w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/watres

Metabolic versatility in full-scale wastewater treatment plants performing enhanced biological phosphorus removal Ana B. Lanham a, Adrian Oehmen a, Aaron M. Saunders c, Gilda Carvalho a,b,*, Per H. Nielsen c, Maria A.M. Reis a a

REQUIMTE/CQFB, Chemistry Department FCT-UNL, 2829-516 Caparica, Portugal Instituto de Biologia Experimental e Tecnolo´gica (IBET), Av. da Repu´blica (EAN), 2784-505 Oeiras, Portugal c Department of Biotechnology, Chemistry and Environmental Engineering, DK-9000 Aalborg University, Denmark b

article info

abstract

Article history:

This study analysed the enhanced biological phosphorus removal (EBPR) microbial community

Received 3 April 2013

and metabolic performance of five full-scale EBPR systems by using fluorescence in situ

Received in revised form

hybridisation combined with off-line batch tests fed with acetate under anaerobiceaerobic

25 July 2013

conditions. The phosphorus accumulating organisms (PAOs) in all systems were stable and

Accepted 27 August 2013

showed little variability between each plant, while glycogen accumulating organisms (GAOs)

Available online 20 October 2013

were present in two of the plants. The metabolic activity of each sludge showed the frequent involvement of the anaerobic tricarboxylic acid cycle (TCA) in PAO metabolism for the

Keywords:

anaerobic generation of reducing equivalents, in addition to the more frequently reported

Polyphosphate accumulating or-

glycolysis pathway. Metabolic variability in the use of the two pathways was also observed,

ganisms (PAO)

between different systems and in the same system over time. The metabolic dynamics was

Glycogen accumulating organisms

linked to the availability of glycogen, where a higher utilisation of the glycolysis pathway was

(GAO)

observed in the two systems employing side-stream hydrolysis, and the TCA cycle was more

TCA cycle

active in the A2O systems. Full-scale plants that showed higher glycolysis activity also

Glycolysis

exhibited superior P removal performance, suggesting that promotion of the glycolysis

Glycogen

pathway over the TCA cycle could be beneficial towards the optimisation of EBPR systems. ª 2013 Elsevier Ltd. All rights reserved.

Return sludge side-stream hydrolysis (RSS)

1.

Introduction

Enhanced biological phosphorus removal (EBPR) has been widely used in wastewater treatment plants (WWTPs) as an efficient, economical and sustainable way to remove phosphorus from wastewater (Oehmen et al., 2007). The process

relies on alternating anaerobic with aerobic and/or anoxic conditions. This strategy imposes a selection pressure promoting the growth of organisms that are able to internally store carbon- and energy-providing polymers, namely polyphosphate, glycogen and polyhydroxyalkanoate (PHA), through the uptake of carbon substrates, mainly volatile fatty acids

* Corresponding author. Instituto de Biologia Experimental e Tecnolo´gica (IBET), Av. da Repu´blica (EAN), 2784-505 Oeiras, Portugal. Tel.: þ351212948571; fax: þ351212948550. E-mail addresses: [email protected] (A.B. Lanham), [email protected] (A. Oehmen), [email protected] (A.M. Saunders), [email protected] (G. Carvalho), [email protected] (P.H. Nielsen), [email protected] (M.A.M. Reis). 0043-1354/$ e see front matter ª 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2013.08.042

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

(VFA), in anaerobic conditions (Oehmen et al., 2007). Two main groups of organisms have been identified: polyphosphate accumulating organisms (PAOs), characterised by their capacity of internally storing phosphate as polyphosphate, and glycogen accumulating organisms (GAOs), with a similar metabolism as PAOs but without the capacity to accumulate polyphosphate and hence negatively contributing to the overall phosphorus removal performance. Several factors that affect the competition between PAOs and GAOs include the carbon source, pH and temperature (Oehmen et al., 2007). The most well-known and studied PAO has been “Candidatus Accumulibacter phosphatis”, a non-isolated organism (Crocetti et al., 2000). Other putative PAOs have also been reported, including organisms belonging to the Tetrasphaera genus, which were shown to accumulate polyphosphate, although it is still unknown which specific storage polymers are being used for phosphorus uptake (Nguyen et al., 2011). The anaerobic metabolism of Accumulibacter-PAOs, in particular with acetate, has been well established (Oehmen et al., 2007). Acetate is taken up and transformed into PHA by using ATP generated by the hydrolysis of polyphosphate. However, the source of reducing equivalents necessary for the conversion of acetate to PHA has been the subject of controversial discussion. Initial results suggested the anaerobic involvement of the tricarboxylic acid (TCA) cycle (Comeau et al., 1986; Wentzel et al., 1986) while others implicated the utilisation of glycogen as a source for reducing equivalents and energy (Mino et al., 1987). However, several findings have indicated that both pathways are active in lab-scale cultures (e.g., Pereira et al., 1996; Hesselmann et al., 2000) as well as in full-scale plants (Pijuan et al., 2008). Zhou et al. (2009) found that a laboratory enrichment of PAOs utilised the TCA cycle instead of the glycolytic pathway under glycogen limited conditions, which demonstrates that the PAO anaerobic metabolism can vary according to external or internal factors

