Formation and characteristics of filamentous granular sludge

In the early 1990s, the phenomenon of aerobic granular sludge (AGS) was first reported (Mishima & Nakamura 1991) and rapidly become the research hotspot in the area of wastewater treatment (McHugh et al. 2003). Compared with conventional activated sludge, AGS has good settling performance, dense structure and high concentration of biomass (Da Motta et al. 2001; Lee et al. 2010). In recent years, many researchers have successfully cultivated aerobic granular sludge in sequencing batch reactors (SBRs) (Adav et al. 2008).

Up to now, the mechanism of occurrence of sludge bulking has not been clear and thus control of sludge bulking is still a big challenge for the conventional activated sludge process. In fact, about 95% of sludge bulking is caused by the overgrowth of filamentous bacteria (Martins et al. 2004). Once sludge bulking caused by overgrowth of filamentous bacteria occurs in the activated sludge process, a large amount of activated sludge will be almost washed out in a short time from the secondary settling tank due to the low settleability of flocs. Normally, the conventional activated sludge system could be destroyed and would take a long time to be restored if the sludge bulking could not be controlled and solved in a timely way (Martins et al. 2004; Liu & Liu 2006). Previous studies showed that the overgrowth of filamentous bacteria was caused by low dissolved oxygen, low F/M ratio, etc. (Madoni et al. 2000). Until now, the chemical methods have often been utilized to control sludge bulking. For example, chlorination and ozonation were applied to inhibit the overgrowth of filamentous bacteria in real engineering (Martins et al. 2004). However, filamentous bacteria, as a common microorganism in activated sludge systems were believed to play a positive role in the formation of a symbiotic microbial ecosystem (Peng et al. 1999; Schuler & Jassby 2007; Wu et al. 2019). Moreover, the presence of some filamentous microorganisms was proved to efficiently remove some pollutants from sewage. For example, some fungi with a filamentous structure can be effectively used for the degradation of cellulose in the treatment of papermaking wastewater (Malaviya & Rathore 2007). At the same time, filamentous bacteria with a high specific surface area could have a high adsorption capacity for heavy metal toxicity. Hence, filamentous microorganisms were often used for heavy metal-polluted wastewater treatment (Yin et al. 2019).

Additionally, filamentous bacteria were also observed in the surface of matured aerobic granular sludge cultured with various organic substrates. Some reports suggested that filamentous bacteria intertwined with each other to form the particle skeleton of AGS (Schwarzenbeck et al. 2005). Furthermore, the existence and growth of filamentous microorganisms was believed to be the key stage during the formation of AGS (Liu & Liu 2006). However, there were few reports on the interaction between filamentous microorganisms and AGS. With the consideration that filamentous microorganisms could be cultivated on the surface of aerobic granular sludge, we tried to firstly enrich the quantity of filamentous microorganisms in the aerobic granular sludge system. The SBR process with A/O/A/O/A/O and three-stage feeding strategy was carried out under reduced aeration, which could maintain the initial influent and dissolved oxygen (DO) at a relatively low concentration. Additionally, sulfamethoxazole (SMZ) was added in the AGS system at the end of the experiment because SMZ is frequently detected in livestock wastewater. Therefore, we focus on: (1) the possibility of formation and stability of the filamentous granular sludge (FGS) system; (2) the performance of the FGS system for the removal of nutrients and organic matter; (3) the evolution of microbial diversity without/with the addition of SMZ.

The inoculated anaerobic granular sludge was extracted from the upflow anaerobic sludge blanket (UASB) system at the local starch wastewater treatment plant. 200 mL of anaerobic granular sludge was washed three times with tap water and then added into the SBR reactor. As shown in Figure 1(a), the SBR reactor (working volume = 2 L) was controlled at 25 °C. As shown in Table 1, the composition of the feeding includes organic carbon sources (glucose, potassium acetate and sodium propionate, chemical oxygen demand (COD) = 1,000 mg/L), ammonium chloride (NH₄⁺ = 50 mg N/L) and phosphate (PO₄³⁻ = 20 mg P/L). From Day-101, the influent phosphate decreased from 20 to 10 mg P/L. As shown in Figure 1(b), the SBR procedure (the single cycle time = 6 hours) includes three internal cycles to achieve three times feeding and alternate anaerobic/oxic processes. The aeration flowrate is 0.1 L/min with relatively low specific air velocity (SAV = 0.085 cm s⁻¹). A small quantity of sulfamethoxazole (SMZ = 1 mg/L) was added in the feeding from Day-201.

The water samples in the experiment were filtered by 0.45 μm filter membrane and immediately stored at 4 °C. Mixed liquid suspended solids (MLSS), mixed liquid volatile suspended solids (MLVSS), COD, total organic carbon (TOC), nitrogen forms (NH₄⁺, NO₂⁻, and NO₃⁻), and phosphate were analyzed according to standard methods (APHA 2002). Sludge morphology was analyzed by optical microscope. The characteristics of matured single granules such as the density and the settling velocity were measured through a series of tests using NaCl solution with six concentrations (0%, 1%, 2%, 3%, 4% and 5%) as in a previous study (Lang et al. 2015).

