Microbial population changes and metabolic shift of candidatus accumulibacter under low temperature and limiting polyphosphate

Enhanced biological phosphorus removal (EBPR) is one of the main processes for phosphorus removal in the activated sludge sewage treatment system. Polyphosphate accumulating organisms (PAOs) are the functional microorganism in the EBPR process. The process utilizes the metabolic characteristics of PAOs, which can release phosphate under anaerobic conditions and store polyphosphate (Poly-P) under aerobic or anoxic conditions. Through the discharge of Poly-P-enriched sludge, the removal of excess phosphorus in sewage is realized. Therefore, the activated sludge in the EBPR system enters the anaerobic environment firstly and then enters the aerobic or anoxic environment, alternately circulating. In the anaerobic phase, the Poly-P hydrolysis and glycogen degradation processes respectively provide ATP and reducing equivalent (nicotinamide adenine dinucleotide, NADH) for PAOs uptake of volatile fatty acid (VFA), and then VFA is transformed into polyhydroxyalkanoates (PHA) and stored in the cells (Mino et al. 1998). During the subsequent aerobic phase, PAOs degrade PHA to produce ATP, which is used to restore glycogen and Poly-P. In the EBPR system, glycogen accumulating organisms (GAOs) often exist, which have similar metabolic behavior to PAOs (Carvalheira et al. 2014). However, the GAOs only take glycogen as the source of ATP and NADH, and this process does not involve the degradation of Poly-P. Therefore, the proliferation of GAOs has adverse effects on the uptake of VFA by PAOs under anaerobic conditions. PAOs and GAOs often coexist in EBPR systems and the highly enriched PAOs culture is expected. Hence the competition between PAOs and GAOs has been paid extensive attention (Winkler et al. 2011; Shen & Zhou 2016; Wang et al. 2021; Ni et al. 2022). Schuler & Jenkins (2003) defined the metabolic patterns of PAOs and GAOs as polyphosphate accumulating metabolism (PAM) and glycogen accumulating metabolism (GAM), respectively.

Recent studies have found that PAOs showed metabolic shift linking GAO behaviors when intracellular Poly-P was in a limiting level (Zhang et al. 2015; Welles et al. 2016a; Acevedo et al. 2017), with VFA taken-up and glycogen consumed but without the release and absorption of phosphate. Limiting Poly-P means that, in the anaerobic phase, Poly-P is not sufficient to generate ATP for PAOs to uptake VFAs from liquid. For example, decreasing the influent P concentration (Acevedo et al. 2017) or increasing the Ca²⁺ concentration (Zhang et al. 2015) both caused the reduction of the available Poly-P in PAOs cells, resulting in the metabolic shift of PAOs from PAM to GAM. The metabolic process of PAOs can be indicated by a number of stoichiometric ratios, and the changes in stoichiometry could directly reflect the metabolic shift of PAOs from PAM to GAM. The typical GAM showed that the P_release/HAc_uptake ratio of PAOs was almost zero, accompanied by higher PHA_synthesized/HAc_uptake ratio and Gly_degraded/HAc_uptake ratio and so on. Combined with microbial population analysis, it was found that the metabolic characteristics of PAO clades were different under the limiting Poly-P conditions. Candidatus Accumulibacter phosphatis (hereinafter referred to as Accumulibacter or Acc) is currently recognized as playing a leading role in PAOs, which can be divided into two main clades of Accumulibacter I type and II type (as Acc I and Acc II) based on gene 16SrRNA and polyphosphate kinase genes (ppk1) (Peterson et al. 2008; Kolakovic et al. 2021).

When intracellular Poly-P was depleted in the short term, the metabolic pathway of Accumulibacter would be changed. At the same time, the population structure of PAOs would be affected by limiting Poly-P level. However, there are different observations about PAOs microbial structure variation in literature, when Poly-P storage was kept in a limited level. Acevedo et al. found that Acc clades changed from Acc I to Acc II as Poly-P decreased (Acevedo et al. 2012). Welles et al. (2015) observed that the HAc-uptake of Acc II was higher than Acc I when Accumulibacter metabolized by GAM. It seemed to suggest that Acc II exhibited higher adaptability than Acc I under Poly-P deficiency. However, some scholars found that GAOs did not proliferate and Acc I could maintain GAM in a long period under lower Poly-P level (Acevedo et al. 2017). Welles et al. (2016b) detected Acc II and Competibacter coexisted in a low-phosphorus enrichment system, indicating that Acc II could behave as GAM for a long time under limiting Poly-P. Even in some studies, Accumulibacter was washed out and GAOs became prevailing with Poly-P restricted (Lv et al. 2014). Therefore, further researches are needed to explore the Acc clades dynamics under limiting Poly-P.