7033

such as glycogen storage levels (Zhou et al., 2010). Furthermore, Majed et al. (2012) demonstrated that PAOs employed both the anaerobic glycolytic and TCA pathways in a lab-scale EBPR process, and suggested that the relative activity of each pathway varied among individual PAO cells. The observation that both the TCA cycle and glycolysis are active under anaerobic conditions leads to the question of what type of metabolism is prevalent in real full-scale EBPR systems and what may contribute to the use of one metabolism over the other. Investigations at the full-scale level are still few, in comparison to lab-scale investigations, and present new challenges on how to deal with a greater number of parameters and a more complex and dynamic system. Some studies have addressed the microbial diversity in WWTPs (e.g., Wong et al., 2005; He et al., 2008; Nielsen et al., 2010), while others have focused on the metabolic performance of these microbial communities (e.g., Tykesson et al., 2006; Gu et al., 2008; LopezVazquez et al., 2008; Pijuan et al., 2008). However, no emphasis has been placed so far in understanding what type of anaerobic metabolic distribution is observed in different plants with different configurations, and if metabolic differences observed can be linked to specific operational parameters. This knowledge is needed in order to improve the existing EBPR models, and to design/operate WWTPs for optimal P removal. This study investigates and compares the microbial community and EBPR activity in WWTPs in Portugal and in Denmark. The main organisms known to date to be involved in the EBPR process were assessed using quantitative fluorescence in situ hybridisation (qFISH) in five EBPR systems. The activity of the sludge from each plant was tested in anaerobic/ aerobic batch tests. The various metabolic capacities of PAOs, known mostly from lab-scale studies, were investigated in full-scale systems in order to study their prevalence and correlate it to the operation of the plants. This work links the knowledge obtained for well defined, laboratory scale systems

Fig. 1 e Schematic representation of the two EBPR configurations studied: (a) conventional anaerobic/anoxic/aerobic (A2O); (b) adapted Biodenitro configuration with side-stream hydrolysis (RSS).

7034

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

Table 1 e Summary of some of the main characteristics of the WWTPs investigated, the incoming influent and their treatment efficiency. All results presented constitute averaged values based on the information provided by the WWTPs for one year of operation at the time the sampling was conducted. Main characteristics and parameters

Wastewater treatment plants Beirolas (PT_1)

Operational conditions

Influent

Efficiency

Sludge

Layout Average flow Supplementary chemical precipitationa pH Temperature range SRT Average COD Suspended solids Average N Average NH3 Average total P COD:N:P COD removal N removal P removal NH3 removal Total P effluent SS PHAb Glycogenb

103 m3/d



C d mg/L mg/L mg-N/L mg-N/L mg-P/L mg % % % % mg-P/L g/L mg-C/gTS mg-C/gTS

Setu´bal (PT_2)

Aalborg West (DK_1)

Aalborg East (DK_2)

A2O 48  7 No

A2O 12  n/a No

Biodenitro þ RSS Biodenitro þ RSS 49  8 17  5 Yes (FeCl3) Yes (FeCl3)

7.5  0.2 15e23 12 253  75 81  23 43  13 33  10 5  0.9 51:9:1 85 64 43 60 21 2.9  0.7 2.5  1 4.3  0.8

7.5  0.2 12e25 5 1032  196 292  59 98  15 68  16 53 206:20:1 91 75 85 92 0.7  0.9 3.7  0.8 4.8  1.3 14.6  1.6

7.0  n/a 8e20 19 198  36 79  19 30  5 n/a 41 50:7:1 88 77 94 n/a 0.2  0.1 4.3  0.8 3.7  0.3 6.1  0.5

7.3  0.1 8e18 30 626  206 182  111 52  14 38  8 10  5 63:5:1 94 91 96 98 0.3  0.2 4.8  0.5 3.1  0.4 5.0  0.3

Hjørring (DK_3) A2O 13  6 Yes (FeCl3) 7.0  0.1 6e18 43 402  272 537  467 37  16 18  10 10  6 40:4:1 92 86 92 96 0.8  0.5 5.0  1.4 1.2  0.8 6.3  0.9

SRT e sludge retention time; COD e chemical oxygen demand; SS e suspended solids; TS e total solids; RSS e return sludge side-stream hydrolysis; n/a: not available. a Chemical precipitants were only added in Danish WWTP as a polishing step for P removal. b Averaged values for aerobic sludge grab samples collected and analysed in the laboratory.

with the more complex and dynamic processes that are occurring in real EBPR systems in order to develop new optimisation strategies.

2.

Materials and methods

2.1.

Sampling campaign and WWTP characteristics

The sampling campaign took place in two different types of bio-P removal systems (Fig. 1): three conventional anaerobic/ anoxic/aerobic systems (A2O) (Beirolas WWTP, SIMTEJO, Lis´ guas do Sado, Setu´bal, bon, Portugal; Setu´bal WWTP, A Portugal; Hjørring WWTP, Hjørring, Denmark) and two adapted Biodenitro configurations (Aalborg West and Aalborg East WWTP, Aalborg, Denmark), which alternate in the same tank anoxic and aerobic conditions (based on online sensors) in two lines running in parallel, combined with a return sludge sidestream hydrolysis tank (RSS) (Vollertsen et al., 2006). The RSS digests anaerobically approximately 20% of the return sludge with a residence time of 20e30 h and recirculates this fermented stream along with the influent to the head of the anoxic tank. All WWTPs treated mainly domestic wastewater, although Setu´bal receives an important industrial contribution from a bread yeast factory which increased the COD load. Beirolas and Setu´bal WWTPs were sampled in two main sampling periods: one from June to September 2010 (five samplings (PT_1) and two samplings (PT_2), Summer), where the typical water temperature was 22  C and one in March

2011 (two samplings (PT_1 and PT_2), Winter), where the typical water temperature was 15  C. Aalborg West WWTP (DK_1), Aalborg East WWTP (DK_2) and Hjørring WWTP (DK_3) were sampled between October and November, 2010 (five, two and two samplings respectively) under Danish winter conditions with a typical water temperature of 13  C. A description of the main characteristics for each plant is listed in Table 1. The values provided were obtained from the plant operators and correspond to averages of the available values during the year the campaign was conducted, i.e., May 2010 to May 2011. For each sampling date, sludge samples were collected from the end of the aerobic phase for the microbial characterisation and for the off-line batch tests. In Portuguese WWTPs, grab samples from all influent and effluent streams, as well as from each biological compartment were also collected.

2.2.