The pretreatment of solid microorganism samples included: (1) adding 2.5% glutaraldehyde (pH = 6.8) and fixing it in a refrigerator for 1.5 hours at 4 °C; (2) washing it three times with phosphoric acid buffer solution (the total concentration of KH₂PO₄ and K₂HPO₄ was 0.1 mol/L, pH = 6.8); (3) dehydrating it three times with 50%, 70%, 80%, 90%, 100% ethanol for 10–15 minutes each time respectively (Sun et al. 2019). Finally, the observation of the sample was analyzed by scanning electron microscope (Nova nanosem450, USA).

The microbial DNA was extracted from the SBR reactor on Day-1, Day-100, Day-200 using the E.Z.N.A.^® soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's protocols. The final DNA concentration and purification was determined by NanoDrop 2,000 UV-vis spectrophotometer (Thermo Scientific, Wilmington, USA). The high-throughput sequencing was performed by Majorbio Co., Ltd in Shanghai using the MiSeq PE300 platform (Illumina, USA). The V3-V4 hypervariable regions of the bacteria 16S rRNA gene were amplified with primers 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′), while the fungi 18S rRNA gene was performed with the primers SSU0817F (5′-TTAGCATGGAATAATRRAATAGGA-3′) and 1196R (5′-TCTGGACCTGGTGAGTTTCC-3′).

Raw fastq files were demultiplexed, quality-filtered by Trimmomatic and merged by FLASH with the following criteria: (i) the reads were truncated at any site receiving an average quality score less than 20 over a 50 bp sliding window; (ii) primers were exactly matched, allowing 2 nucleotide mismatching, and reads containing ambiguous bases were removed; (iii) sequences with an overlap longer than 10 bp were merged according to their overlap sequence. Operational taxonomic units (OTUs) were clustered with 97% similarity cut off using UPARSE (http://drive5.com/uparse/) and chimeric sequences were identified and removed using UCHIME.

DNA samples were extracted from the SBR system using a fast DNA kit for Soil (MP Biomedicals, USA) according to the manufacturer's instructions. Standard polymerase chain reaction (PCR) was performed to determine the presence of 2 sulfonamide resistances genes (sul1 and sul2) (Hoa et al. 2008). Then, these two ARGs were analyzed further by qPCR using lightcycler480 Real-Time PCR detection system (Roche). The reaction mixture system comprised a volume of 10 μL containing: 5 μL SybrGreen qPCR Master Mix (2X); 0.2 μL Primer F (10 μM); 0.2 μL Primer R (10 μM); 3.6 μL ddH2O; 1.0 μL Template(DNA). Light Cycler 480 II steps included: (1) Initial Steps at 95 °C for 3 min; (2) Melt at 95 °C for 5s; (3) Anneal/Extend at 60 °C for 30s; (4) Plate read, where steps (2) to (3) were repeated 45 times. After completing the above steps, the sampled 96/384-well plate was placed in a Light Cycler 480 II (Roche Roche) for reaction. The absolute abundance (AA) of the ARGs was expressed as copies/mg of sludge.

As shown in Figure 2(a), the COD and TOC removal efficiency of the SBR system fluctuated during the first phase. When the phosphate concentration decreased from 20 to 10 mg/L, the removal of COD and TOC tended to be improved. At phase III, with the addition of SMZ, the degradation of organics in the system became stable. In fact, the filamentous granular sludge was firstly observed at the end of phase I. As shown in Figure 2(b), the removal efficiency of total nitrogen fluctuated from 60% to 75% at phase I. From phase II, the effluent ammonium increased probably due to the rapid growth of filamentous bacteria on the surface of the granular sludge. The filamentous bacteria have a higher ability to capture oxygen compared with ammonium oxidizing bacteria (AOB) under low concentration of DO (0.5–1.0 mg/L) (Wu et al. 2019). At the same time, the phosphorus removal gradually increased, the removal efficiency of phosphorus was over 90% at the end of phase II and at the beginning of phase III. As shown in Figure 2(c), the SMZ in the effluent was stabilized at ca. 0.8 mg/L during phase III. Figure 2(d) showed that the MLSS fluctuated between 2 and 5 g/L in phase I and II. Once the SMZ was added in the system, the MLSS decreased gradually from 2 to 1 g/L, suggesting SMZ has a negative effect on the quantity of microorganisms in the SBR system.