Temperature and carbon source are known as two key factors for EBPR performance and PAOs microbial community structure (Wang et al. 2020a; Ni et al. 2021). Temperature can influence the metabolic pathway of PAOs and competition between PAOs and GAOs (Winkler et al. 2011). An EBPR system operating at SRT = 5 d showed that the phosphorus removal activity was lost when the temperature dropped to 10 °C (Chan et al. 2020). Nevertheless, efficient EBPR was successfully achieved at 10 ± 1 °C (Wang et al. 2020b). It was found that temperature was also an important factor affecting the metabolism characteristics of Accumulibacter with Poly-P limitation. Under the restriction of Poly-P, Acc I did not seem to switch to GAM at low temperature (Tian et al. 2013). However, there are few studies on the metabolism shift of Accumulibacter at low temperature, especially for a long time. Moreover, the longer-term effect of low temperature on the population structure and metabolic behavior of Accumulibacter is not well understood.

The main objective of this research is to investigate the microbial community dynamics of Accumulibacter at low temperature and the metabolic shift to limiting Poly-P. The results of this research could help for in-depth understanding of Accumulibacter metabolism and activity variation at lower temperature under Poly-P limitation.

Two identical SBR reactors (Figure 1) with 12 L capacity were operated for 60 days at 10 ± 1 °C, recorded as SBR-1 and SBR-2 respectively. The temperature was controlled by a water bath. The two systems have operated for 180 days at 20 ± 1 °C and enriched with Accumulibacter, which were fed with sodium acetate (SBR-1, HAc system) and sodium propionate (SBR-2, HPr system) as the sole carbon source, respectively. Two constant temperature control devices were set up to ensure that the temperature in both reactors was 10 ± 1 °C. The sludge was discharged at the end of every aerobic phase, controlling the sludge retention time (SRT) at 12 days. The dissolved oxygen (DO) in the aerobic phase varied between 2.5 and 3.0 mg/L. The initial pH for each cycle was 7.5 ± 1 and ranged from 7.0 to 8.8 during each cycle. The reactors were operated for four cycles per day with a weekly period of 6 h including 5 min for filling, 90 min for stirring to simulate the anaerobic condition, 150 min for aerating, 80 min for settling and withdrawing, and the rest of the time for static. Each cycle was flooded with 6 L synthesized wastewater, which were the same except for the carbon sources. The main influent components were as follows (Wang et al. 2020b): 400 mg COD/L (512.5 mg/L NaAc for SBR-1 and 342.86 mg/L NaPr for SBR-2), 25 mg P/L (109.68 mg/L KH₂PO₄), 20 mg N/L (76.43 mg/L NH₄Cl), 153.75 mg/L MgSO₄·7H₂O and 15 mg/L Ca²⁺ supplied by tap water. In addition, 0.5 mL trace element solution was added to the influent per liter, which was prepared as described by Smolders et al. (1994). The trace element solution included (in g/L): KI (0.18), H₃BO₃ (0.15), CoCl₂·6H₂O (0.15), MnCl₂·4H₂O (0.12), Na₂MoO₄·2 H₂O (0.06), ZnSO₄·7H₂O (0.12), CuSO₄·5H₂O (0.03), FeCl₃·6H₂O (1.50) and EDTA (6.00).