Quantitative fluorescence in situ hybridisation

Quantitative fluorescence in situ hybridisation (FISH) was carried out according to Nielsen (2009) and Mielczarek et al. (2012). Sludge samples were collected for each sampling date from the end of the aerobic tank, fixed in 4% paraformaldehyde (PFA) for gram-negative bacteria and in ethanol for gram-positive bacteria and stored at 18  C. Target organisms were quantified by their biovolume against the total bacterial biovolume as determined using a generic probe targeting all bacteria (EUBmix, containing a mixture of EUB338, EUB338II and EUB338III) (Amann et al.,

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

1990; Daims et al., 1999). Specific oligonucleotide probes included PAOmix (PAO651, PAO462 and PAO846) targeting Accumulibacter-PAOs (Crocetti et al., 2000); Tet1-266 (clade I), Tet2-174 and Tet2-892 (clade II) and Tet3-654 (clade III) targeting Tetraesphaera putative P-accumulating organisms (Nguyen et al., 2011); GAOmix (GAOQ431, GAOQ989 and GB_G2) targeting Competibacter GAOs (Crocetti et al., 2002; Kong et al., 2002); DEF1mix (TFO_DF218 and TFO_DF618) targeting Defluviicoccus vanus related GAOs cluster I (Wong et al., 2004); DEF2mix (DF988 and DF1020) targeting D. vanus related GAOs cluster II (Meyer et al., 2006); DF1013 and DF1004 targeting phylotypes within cluster III Defluviicoccus (Nittami et al., 2009), indicated as putative GAOs (McIlroy et al., 2010). The results are presented as a percentage of all bacteria and given as an average of duplicate samples. The standard deviation associated with the quantitative FISH method has been shown to be 20% (Nielsen et al., 2010).

2.3.

Batch tests

Five litre sludge samples were taken from the end of the aerobic phase and stored at 4  C overnight. The sludge was diluted with mineral medium to a final concentration of approximately 3 g/L volatile suspended solids (VSS). The sludge was washed (3, mineral medium) and sparged with argon (Portugal) or nitrogen (Denmark) for at least 15 min to attain anaerobic conditions. The sludge was subjected to a sequence of anaerobiceaerobic conditions at pH 7.0  0.2, 20  1  C and with oxygen levels close to saturation in the aerobic phase. For batch tests performed in Denmark and Portugal (winter), anaerobic maintenance tests were also carried out, where no acetate was provided. Samples were periodically taken for analysis of acetate, phosphate, ammonia concentration in the supernatant and to determine PHA, glycogen and total phosphorus in the biomass. Samples were also taken at the beginning and end of each phase for the determination of the mixed liquor suspended solids concentration and the volatile suspended solids (VSS). The mineral medium used was similar to the one used in Carvalho et al. (2007) without acetate nor phosphate. Each test commenced with the addition of an acetate (10e15 mg-C/L) and phosphate (30e40 mg-P/L) pulse. The concentration of the nutrients was optimised so that it would be possible to achieve in a 4e6 h period the maximal utilisation of the PHA internal reserves of PAOs (or GAOs) as recommended in Oehmen et al. (2010), obtained when a stabilisation of phosphorus levels was observed.

2.4.

7035

suspended solids and volatile suspended solids (TSS and VSS) were analysed according to standard methods (APHA, 1995). Glycogen was analysed in triplicate, while all other analyses were performed in duplicate.

2.5.

Calculations

The chemical concentrations given in this study were calculated per C-mol of active biomass (X). The concentration of active biomass was calculated by subtracting the amount of glycogen and PHA from the VSS. The general biomass formula used was CH1.84O0.5N0.19 as reported by Zeng et al. (2003).

3.

Results and discussion

3.1. EBPR performance and microbial characterisation in the different WWTPs Five full-scale EBPR systems were studied at different time periods: two Portuguese WWTPs (PT_1 and PT_2) sampled

Chemical analysis

Samples for the analysis of acetate, phosphate, ammonia and nitrate or nitrite were filtered (0.2 mm) and acidified (40 mM H2SO4). Samples for phosphate were stored at 4  C prior to analysis and samples for acetate, ammonia and nitrate or nitrite were stored at 18  C. Acetate and phosphate were analysed as described in Carvalho et al. (2007). Ammonia was analysed as described in Lanham et al. (2011). Samples for PHA and glycogen analysis were fixed for at least 1 h (8% formaldehyde), washed in 0.9% NaCl, then freeze-dried before analysis as described in Lanham (2012) and Lanham et al. (2012), respectively. Total

Fig. 2 e Microbial composition of the EBPR-related organisms in the sampled WWTP in Portugal and in Denmark; error bars represent the standard deviation between different sampling dates. Win e Winter; Sum e Summer.

7036

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

Fig. 3 e Chemical transformations occurring in the batch tests conducted, illustrating the cycling of phosphorus (circles), PHA (triangles), glycogen (squares) and the uptake of acetate (diamonds) in sludge during the sampling campaigns: PT_1 Summer, PT_2 Winter, DK_1 and DK_2.

both in Summer and in Winter and three Danish WWTPs sampled in wintertime (DK_1, DK_2 and DK_3). The operational characteristics of these full-scale systems are presented in Table 1. The configuration of the plants were different: PT_1, PT_2 and DK_3 were operated with a sequence of anaerobiceanoxiceaerobic tanks in a conventional A2O system, whereas DK_1 and DK_2 were operated with a Biodenitro system. Furthermore, DK_1 and DK_2 were coupled to a return sludge side-stream hydrolysis (RSS) system that provides an anaerobic sludge hydrolysis step in order to increase the fraction of available volatile fatty acids (Vollertsen et al., 2006). The Danish WWTPs also use a polishing step with chemical precipitants for additional phosphorus removal. There were also differences in climate, where the lower temperatures to which the Danish WWTPs are exposed led to lower biomass growth rates, resulting in higher sludge retention times. The average yearly removal efficiencies for P were better in the Danish plants (>90%), probably a result from the combined biological and chemical treatment, whereas the Portuguese plants had an average yearly efficiency of 43% and 85% for PT_1 and PT_2, respectively. Plant PT_2 also had a significant industrial discharge that increased the influent COD/P ratio. The microbial composition of the main organisms relevant to EBPR was assessed via quantitative FISH (qFISH), and is shown in Fig. 2. The populations of all the biomass samples collected in the different plants comprised significant quantities of Accumulibacter and of Tetrasphaera putative PAOs. Accumulibacter was observed within the range of 3.5e6% of total bacteria, with a marginally higher abundance in DK_2. Tetrasphaera were present in higher numbers, which agrees with previous reports in literature (Nguyen et al., 2011),

