycetes was the major phylum in soil samples. Across different activated sludge samples, the relative abundances of some bacterial phyla, such as Nitrospirae and Chlamydiae, varied significantly. For instance, the relative abundance of the phylum Nitrospirae in sample collected on November 17, 2015 was significantly higher than in other samples. In contrast, the relative abundance of Firmicutes did not change significantly throughout the experimental period. Moreover, Firmicutes and Gemmatimonadetes have lower abundance in all activated sludge samples. This finding differed from those of previous studies, which showed Firmicutes and Gemmatimonadetes were the dominant phyla [30–32]. Cyanobacteria was a rare group in this study. Most Cyanobacteria can produce extracellular polymeric substances, mainly polysaccharide, which could adsorb heavy metals dispersed in the environment [33,34]. Planctomycetes and Nitrospirae increased greatly with operation time. Proteobacteria, Chloroflexi, Bacteroidetes and Actinobacteria are predominant phyla in the activated sludge.

Fig 7. Abundances of different phyla in 12 activated sludge samples.

The abundance is presented in terms of the percentage of total effective bacterial sequences in a sample.

The distribution of sequences at the genus level in each sample is shown in Fig 8. The top 10 genera of each sample were selected, while all remaining genera were categorized as “other.” Anaerolineaceae_uncultured was the major genus in sample collected on May 25, 2015, while Novosphingobium and Nitrospira were dominant in most samples, especially during the latter portion of the experiment. These findings are consistent with those of previous studies that showed Nitrospira was the major phylum in the recirculating aquaculture system [35]. Nitrosomonas could be a minor component in activated sludge samples. Nitrosomonas are regarded as the dominant ammonia-oxidizing bacteria, while Nitrospira are the main nitrite-oxidizing bacteria in wastewater treatment plants [36]. The bacterial community composition showed little difference as the operating time increased in this study. Anaerolineaceae_uncultured and Ottowia decreased significantly as operating time increased. Coxiellaceae_uncultured was undetected in most samples. Mycobacterium was a minor component in most activated sludge samples, which was much different from the results of a previous study [37]. Some species of Blastocatella are aerobic, chemoorganotrophic organisms with a strictly respiratory type of metabolism [38]. The sulphur oxidizing bacteria genus Thiothrix, which increased significantly during the operation period, has also been found in aerobic granules incubated in brewery wastewater [39]. The relative abundance of Caldilineaceae_uncultured, Saprospiraceae_uncultured and Anaerolineaceae_uncultured decreased significantly as operation time increased, while Nitrospira increased greatly with increased operation time. OM190_norank did not change significantly in the first 8 samples, while it increased after October 15, 2015.

Fig 8. Taxonomic compositions of bacterial communities at the genus level in each sample retrieved from Illumina MiSeq pyrosequencing.

The 10 most abundant genera in each sample were selected (a total of 47 genera for all 12 samples), and their abundances were compared with those in other samples (Fig 9). The bacterial population at the genus level differed significantly among samples from different time points. Most genera comprised a low proportion of the activated sludge samples. However, Anaerolineaceae_unclassified and Anaerolineaceae_uncultured were found in relatively high abundance in sample collected on May 25, 2015. Moreover, Caldilineaceae_uncultured had higher relative abundance in samples collected on May 25 and June 5, 2015. Saprospiraceae_uncultured had higher abundance in sample 20150525 than in other samples. OM190_norank was more abundant in samples collected on November 17 and 20 and December 12, 2015. In addition, OM190_norank clearly increased over the eight-month experimental period. JG30-KF-CM45_norank was abundant in samples collected on June 25 and July 29, 2015. Nitrospira had relatively high abundance in most samples. The relative abundance of PHOS-HE51_norank on June 5, 2015 was higher than that of other samples. Saccharibacteria_norank showed high relative abundance on June 5 and July 29, 2015.

Fig 9. Heat map of top 10 genera in each sample.

A total of 47 genera were selected from 12 samples. The color intensity in each panel shows the percentage of a genus in a sample based on the color key at the bottom of the figure.