DO and pH were conducted to ensure SBR-1 and SBR-2 in the stable conditions after changing the temperature during the operation. The other routine monitoring parameters included orthophosphate (PO₄³⁻-P), COD, mixed liquor suspended solids (MLSS), mixed liquor volatile suspended solids (MLVSS). The reactor operated for 60 days (5 SRT) at 10 ± 1 °C. It was considered that the period was long enough to reflect the influence of low temperature on the microbial community structures and metabolic shift characteristics of the EBPR systems. After stable operation for 60 days, the population identification and cycle metabolic characterization (Cycle 1) for SBR-1 and SBR-2 were carried out, calculating the related biological kinetic rates and stoichiometry during the conversions. Cycle 1 represented the metabolic characteristics of SBR-1 and SBR-2 under high Poly-P condition at 10 ± 1 °C. And then an extraction process was added at the end of the anaerobic phase during each operation cycle. After the reactors were static for 30 min, the supernatant was removed and the synthesized wastewater was replaced without VFA and P. The extraction process was followed by Acevedo et al. (2017). After five extraction cycles, Cycle 2 was conducted to characterize the metabolic characteristics of Accumulibacter under short term limiting Poly-P condition. In Cycle 2, there is no phosphorus in influent and other components were the same as before. The measured parameters included PO₄³⁻-P, VFA, PHA, Poly-P, glycogen during the whole of the cycles. In addition, MLSS and MLVSS at the initial and end of the aerobic phase and end of the anaerobic phase were determined.

The determinations of MLSS, MLVSS, PO₄³⁻-P and COD were carried out according to Standard Methods (APHA 1998). PHA were determined by gas chromatography, according to the methods of Oehmen et al. (2005). The method of glycogen determination was modified as described in Oehmen et al. (2004). A certain amount of freeze-dried solid was taken and digested for 5 h at 100 °C after adding 5 mL 0.6 mol/L HCl solution. The content of Poly-P was calculated on the basis of TP and MLVSS (Acevedo et al. 2015). The DO, pH, and temperature were measured using DO and pH probes (Multi 340i, WTW company, Germany).

Fluorescence in situ hybridization (FISH) technique was used to identify the population structures of Accumulibacter and GAOs in SBR-1 and SBR-2, following the procedure described by Winkler et al. (2011). The population structures of inoculated sludge were also analyzed to investigate the influence of temperature on the population structure of the EBPR system. The ‘all bacteria’ corresponded to the EUB mix probe, which was a mixture of the EUB, EUB I and EUB II probes. Acc I and Acc II were targeted by the probes Acc-1-444 and Acc-2-444, respectively. The GAO mix probe (mixture of probes GAOQ431 and GAOQ989) targeted ‘Candidatus Competibacter phosphatis’ (belonging to GAOs). The formamide concentration of all probes was 35%.

After hybridization, 10 randomly selected fluorescent images were analyzed by images software. The abundances of Acc I, Acc II and GAOs were expressed as the proportion of relative probes to the all bacteria probe based on the optical density calculation. The standard error of the mean (SEM) is calculated as the standard deviation divided by the square root of the number of images. The particle size distribution was identified, combined the wet sieving method (Laguna et al. 1999).

The SBR-1 and SBR-2 reactors were operated with sodium acetate and sodium propionate at 10 °C for 60 days. Figure 2 represented the influent and effluent COD, PO₄³⁻-P, MLSS and VSS/SS in two systems.

It was observed that the P removal performance tended to a stable state when SBR-1 and SBR-2 operated for 8 days. The average PO₄³⁻-P concentration of effluent was 0.887 and 0.148 mg/L, and COD concentration was 0.254 and 0.414 C mmol/L, respectively. During the whole operation, the average MLSS of SBR-1 and SBR-2 was 4,200 ± 341 mg/L and 4,500 ± 362 mg/L respectively and VSS/SS kept between 0.60–0.70 mg/mg, indicating that microorganisms in the systems could maintain good microbial activity and phosphorus removal ability at 10 ± 1 °C.