ranging from 17% in PT_2 to 28% in DK_1. However, these Tetrasphaera organisms are likely to be less competitive for acetate than for other substrates, as compared to Accumulibacter and Competibacter (Nguyen et al., 2011), thus likely contributing less to the anaerobic/aerobic metabolism observed in the batch tests. Competibacter biovolumes were 0.4% and Defluviicoccus-related GAOs were 0.2% in PT_1, DK_1 and DK_2, suggesting minimal GAO activity in these systems. DK_3 had >3% of D. vanus related GAOs, mainly belonging to cluster II, but minimal Competibacter (0.3%). PT_2 had a relatively high population of Competibacter in the winter months (5.1%), while the D. vanus related GAOs, also mainly cluster II, ranged from 3 to 5%. No correlation was found in this study between temperature and the abundance of GAOs in the system. While GAOs, particularly Competibacter, proliferate at high temperatures (Lopez-Vazquez et al., 2009), other factors likely contributed to the relatively high GAO abundance observed in PT_2 and DK_3. PT_2 had a very high COD/P ratio (204:1), which likely stimulated the proliferation of GAOs (Mino et al., 1998). Further, DK_3 had the highest average SRT during the study period (43 d), which is another factor known to favour GAOs (Whang and Park, 2006).

3.2. Anaerobic and aerobic stoichiometry e glycolysis vs TCA cycle At each sampling time point, an off-line anaerobic/aerobic batch test was conducted. As illustrated in Fig. 3, all systems showed a typical EBPR profile with anaerobic acetate uptake coupled with the anaerobic/aerobic cycling of phosphorus (release/uptake), PHA (formation/consumption) and glycogen (degradation/formation).

1.3

0.5

0.9

0.0

0.65 0.41

0.5 0.8

1.1  0.0 1.04e1.06 0.9  0.2 0.70e1.01 0.2  0.1 0.13e0.31 1.6  0.3 1.4e1.9 0.8  0.6 0.4e1.2 1.0  0.4 0.7e1.3 1.4  0.4 1.07e1.64 0.4  0.0 0.35 0.7  0.2 0.6e0.9 0.3  0.0 0.3 1.1  0.2 0.8e1.3 1.3  0.2 1.16e1.55 0.3  0.1 0.21e0.48 0.8  0.1 0.8e0.9 0.3  0.1 0.3e0.4 0.5  0.0 0.50e0.51 1.8  0.4 1.47e2.10 0.7  0.2 0.49e0.82 0.3  0.0 0.3 0.3  0.2 0.2e0.4

c

b

a

Glyc/PHA

P/PHA Aerobic

Glyc/HAc

PHA/HAc

Anaerobic PAO models developed by Smolders et al. (1994). Aerobic PAO model developed by Smolders et al. (1995). GAO model developed by Zeng et al. (2003).

0.3  0.1 0.30e0.38 1.6  0.2 1.40e1.74 0.2  0.2 0.02e0.35 0.2  0.1 0.2e0.3 0.2  0.0 0.2 0.9  0.1 0.85e0.96 0.8  0.1 0.72e0.82 0.1  0.0 0.04e0.07 0.9  0.1 0.8e1.0 0.4  0.3 0.2e0.6 0.6  0.2 0.46e0.87 0.8  0.2 0.67e1.02 0.2  0.1 0.04e0.31 1.3  0.4 0.8e1.8 0.7  0.3 0.4e1.1 Average Range Average Range Average Range Average Range Average Range P/HAc Anaerobic

7037

The anaerobic stoichiometric data obtained for the batch tests conducted with sludge from each plant (Table 2) were compared with metabolic models, which condense the information known so far on the metabolism of PAOs and GAOs. As reflected in the metabolic models, GAOs rely on glycogen as their sole energy source; therefore their Glyc/HAc ratios are higher than PAOs. The highest Glyc/HAc value was obtained in PT_2 in the winter period, which also contained the highest abundance of Competibacter. According to FISH analysis (Fig. 2), this represented the only case where Competibacter comprised >0.5% of the biomass. PT_2 in the summer period and DK_3 represented the other two cases where >0.5% of GAOs were present, but these sludges contained mostly Defluviicoccus, and also exhibited a much lower Glyc/HAc ratio. This suggests that Competibacter are more effective at competing with PAOs for acetate uptake as compared to Defluviicoccus. Indeed, most previous studies performed with lab-scale enrichments have found that PAOs compete principally with Competibacter for anaerobic acetate uptake (Crocetti et al., 2002; Oehmen et al., 2005) while Defluviicoccus are more effective at competing for propionate uptake (Oehmen et al., 2005; Dai et al., 2007). Thus, the results from full-scale sludges obtained in this study validate the findings previously observed in lab-scale enrichments. Three systems, PT_1, DK_1 and DK_2, had only residual quantities of GAOs according to FISH analysis (Fig. 2). The anaerobic stoichiometry appeared to vary between these different sludges, where dynamics was also observed within the same plant over time. Most of the stoichiometric yields are in between the anaerobic metabolic model predictions based on the use of the TCA cycle and glycolysis for the anaerobic production of reducing equivalents by PAOs. In some cases, such as in most PT_1 samples, the Glyc/HAc ratios were close to 0, in accordance with the utilisation of the anaerobic TCA cycle, while in some other cases, such as in most samples of DK_1 and DK_2, they were closer to what is expected when

0.42

1.1

1.9

0.0

GAOc O2 PAOb AnO2 Glyca PAO PT_1 winter (n ¼ 2) PT_1 summer (n ¼ 4)

PT_2 summer (n ¼ 2)

PT_2 winter (n ¼ 2)

DK_1 (n ¼ 4)

DK_2 (n ¼ 2)

DK_3 (n ¼ 2)

AnO2 TCAa PAO

Models Experimental results Yield (C or P-mol)

Table 2 e Anaerobic and aerobic yields for chemical transformations involving P, PHA and glycogen (Glyc) cycling, and comparison with metabolic model predictions for PAO and GAO. The average, standard deviation and range of values observed in the repeated batch tests (n [ number of tests) are indicated.