At the end of phase III, the dynamic performances of the FGS system were tested on Day-249, as shown in Figure 3. After the first feeding, COD and ammonium increased and stabilized at around 280 and 12 mg/L, while phosphate showed a very slight upward trend during the first anaerobic stage. In the following aerobic stage, the COD rapidly decreased from 385 to 65 mg/L, while a small part of ammonium and phosphate was removed. The similar variation of COD, ammonium and phosphate was observed in the second and third anaerobic and aerobic stages. However, nitrite and nitrate were always at very low level (<0.5 mg/L) during the whole cycle. In fact, the dissolved oxygen (DO) was kept at between 0.5 and 2.5 mg/L in the aerobic stage under low aeration (0.1 L/min). In fact, most of the DO could be consumed by the filamentous bacteria on the surface of the aerobic granules, which possibly inhibited the growth of ammonium-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB). Moreover, the excessive EPS produced by the filamentous bacteria could change the viscosity of water and thus hinder the transfer of DO in the aerobic stage (Wu et al. 2019). Hence, the produced nitrite or nitrate should be utilized as electron acceptors by denitrifiers inside the aerobic granule during the aerobic stage. Additionally, the influent sulfate (around 300 mg/L) was much higher than that (between 40 and 150 mg/L) in the FGS system during the whole cycle, suggesting sulfate was consumed as electron acceptors for the oxidation of organic pollutants by some sulfate-reducing bacteria (SRBs). In fact, Desulfobulbus, as one of typical SRBs, was confirmed by the microbial diversity analysis (Yan et al. 2018).

The morphology of FGS was characterized via SEM and optical microscope observation, as shown in Figure 4 Filamentous granular sludge was between 0.64 mm and 2 mm. Filamentous bacteria were growing on the surface of granular sludge and the diameter of the filamentous bacteria was 0.4–0.7 μm and the length of the filamentous bacteria was 1–2 mm. A large number of filamentous microorganisms was intertwined on the surface of the granular sludge. Moreover, the core of FGS was very dense with a clear and smooth edge after a cross-section of FGS, while a large amount of white filaments surrounded on the outside of FGS.

The physical settleability of the single filamentous granule was tested in a series of batch experiments with different concentrations of NaCl solution (Figure 5). At 0–1% NaCl concentration, the FGS settled immediately because the density of the NaCl solution is lower than that of FGS granular sludge. With the increase in the NaCl concentration, the settling velocity of FGS correspondingly decreased. When the concentration of NaCl was higher than 1.5%, the filamentous granule was firstly suspended and then settled, which means the initial density of FGS is lower than that of the NaCl solution. The density of FGS gradually increased because of the diffusion of NaCl into the FGS. The concentration of NaCl in the aqueous solution was higher, which resulted in the diffusion time becoming longer, and the macro behavior was the suspension time became longer. At the same time, the settling velocity of FGS was constant at around 18 m/h. Thus, the density of FGS should be in the range of 1.007–1.015 g/cm³.

As shown in Figure 6(a), both Thiothrix and Trichococcus as typical filamentous species were detected and dominant in FGS without the addition of SMZ. When SMZ was added in the feeding, the abundance of Thiothrix increased from 15% to 34.69%, while Trichococcus decreased from 23.2% to 3.18%. At the same time, a diversity of fungi were also detected in FGS system, as shown in Figure 6(b).

As shown in Figure 6(a), the most dominant groups of microorganisms in FGS are Thiothrix and Trichococcus. Thiothrix and Trichococcus are typical filamentous species (Mielczarek et al. 2012). With the addition of SMZ in the feeding, the dominant strain of filamentous bacteria changed from Trichococcus to Thiothrix, which was probably due to their various characters (ex. Thiothrix is Gram negative while Trichococcus is Gram positive) and different resistance with SMZ. Thiothrix is characterized by filament length less than 500 μm, round ended rod and cylindrical cell shape, and colony size of 1–2 mm. It was also confirmed that the Thiothrix is controlled by anaerobic and oxic stages (Deepnarain et al. 2019). Currently, Trichococcus genus includes 5 species: Trichococcus flocculiformis, T. palustris, T. pasteurii, T. collinsii and T. patagoniensis. All Trichococcus species share a high similarity of their 16S rRNA gene sequences. They are described as facultative anaerobes able to create redox conditions and to reduce resazurin in aerobic media during growth. They are all oxidase and catalase negative, and can grow with a wide variety of sugars and other substrates (Vasmara et al. 2018). As shown in Figure 6(c), after the addition of SMZ in the feeding, the content of sulfonamide resistance genes (sul1 and sul2) in sludge also increased, especially for sul1. SMZ did increase the resistance genes in the FGS system, which was consistent with the previous study (Ren et al. 2019).

The main conclusions were included as follows: (1) FGS was successfully cultivated with alternate anaerobic/oxic process and multi-feeding strategy under a low aeration rate after 2 months' operation; (2) the FGS system could maintain a relatively stable removal efficiency for the carbon and phosphorus, while nitrogen removal was inhibited; (3) both Thiothrix and Trichococcus were the main species of filamentous bacteria detected in the FGS system, while the dominant filamentous bacteria changed from Trichococcus to Thiothrix in the presence of SMZ; (4) the role of filamentous microorganisms in FGS should be investigated in future work, especially on the treatment of antibiotics and refractory organics.

This work was financed by the National Natural Science Foundation of China (No. 21107100) and the Outstanding Young Talent Research Fund of Zhengzhou University (No. 1421324067).