Low temperature environment could affect the specific microorganisms with phosphorus removal ability (Li et al. 2019). FISH was used to estimate the abundances of Accumulibacter and GAOs (Figure 3). The abundances of Acc I, Acc II and GAOs were represented as the proportions of relative probes to the all bacteria probe, as shown in Table 1. After running for 60 days at 10 ± 1 °C, the abundance of Acc I and Acc II in SBR-1 were 10 ± 2% and 16 ± 4%, and that in SBR-2 were 49 ± 11% and 26 ± 3% respectively. Compared with the population structure in the seed sludge, Accumulibacter clades decreased in SBR-1, while Accumulibacter in SBR-2 still remained at the original level and Acc I become the dominant population in the HPr system. Tian et al. (2013) only detected the presence of Acc I in Accumulibacter culture (NaAc·3H₂O as carbon source, SRT = 16d, 36d at different phases) enriched at 10 ± 1 °C. Combined with the results of this study, it seemed that low temperature and HPr with a SRT of 12d were more beneficial to the enrichment of Acc I. Previous studies suggested that low temperature would reduce the aerobic metabolic rate of PAOs and then influence minimal SRT for PAOs growth (Chan et al. 2020). Therefore, low temperature and HPr can be used as the key factors to enrich Acc I with a proper SRT. The abundance of GAOs became lower at 10 ± 1 °C, which verified the previous opinion that low temperature was not conducive to GAOs (Tian et al. 2013). Besides the effect of GAOs proliferation could be neglected when exploring the characteristics of GAM shift in the systems.

In addition, it was found that the biomass morphology of the systems changed significantly during the low-temperature operation, gradually forming granular sludge. Figure 4(a) showed the particle size distribution of the sludge particles in SBR-1 and SBR-2. The particle size of the sludge in SBR-1 was relatively small, and the maximum particle size was between 1.0 and 1.6 mm, accounting for 16.1% (/TSS). The proportion of particles of 0.5–1.0 mm was highest, accounting for 53%. However, the distribution of sludge particles in SBR-2 was inhomogeneous, and the diameter of certain sludge particles could reach 4.0 mm. Although the proportion of particles below 0.5 mm was highest (46.4%), the particles (38.4%) larger than 1.0 mm increased significantly compared with the original seed sludge in SBR-2. Figure 4(b) showed the sludge particles with diameters of 2.5–4.0 mm in SBR-2. It can be seen that during the low-temperature operation, more large size sludge particles were formed in SBR-2. Although fine particles in SBR-1 was not as obvious as in SBR-2, the change of the average particle size of SBR-1 in the process could not be ignored. Compared with parent systems, temperature was the crucial factor that was changed in both reactors. Therefore, it can be concluded that low temperature was beneficial to the sludge granulation in the EBPR system, which was consistent with previous researches (Tian et al. 2017; Wang et al. 2020b). It probably resulted from the formation of alginate-like exopolysaccharides (ALE) at low temperature, which could promote the granulation of sludge (Lin et al. 2013; Tian et al. 2017). Interestingly, there were some differences of sludge granulation in the two EBPR systems at low temperature. The sludge granule size in HAc system was more homogeneous than that in HPr system, while the proportion of larger size (>1.0 mm) granules in HPr system was higher than that in HAc system. The different organic substrate was probable the main reason which led to different granulation, and further research is needed to investigate the mechanisms (Cai et al. 2019). The proportion of particles below 0.5 mm was highest (46.4%) in SBR 2, and that in 0.5–1.0 mm was highest (53%) in SBR 1, which indicated that the granule in SBR1 was more prominent than in SBR 2. In this study, larger size granules were achieved by feeding with HPr, which was a little bit different from the observation of Cai et al. (2016). Cai et al. (2016) observed that the larger and more stable granular sludge was achieved by mainly feeding HAc. This may be due to the different temperatures (the operating temperature of Cai et al. (2016) was 25 ± 2 °C). And the other possibility was that the Accumulibacter abundance of the HPr system was significantly higher than that of the HAc system in this study, which was more conducive to the formation of ALE.