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

Fig. 4 e Correlation between the Glyc/HAc and PHA/HAc ratios in anaerobic conditions for the different WWTPs without GAOs e experimental results of PT_1 (circles e summer, diamonds e winter), DK_1 (squares) and DK_2 (triangles), with the studies of Pijuan et al. (2008) (plus markers) in two WWTPs and by Zhou et al. (2009) (stars) in a lab-scale reactor also indicated. Note that the experimental values represented reflect only the sampling points for which the redox balance closed within less than 10%.

7038

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

glycolysis is used (i.e., 0.5 C-mol/C-mol VFA (Smolders et al., 1994)). This suggests metabolic variability of PAO activity in full-scale sludges. According to metabolic models, a higher utilisation of glycogen anaerobically by PAOs leads to a higher PHA/HAc yield. Excluding the sludges that contain GAOs, a trend was observed whereby higher PHA/HAc yields were positively correlated with higher Glyc/HAc ratios (Fig. 4). This provides additional evidence supporting the hypothesis that the glycolysis and/or the TCA cycle were active in the different systems, where the intermediate values suggest a joint utilisation of both metabolic pathways by the sludges. Thus, the experimental data of two independent parameters, glycogen and PHA (analysed in triplicate and duplicate, respectively), reflected the metabolic variability regarding the anaerobic generation of reducing equivalents by PAOs. Furthermore, the experimental results of this study agree well with observations by Pijuan et al. (2008) and Zhou et al. (2009) that observed the simultaneous activity of the glycolysis and TCA cycle pathways in their systems. The P/HAc ratios observed with the sludges from the Portuguese plants were generally within the range predicted by the glycolysis and TCA cycle anaerobic PAO models. PT_2 showed a lower P/HAc ratio than PT_1, which is in accordance with the higher GAO population observed in PT_2. It should be noted that the P/HAc ratios resulted in higher values than expected for the Danish plants, which was not concordant with the Glyc/HAc and PHA/HAc ratios (Table 2). A series of anaerobic maintenance tests was conducted without the addition of acetate, where the P release rate in all of the Danish plants was approximately the double of the value obtained in the Portuguese anaerobic maintenance tests (1.3  0.2 and 0.6  0.1 P-mmol/C-molX.h, respectively). Considering that the PHA and glycogen variations were equivalent in the Portuguese and Danish anaerobic maintenance tests, another mechanism, other than cell maintenance, might have taken place in the Danish tests that influenced the P release observed. One possibility is the chemical release of phosphate from iron-phosphate precipitates due to the activity of iron reducing bacteria (Nielsen, 1996). As shown in Table 1, only the Danish plants use chemical precipitation and therefore it is plausible that ironphosphate precipitates present in these sludges could be redissolved under anaerobic conditions, partly contributing to some over-estimation of the biological P/HAc ratios for these plants. The aerobic yields are also presented in Table 2. While the Glyc/PHA yields agree fairly well with PAO model predictions in most plants, the P/PHA yields were significantly higher (2e3 times the model predictions) in all systems except for PT_2. PT_2 presented P/PHA yields that were lower than PAO model prediction, which is consistent with the fact that this plant had the highest GAO population observed by FISH. The fact that the P uptake for the other sludges was much higher than expected could indicate that the uptake of phosphorus by PAOs other than Accumulibacter (that rely on PHA for their P uptake), were actively removing P as well. Indeed, the models were formed based on lab-scale enrichments of Accumulibacter, thus neglecting the potential contribution of other PAOs. It should be noted that Tetrasphaera-related PAOs,

which have been suggested to employ storage polymers other than PHA (Nguyen et al., 2011), were >3 times more abundant than Accumulibacter in all systems studied, further supporting this hypothesis.

3.3. Link between PAO metabolism and EBPR plant operation The observation that both glycolysis and the TCA cycle are active to different extents in WWTP sludge leads to the question of which pathway is more beneficial for EBPR. Theoretically, the glycolysis metabolism should be more efficient and beneficial for EBPR, since at the end of the anaerobic period cells will contain more PHA (1.33 C-mol PHA/C-mol acetate with glycolysis vs 0.89 C-mol PHA/C-mol acetate with the TCA cycle) and will have released less phosphate anaerobically, thus necessitating a lower level of aerobic P uptake to achieve good P removal. Table 3 shows the correlation between the relative glycolysis/TCA cycle activity with the P uptake/release ratio observed in the batch tests. High Glyc/HAc ratios led to a significantly higher P uptake/release ratio as compared to sludges where the TCA cycle was prevalent. This shows that by using the TCA cycle to produce reducing equivalents anaerobically, PAOs may have lower potential phosphorus removal capacity under aerobic conditions. The results of this study show that, unlike most laboratoryenriched cultures, the TCA cycle can play a critical role in the generation of anaerobic reducing power in full-scale EBPR systems. One study showed that the anaerobic TCA cycle activity in an enriched culture was stimulated after glycogen was depleted under starvation conditions, leading to a significantly higher P/HAc ratio (Zhou et al., 2009). Thus, we investigated the link between P/HAc ratio and the initial glycogen concentration in PT_1, where GAOs were scarce and there was no contribution to P release from phosphorus precipitates. The results showed that the anaerobic P/HAc ratio was inversely correlated with the initial glycogen concentration (Fig. 5). It can be hypothesised that, when glycogen is available, glycolysis is the preferred pathway over the TCA cycle, but cells with lower glycogen levels are forced to rely more heavily on polyphosphate consumption as their main energy source, leading to higher P/HAc levels. Therefore, the fact that anaerobic TCA cycle activity was so prevalent in this study was likely due to the limiting glycogen levels in the