Multiple samples similarity tree analysis was used to identify similarities and differences among the 12 bacterial community structures (Fig 10). Overall, samples collected from May, June and July formed one cluster, while those from August, September, October, November and December formed another cluster. The samples which were clustered together also indicated a more similar community structure between these samples. Additionally, the two clusters were well separated from each other, suggesting clear distinctions in the bacterial community structure of the two clusters.

Fig 10. Multiple samples similarity tree.

Principal component analysis (PCA) was conducted using high-throughput sequencing data to further corroborate the results of DNA fingerprinting [21] (Fig 11). The first axis explained 28.05% of the cumulative variance of species, while the second axis explained 19.99%. Samples collected on August 13 and 28, 2015 were closely related to sample collected on September 9, 2015, while sample collected on September 25 were closely related to sample collected on October 15 and those obtained on November 17, November 20 and December 15 clustered together.

Fig 11. Principal component analysis (PCA) of bacterial communities from samples based on sequencing of Illumina MiSeq.

A canonical correspondence analysis (CCA) was used to illustrate relationship among operational conditions (S1 Table) and bacterial community (Fig 12). Based on the length of connecting wire, temperature, DO, pH and SO₄^2- were found to be significantly correlated with the observed microbial composition both in phylum and genus level. In phylum level, Proteobacteria was positively correlated with DO and pH, and negatively with temperature. Phylum Planctomycetes, Nitrospirae. Firmicute and Gemmatimonadetes positively correlated with DO and pH, and SO₄^2- and negatively with temperature. Chloroflexi positively correlated with temperature, and negatively with SO₄^2-,pH and DO. Bacteroidetes positively correlated with pH, and negatively with DO and SO₄^2-. Actinobacteria positively correlated with temperature and SO₄^2-, and negatively with pH. Besides, Chlamydiae positively correlated with temperature, SO₄^2- and DO, and negatively with pH. As shown in Fig 12B, the abundance of Nitrospira (G1) positively correlated with pH and DO, and negatively with SO₄^2- and temperature. The correlation of Caldilineaceae_uncultured (G2) was similar to Saccharibacteria_norank (G3) and PHOS-HE51_norank (G28),with positive correlation with temperature and SO₄^2-, and negative correlation with pH and DO. Abundance of JG30-KF-CM45_norank (G11) positively correlated with temperature,SO₄^2- and DO, and negatively with pH, while abundance of Thiothrix (G17) had the opposite result. Genus Novosphingobium (G4) positively correlated with DO and SO₄^2-. Genus Anaerolineaceae_uncultured (G5) positively correlated with temperature, and negatively with SO₄^2-,pH and DO. The correlation of genus OM190_norank (G8) and Saprospiraceae_uncultured (G9) was similar, with positive correlation with pH, and negative correlation with temperature. Besides, Saprospiraceae_uncultured (G9) and Anaerolineaceae_unclassified (G41) also negatively with SO₄^2- and DO.

Fig 12. Correspondence canonical analysis (CCA) diagrams with bacterial community structure and operational conditions.

(a) at phylum level (b) at genus level.

Conclusion

The combined processes of multi-stage anaerobic bioreactor + micro-aerobic + anoxic/aeration + biological contact oxidation were used to treat gibberellin wastewater. The A/O process led to good removal of COD, NH₃-N and TP. Proteobacteria was the dominant phylum in all samples, followed by Chloroflexi, Bacteroidetes and Actinobacteria. Novosphingobium and Nitrospira were dominant genus in most samples. Illumina MiSeq sequencing showed that bacterial community structure and diversity changed with DO, temperature and SO₄^2-.

Supporting information

S1 Table. Operational conditions in 12 samples.

(PDF)

Data Availability

All raw reads files are available from the NCBI Sequence Read Archive (SRA) database (https://www.ncbi.nlm.nih.gov/sra, accession number SRP077353).

Funding Statement

This work was supported by Implement Project of Science and Technology of Colleges and Universities in Jiangxi Province, grant number KJLD13007 to Erming Ouyang (website: http://www.jxedu.gov.cn/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.