Figure 5 (a) and 5(b) presented the metabolic conversions of SBR-1 and SBR-2 in high Poly-P cycle (Cycle 1) at 10 ± 1 °C. It can be seen that both systems conducted typical PAM. Under the anaerobic phase, Accumulibacter took up VFA to synthesize PHA, with degradation of Poly-P and glycogen. And then during the aerobic phase, Accumulibacter utilized ATP produced by PHA degradation to resynthesize Poly-P and glycogen. However, there were some differences of parameters in the cycles. The metabolism of SBR-1 tended to suspend after 50 min during the anaerobic phase, basically completing the process of phosphate release and HAc uptake. The release of phosphate in the anaerobic phase was 2.897 P mmol/L and the HAc uptake was 3.783 C mmol/L with the rate of 1.184 (C mol/mg VSS·h). The release of phosphate in SBR-2 lasted to the end of the anaerobic phase, and the content of PHA increased continuously. In the anaerobic process, the release of phosphate was 3.408 P mmol/L and the uptake of HPr was 4.953 C mmol/L with the rate of 0.910 (C mol/mg VSS·h). In the anaerobic phase, the degradation of glycogen and Poly-P in SBR-1 were 12.152 C mmol/L and 0.639 P mmol/L respectively and that in SBR-2 were 6.153 C mmol/L and 2.044 P mmol/L. Besides, this may be due to the diverse metabolic pathways of Accumulibacter utilizing different carbon sources, resulting in different components and quantities of PHA. The amount of PHA synthesized in the HPr system was greater. According to the stoichiometry of the anaerobic phase (Table 2), it was found that more glycogen and Poly-P degradation were needed in the HAc system than that in the HPr system for the synthesis of per PHA.

During the aerobic phase, PHA degradation and resynthesis of Poly-P and glycogen could be observed. After 150 min, the Poly-P synthesis in SBR-1 was basically completed, but PHA continued to be consumed with glycogen increasing remarkably. In SBR-2, Poly-P was in an increasing trend, while glycogen synthesized slowly. This phenomenon indicated that Accumulibacter showed a preference for Poly-P synthesis at the aerobic phase. The ATP produced by PHA was firstly used for the synthesis of Poly-P, followed by glycogen regeneration. This was in accordance with Welles et al. (2016a) that Poly-P was the preferred storage polymer for PAOs.

Figure 6(a) and 6(b) show the metabolic conversions of SBR-1 and SBR-2 in limiting the Poly-P cycle (Cycle 2) at 10 ± 1 °C. After the extraction period, the intracellular Poly-P level of Accumulibacter decreased greatly. The P release in SBR-1 stopped at 70 min in the anaerobic phase, with the P release 0.772 P mmol/L and HAc uptake rate 0.883 (C mol/mg VSS·h). HAc was not completely uptaken, but glycogen was no longer consumed. Therefore, the surplus HAc entered the aerobic phase. The P_release/VFA_uptake ratio of the anaerobic phase was 0.286 P mol/C mol, which was lower than that in Cycle 1. At the aerobic stage, Accumulibacter still favored Poly-P storage, increasing the synthesis of glycogen after Poly-P synthesis completed. The glycogen synthesis was 3.947 and 10.851 C mmol/L, respectively. Compared with 12.246 and 10.378 C mmol/L in Cycle 1, the glycogen synthesis of SBR-1 was drastically reduced in Cycle 2. After five extraction cycles, the glycogen level of SBR-1 was reduced from the initial 60 C mmol/L to 15 C mmol/L. On the contrary, under limiting Poly-P, the glycogen content in SBR-2 almost didn't decrease, and the uptake rate of HPr was not slowed down. The uptake of HPr was at a rate of 0.989 (C mol/mg VSS·h) with the P release of 1.143 P mmol/L and P_release/VFA_uptake ratio of 0.259 P mol/C mol. It suggested that Accumulibacter in SBR-1 could not behave as GAM under the limiting Poly-P condition at 10 ± 1 °C, while that in SBR-2 could perform GAM. Previous study concluded that, under P-limiting conditions, both Acc I and II could shift to a GAM, while Acc II has a competitive advantage over Acc I (Welles et al. 2015). Moreover, it was believed that the metabolic pathway of Accumulibacter is dependent on the clade and environmental conditions. In this study, total abundance of Acc in the HPr system (SBR-2) was higher than that in the HAc system (SBR-1) with a similar abundance of Acc II for the two systems. However, the clades of two types (Acc I, II) were not identified in this study. So, probably, there are some clade differences for the two systems.

After five extraction cycles, the intracellular Poly-P of Accumulibacter in SBR-1 and SBR-2 decreased significantly from 0.036 and 0.043 mg P/mg VSS to 0.013 and 0.017 mg P/mg VSS, respectively (Table 3). Poly-P was the main component of sludge ash, which can reflect the proportion change of inorganic suspended solids (ISS) in total suspended solids (TSS). The Poly-P variation was in accordance with the variation of ISS/TSS ratio.