Table 3 e Comparison of the anaerobic Glyc/HAc yields with ratio of aerobic P uptake to anaerobic P release (PAE/ AN). The % number of tests from RSS plants and non-RSS plants that were in the “Low”, “Medium” and “High” range of glycolysis activity is also provided. PT_2 winter tests were not considered due to their relatively high abundance of Competibacter. Parameter

Units

Low

Med

High

Glyc/HAc PAE/AN RSS plants Non-RSS plants

C-mol/C-mol P-mol/P-mol % %

0.07  0.04 1.06  0.23 0 71

0.25  0.06 1.38  0.02 33 14

0.39  0.06 1.70  0.21 67 14

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

WWTP sludges. Indeed, the batch tests showed that the amount of glycogen consumed anaerobically constituted on average only 30e60% of the glycogen produced during the aerobic phase (see Fig. 3), which suggests that glycogen was limiting in the sludge when it was collected from the WWTPs. Two factors could contribute to low glycogen availability for substrate uptake in full-scale plants. Firstly, in WWTPs, the aerobic phases have longer durations (ranging in this study from 5 to 15 h) than in laboratory reactors (usually ranging from 2 to 4 h) (e.g., Smolders et al., 1994; Carvalho et al., 2007; Zhou et al., 2009; Flowers et al., 2009; Lanham et al., 2011). This longer aerobic phase may lead to glycogen depletion once the PHA reserves are exhausted, as shown by Lopez et al. (2006), where it serves as a source of energy for cell maintenance. Additionally, sludge in WWTPs normally has to endure longer starvation periods during settling and recycling, where glycogen pools can be further depleted. Alternatively, low levels of glycogen storage could also result from low availability of external carbon substrates. As shown in Fig. 6, a correlation was observed in PT_1 between the specific P uptake observed within the WWTP and the amount of VFA in the primary settler (R2 of 0.85). This shows that the biological P removal process in this plant was highly limited by VFA, which may have led to the low P removal efficiency of 43% observed in this plant. Interestingly, a correlation was also observed between the level of glycolysis activity with the EBPR process configuration (Table 3). EBPR plants employing side-stream hydrolysis (RSS: DK_1 and DK_2) exhibited relatively high glycolysis activity, unlike the A2O processes (PT_1, PT_2 and DK_3). This suggests that RSS processes may aid in facilitating the activity of the glycolysis pathway in PAOs, probably due to the fact that sludge hydrolysis generates additional VFAs. Further investigation regarding strategies that favour glycolysis activity over the TCA cycle of PAOs in EBPR systems is warranted. Despite the fact that the two RSS plants were located in Denmark, no correlation between glycolysis activity and temperature (due to seasonal/geographic variations) was found. The incorporation of the TCA cycle and glycolysis pathways into EBPR models could enable the description and prediction of PAO performance in response to different

7039

Fig. 6 e Correlation between the VFA concentration in the primary settler and the resulting specific P uptake observed in WWTP PT_1.

operational conditions in EBPR plants. Additionally, it is important to stress that the role and metabolism of other PAOs is still unclear, e.g., little could be inferred regarding the role of Tetrasphaera, which could potentially affect the stoichiometry and kinetics of the EBPR process. Future work should investigate the metabolism of other EBPR-related organisms, and integrate this knowledge with that of wellknown PAOs and GAOs in order to better understand fullscale EBPR processes.

4.

Conclusions

This study shows that PAOs in full-scale EBPR plants often employ the anaerobic TCA cycle in addition to, or instead of, the glycolysis pathway for the anaerobic generation of reducing equivalents. Limited glycogen availability tended to promote anaerobic TCA cycle activity, perhaps caused by a low abundance of VFAs and/or glycogen depletion as an energy source for cell maintenance, where both situations are commonly found in full-scale systems. Nevertheless, it was found that PAOs exhibiting a higher glycolysis activity had more efficient P removal than those employing the TCA cycle. One means of favouring glycolysis activity was through the EBPR plant configuration, where systems employing a side-stream hydrolysis process had a substantially higher level of glycolysis activity and more efficient biological P removal.

Acknowledgements

Fig. 5 e Correlation between the initial glycogen concentration and the resulting P/HAc ratio obtained for the Portuguese WWTP PT_1.

The authors acknowledge the financial support of the Fundac¸a˜o para a Cieˆncia e Tecnologia, namely through project PTDC/AAC-AMB/120581/2010 and grants SFRH/BD29477/2006 and SFRH / BPD / 88382 / 2012. The authors would also like to thank the Danish and Portuguese WWTPs for allowing us to sample the plants and collect data, in particular Dra. Ana Paula Teixeira, Eng Pedro Po´voa and Dra. Ana Nobre from ´ guas do Sado, and Enga SIMTEJO, Enga Ana Quinta˜o, from A ´ guas do Algarve. Thanks to Artur T. Sara Barreto, from A Mielczarek, for helping with FISH analyses and to Paolo Siano, for helping with the experimental work.