The kinetic rate and stoichiometry during the metabolic process of Accumulibacter can correspond to different metabolic patterns. Therefore, the changes in these stoichiometric parameters can be used as a method for estimating metabolic pathway (PAM or GAM) in an EBPR system. The difference between PAM and GAM is that the energy required for VFA absorption is supplied by Poly-P(PAM) or only by glycogen (GAM). The phosphate was released into the liquid during the Poly-P degradation. Therefore, the change trend of phosphate and glycogen is the key to investigation of the metabolic shift in the system. The metabolic shift of PAM to GAM will reduce the P_release/VFA_uptake ratio. Besides, in the anaerobic phase, the utilization of glycogen will be increased, and more glycogen degradation will provide more ATP and NADH, showing Gly_degraded/VFA_uptake ratio increasing, when PAM shifts to GAM. Combined with the stoichiometric parameters in Table 2, P_release/VFA_uptake ratios of SBR-1 and SBR-2 were reduced from 0.765 and 0.685 P mol/C mol of Cycle 1 to 0.286 and 0.259 P mol/C mol, respectively. From higher to limiting Poly-P condition, Gly_degraded/VFA_uptake ratio of SBR-2 increased from 1.236 to 2.466 C mol/C mol, while that of SBR-1 decreased from 3.125 to 1.528 C mol/C mol. It indicated that under Poly-P limitation, Accumulibacter in SBR-1 cannot get ATP from glycogen degradation.

The studies of metabolic shift for Accumulibacter were usually based on HAc (Acevedo et al. 2012; Acevedo et al. 2015; Welles et al. 2015; Acevedo et al. 2017), accompanied by the higher PHA_synthesized/VFA_uptake ratio. However, in this study PHA_synthesized/VFA_uptake, PHB_synthesized/VFA_uptake and PHV_synthesized/VFA_uptake of the HAc system were reduced under the condition of Poly-P deficiency. The components of PHA were different in the HAc system and HPr system. Researches on metabolic shift using HPr as the carbon source are lacking. Although PHA_synthesized/VFA_uptake ratio of SBR-2 during the anaerobic phase was also decreased in Cycle 2, the stoichiometric ratio of the aerobic phase could still reflect GAM occurrence in aerobic phase, P_uptake/PHA_degraded ratio decreased and Gly_synthesized/PHA_degraded ratio increased in SBR-2. This may be as a result of the low phosphate concentration in the liquid. The ATP used to synthesize Poly-P was reduced, and the excess ATP produced by PHA degradation promoted the synthesis of glycogen. Contrasting with SBR-2, the Gly_synthesized/PHA_degraded ratio in SBR-1 decreased.

The changes mentioned above demonstrated that Accumulibacter in the HAc system cannot behave as GAM under limiting Poly-P condition at 10 ± 1 °C, but that in the HPr system it has the ability to get ATP through GAM, so as to avoid being in a disadvantaged position. Two Accumulibacter types (I and II) coexisted in SBR-1 and SBR-2. The difference between SBR-1 and SBR-2 is the carbon source. Maybe different conversion pathways of carbon sources lead to the distinction in the utilization of glycogen. Another possibility is that Accumulibacter has numerous subclades with diverse metabolic characteristics, and it is unclear whether the different subclades play different roles in the metabolic process (Kolakovic et al. 2021). Further research is required to elucidate that.

After operating for 60 days at 10 ± 1 °C, the abundance of Accumulibacter in the HAc system reduced, while that of Acc I in the HPr system increased significantly, becoming dominant PAOs. Phosphorus removal performance of both HAc and HPr systems was stable and granular sludge formed gradually at 10 ± 1 °C. The sludge granulation in the two systems was different and there were more larger granules in the HPr system. Accumulibacter in the HPr system has the ability to get ATP through GAM, while that in the HAc system cannot do this under limiting Poly-P condition at 10 ± 1 °C. This work provides a deep insight into the metabolic behavior of PAOs under limiting Poly-P at low temperature, which is meaningful for maintaining a stable performance of an EBPR system.

This work was supported by the National Natural Science Foundation of China. (Grant No. 51678388).

All relevant data are included in the paper or its Supplementary Information.