7040

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

references

Amann, R., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A., 1990. Combination of 16S ribosomal-RNA-targeted oligonucleotide probes with flow-cytometry for analysing mixed microbial populations. Appl. Environ. Biotechnol. 56 (6), 1919e1925. APHA, AWWA, and WPCF, 1995. Standard methods for the examination of water and wastewater. In: Eaton, A.D., Clesceri, L.S., Greenberg, A.E. (Eds.), Standard Methods for the Examination of Water and Wastewater. Port City Press, Baltimore. Carvalho, G., Lemos, P.C., Oehmen, A., Reis, M.A.M., 2007. Denitrifying phosphorus removal: linking the process performance with the microbial community structure. Water Res. 41 (19), 4383e4396. Comeau, Y., Hall, K., Hancock, R., Oldham, W., 1986. Biochemical model for enhanced biological phosphorus removal. Water Res. 20 (12), 1511e1521. Crocetti, G.R., Banfield, J.F., Keller, J., Bond, P.L., Blackall, L.L., 2002. Glycogen-accumulating organisms in laboratory-scale and full-scale wastewater treatment processes. Microbiology 148, 3353e3364. Crocetti, G.R., Hugenholtz, P., Bond, P.L., Schuler, A., Keller, J., Jenkins, D., Blackall, L.L., 2000. Identification of polyphosphate-accumulating organisms and design of 16S rRNA-directed probes for their detection and quantitation. Appl. Environ. Microbiol. 66 (3), 1175e1182. Dai, Y., Yuan, Z., Wang, X., Oehmen, A., Keller, J., 2007. Anaerobic metabolism of Defluviicoccus vanus related glycogen accumulating organisms (GAOs) with acetate and propionate as carbon sources. Water Res. 41 (9), 1885e1896. Daims, H., Bruhl, A., Amann, R., Schleifer, K.H., Wagner, M., 1999. The domain-specific probe EUB338 is insufficient for the detection of all bacteria: development and evaluation of a more comprehensive probe set. Syst. Appl. Microbiol. 22 (3), 434e444. Flowers, J.J., He, S., Yilmaz, S., Noguera, D.R., McMahon, K.D., 2009. Denitrification capabilities of two biological phosphorus removal sludges dominated by different “Candidatus Accumulibacter” clades. Environ. Microbiol. Rep. 1 (6), 583e588. Gu, A.Z., Saunders, A.M., Neethling, J.B., Stensel, H.D., Blackall, L.L., 2008. Functionally relevant microorganisms to enhanced biological phosphorus removal performance at fullscale wastewater treatment plants in the United States. Water Environ. Res. 80 (8), 688e698. He, S., Gu, A.Z., McMahon, K.D., 2008. Progress toward understanding the distribution of Accumulibacter among fullscale enhanced biological phosphorus removal systems. Microb. Ecol. 55 (2), 229e236. Hesselmann, R., Von Rummell, R., Resnick, S.M., Hany, R., Zehnder, A., 2000. Anaerobic metabolism of bacteria performing enhanced biological phosphate removal. Water Res. 34 (14), 3487e3494. Kong, Y., Ong, S.L., Ng, W.J., Liu, W.-T., 2002. Diversity and distribution of a deeply branched novel proteobacterial group found in anaerobic-aerobic activated sludge processes. Environ. Microbiol. 4 (11), 753e757. Lanham, A.B., 2012. Full-scale Biological Phosphorus Removal: Quantification of Storage Polymers, Microbial Performance and Metabolic Modelling. PhD thesis. Universidade Nova de Lisboa, Lisboa, pp. 51e74. Lanham, A.B., Moita, R., Lemos, P.C., Reis, M.A.M., 2011. Longterm operation of a reactor enriched in Accumulibacter clade I DPAOs: performance with nitrate, nitrite and oxygen. Water Sci. Technol. 63 (2), 352e359.

Lanham, A.B., Ricardo, A.R., Coma, M., Fradinho, J., Carvalheira, M., Oehmen, A., Carvalho, G., Reis, M.A.M., 2012. Optimisation of glycogen quantification in mixed microbial cultures. Bioresour. Technol. 118, 518e525. Lopez, C., Pons, M.N., Morgenroth, E., 2006. Endogenous processes during long-term starvation in activated sludge performing enhanced biological phosphorus removal. Water Res. 40 (8), 1519e1530. Lopez-Vazquez, C.M., Hooijmans, C.M., Brdjanovic, D., Gijzen, H.J., Van Loosdrecht, M.C.M., 2008. Factors affecting the microbial populations at full-scale enhanced biological phosphorus removal (EBPR) wastewater treatment plants in The Netherlands. Water Res. 42 (10e11), 2349e2360. Lopez-Vazquez, C.M., Hooijmans, C.M., Brdjanovic, D., Gijzen, H.J., van Loosdrecht, M.C.M., 2009. Temperature effects on glycogen accumulating organisms. Water Res. 43 (11), 2852e2864. Majed, N., Chernenko, T., Diem, M., Gu, A.Z., 2012. Identification of functionally relevant populations in enhanced biological phosphorus removal processes based on intracellular polymers profiles and insights into the metabolic diversity and heterogeneity. Environ. Sci. Technol. 46, 5010e5017. McIlroy, S.J., Nittami, T., Seviour, E.M., Seviour, R.J., 2010. Filamentous members of cluster III Defluviicoccus have the in situ phenotype expected of a glycogen-accumulating organism in activated sludge. FEMS Microbiol. Ecol. 74 (1), 248e256. Meyer, R.L., Saunders, A.M., Blackall, L.L., 2006. Putative glycogenaccumulating organisms belonging to the Alphaproteobacteria identified through rRNA-based stable isotope probing. Microbiology 152 (Pt 2), 419e429. Mielczarek, A., Kragelund, C., Eriksen, P., Nielsen, P., 2012. Population dynamics of filamentous bacteria in Danish wastewater treatment plants with nutrient removal. Water Res. 46, 3781e3795. Mino, T., Arun, V., Tsuzuki, Y., Matsuo, T., 1987. Effect of phosphorus accumulation on acetate metabolism in the biological phosphorus removal process. In: Ramadori, R. (Ed.), Biological Phosphate Removal from Wastewaters: Proceedings of An IAWPRC Specialized Conference Held in Rome, Italy, 28e30 September, 1987. Pergamon Press, Oxford, pp. 27e38. Mino, T., Loosdrecht, M., Heijnen, J.J., 1998. Microbiology and biochemistry of the enhanced biological phosphate removal process. Water Res. 32 (11), 3193e3207. Nguyen, H.T.T., Le, V.Q., Hansen, A.A., Nielsen, J.L., Nielsen, P.H., 2011. High diversity and abundance of putative polyphosphate-accumulating Tetrasphaera-related bacteria in activated sludge systems. FEMS Microbiol. Ecol. 76 (2), 256e267. Nielsen, J.L., 2009. Protocol for fluorescence in situ hybridization (FISH) with rRNA-targeted oligonucleotides. In: Nielsen, P.H., Lemmer, H., Daims, H. (Eds.), FISH Handbook of Wastewater Treatment. IWA Publishing, London, pp. 73e84. Nielsen, P., 1996. The significance of microbial Fe (III) reduction in the activated sludge process. Water Sci. Technol. 34 (5e6), 129e136. Nielsen, P.H., Mielczarek, A.T., Kragelund, C., Nielsen, J.L., Saunders, A.M., Kong, Y., Hansen, A.A., Vollertsen, J., 2010. A conceptual ecosystem model of microbial communities in enhanced biological phosphorus removal plants. Water Res. 44 (17), 5070e5088. Nittami, T., McIlroy, S., Seviour, E.M., Schroeder, S., Seviour, R.J., 2009. Candidatus Monilibacter spp., common bulking filaments in activated sludge, are members of cluster III Defluviicoccus. Syst. Appl. Microbiol. 32 (7), 480e489. Oehmen, A., Carvalho, G., Freitas, F., Reis, M.A.M., 2010. Assessing the abundance and activity of denitrifying polyphosphate

w a t e r r e s e a r c h 4 7 ( 2 0 1 3 ) 7 0 3 2 e7 0 4 1

accumulating organisms through molecular and chemical techniques. Water Sci. Technol. 61 (8), 2061e2068. Oehmen, A., Lemos, P.C., Carvalho, G., Yuan, Z., Keller, J., Blackall, L.L., Reis, M.A.M., 2007. Advances in enhanced biological phosphorus removal: from micro to macro scale. Water Res. 41 (11), 2271e2300. Oehmen, A., Yuan, Z., Blackall, L., Keller, J., 2005. Comparison of acetate and propionate uptake by polyphosphate accumulating organisms and glycogen accumulating organisms. Biotechnol. Bioeng. 91 (2), 162e168. Pereira, H., Lemos, P.C., Reis, M.A.M., Crespo, J.P.S.G., Carrondo, M.J.T., Santos, H., 1996. Model for carbon metabolism in biological phosphorus removal processes based on in vivo13C-NMR labelling experiments. Water Res. 30 (9), 2128e2138. Pijuan, M., Oehmen, A., Baeza, J., Casas, C., Yuan, Z., 2008. Characterizing the biochemical activity of full-scale enhanced biological phosphorus removal systems: a comparison with metabolic models. Biotechnol. Bioeng. 99 (1), 170e179. Smolders, G., Van der Meij, J., Van Loosdrecht, M., Heijnen, J., 1994. Model of the anaerobic metabolism of the biological phosphorus removal process: stoichiometry and pH influence. Biotechnol. Bioeng. 43 (6), 461e470. Smolders, G.J.F., Van der Meij, J., Van Loosdrecht, M.C.M., Heijnen, J.J., 1995. A structured metabolic model for anaerobic and aerobic stoichiometry and kinetics of the biological phosphorus removal process. Biotechnol. Bioeng. 47 (3), 277e287. Tykesson, E., Blackall, L.L., Kong, Y., Nielsen, P., La Cour Jansen, J., 2006. Applicability of experience from laboratory reactors with

7041

biological phosphorus removal in full-scale plants. Water Sci. Technol. 54 (1), 267. Vollertsen, J., Petersen, G., Borregaard, V.R., 2006. Hydrolysis and fermentation of activated sludge to enhance biological phosphorus removal. Water Sci. Technol. 53 (12), 55. Wentzel, M., Lotter, L., Loewenthal, R., 1986. Metabolic behaviour of Acinetobacter spp. in enhanced biological phosphorus removal e a biochemical model. Water SA 12 (4), 209e224. Whang, L.-M., Park, J.K., 2006. Competition between polyphosphate- and glycogen-accumulating organisms in enhanced-biological-phosphorus-removal systems: effect of temperature and sludge age. Water Environ. Res. 78 (1), 4e11. Wong, M., Mino, T., Seviour, R.J., Onuki, M., Liu, W., 2005. In situ identification and characterization of the microbial community structure of full-scale enhanced biological phosphorous removal plants in Japan. Water Res. 39 (13), 2901e2914. Wong, M.-T., Tan, F.M., Ng, W.J., Liu, W.-T., 2004. Identification and occurrence of tetrad-forming Alphaproteobacteria in anaerobic-aerobic activated sludge processes. Microbiology 150, 3741e3748. Zeng, R.J., Van Loosdrecht, M.C.M., Yuan, Z., Keller, J., 2003. Metabolic model for glycogen-accumulating organisms in anaerobic/aerobic activated sludge systems. Biotechnol. Bioeng. 81 (1), 92e105. Zhou, Y., Pijuan, M., Oehmen, A., Yuan, Z., 2010. The source of reducing power in the anaerobic metabolism of polyphosphate accumulating organisms (PAOs) e a minireview. Water Sci. Technol. 61 (7), 1653e1662. Zhou, Y., Pijuan, M., Zeng, R.J., Yuan, Z., 2009. Involvement of the TCA cycle in the anaerobic metabolism of polyphosphate accumulating organisms (PAOs). Water Res. 43 (5), 1330e1340.