Proton motive force generation from stored polymers for the uptake of acetate under anaerobic conditions


  • Present address: Aaron M. Saunders, Danish Technological Institute, Kongsvang Allé 29, Aarhus C, 8000, Denmark.

  • Editor: Anthony George

Correspondence: Linda L. Blackall, Advanced Wastewater Management Centre, The University of Queensland, St Lucia 4072, Qld, Australia. Tel.: +61 7 3365 4645; fax: +61 7 3365 4726;


The bacteria facilitating enhanced biological phosphorus removal gain a selective advantage from intracellularly stored polymer-driven substrate uptake under anaerobic conditions during sequential anaerobic : aerobic cycling. Mechanisms for these unusual membrane transport processes were proposed and experimentally validated using selective inhibitors and highly-enriched cultures of a polyphosphate-accumulating organism, Accumulibacter, and a glycogen-accumulating organism, Competibacter. Acetate uptake by both Accumulibacter and Competibacter was driven by a proton motive force (PMF). Stored polymers were used to generate the PMF –Accumulibacter used phosphate efflux through the Pit transporter, while Competibacter generated a PMF by proton efflux through the ATPase and fumarate reductase in the reductive TCA cycle.


Polyphosphate-accumulating organisms (PAOs) facilitate the enhanced biological phosphorus removal (EBPR) wastewater treatment process (Mino et al., 1998). The process is characterized by an anaerobic zone with organic carbon addition before the supply of an external electron acceptor. While PAOs cannot grow under anaerobic conditions, they are able to sequester organic carbon compounds, such as volatile fatty acids (VFAs), and reduce this carbon to polyhydroxyalkanoates using energy and reducing equivalents derived from internally stored polyphosphate and glycogen. The anaerobic breakdown of polyphosphate during the acetate uptake by PAOs leads to a release of inorganic phosphate. Anaerobic uptake of acetate gives PAOs a selective advantage over other relatively fast-growing, aerobic heterotrophs. The cells oxidize the stored polyhydroxyalkanoates when an external electron acceptor becomes available in subsequent zones of the treatment process and the PAOs grow and replenish their intracellular polyphosphate and glycogen stores. It is this storage of polyphosphate that facilitates phosphorus removal from the wastewater. Acetate also induces the release of phosphate in the sulphur bacteria, Thiomargarita namibiensis (Schulz & Schulz, 2005), which suggests that this aspect of their physiology may be similar to that of PAOs.

Glycogen-(nonpolyphosphate)-accumulating organisms (GAOs) can also sequester VFAs in the absence of an external electron acceptor and store this carbon as polyhydroxyalkanoates (Seviour et al., 2003). Therefore, while GAOs are abundant in many EBPR plants they derive the required energy and reducing equivalents from glycogen degradation alone and do not require polyphosphate for energy. As both PAOs and GAOs can assimilate VFAs under anaerobic conditions but only PAOs use this carbon to remove phosphorus from the wastewater; competition for carbon between these organisms is believed to be an important factor affecting the efficacy and stability of the EBPR process (Mino et al., 1998; Saunders et al., 2003).

Traditionally, pure cultures of planktonic cells have been used to study bacterial physiology. However, in the absence of isolates of PAOs or GAOs, highly enriched PAO and GAO cultures were used to study these unique and industrially significant bacteria. The activity of glycolysis in ‘Candidatus Accumulibacter phosphatis’ (henceforth called Accumulibacter), a PAO, and ‘Candidatus Competibacter phosphatis’ (henceforth called Competibacter), a GAO, has been confirmed by the inhibition of acetate uptake in the presence of the glycolytic inhibitor, iodoacetate (Kong et al., 2004, 2006). Experimentally based metabolic models have also been developed to describe the PAO (Smolders et al., 1994; Bond et al., 1999; Oehmen et al., 2005c) and GAO (Zeng et al., 2003) phenotypes. Ambient pH has been identified as a critical factor determining the stoichiometry of anaerobic substrate uptake as increased pH leads to an increase in the energy required for VFA uptake. PAOs meet this demand by an increased consumption of polyphosphate while GAOs meet this energy demand by an increased consumption of glycogen. This increased glycogen consumption, however, produces an excess of reducing equivalents that are balanced by producing more reduced polyhydroxyalkanoates. At relatively high pH values (above pH 8), these metabolic differences give PAOs a selective advantage over GAOs (Filipe et al., 2001; Oehmen et al., 2005a), demonstrating the importance of the anaerobic substrate uptake mechanism to the competition between PAOs and GAOs.

The actual membrane transport mechanisms that mediate this important phenotype have not been investigated. Even the general mechanisms for the uptake of short-chain fatty acids, such as acetate, are not well known. In 2003, a gene, ylcG, encoding an acetate permease, ActP, was cloned and characterized for the first time (Gimenez et al., 2003). The enzyme was found to be sensitive to the proton-ionophore carbonyl cyanide m-chlorophenylhydrazone (CCCP), suggesting that the process is driven by the proton motive force (PMF). A homologue of ylcG was also identified in the Accumulibacter genome sequence (Martin et al., 2006), located together with the acetyl-CoA transferase in an arrangement similar to that described for the acetate permease in Escherichia coli (Gimenez et al., 2003).

In this study, very highly enriched Accumulibacter, a PAO, and Competibacter, a GAO, biomasses were used to investigate anaerobic substrate uptake processes in the presence of inhibitors that alter bacterial membrane transport processes. The results of these experiments have been used to propose a membrane transport model which explains the physiological properties of Accumulibacter and Competibacter under anaerobic conditions.

Materials and methods

Characterization of the enrichment cultures

Two mixed microbial biomasses were investigated. The first was enriched in Accumulibacter in a EBPR bioreactor operated with a feed alternating between acetate and propionate as the carbon source every two sludge ages (16 days). The samples were taken from the reactor during Phase II (March 2005) of the study published by Lu et al. (2006). The second biomass was enriched in Competibacter in an acetate-fed, anaerobic–aerobic bioreactor. The operational parameters of this bioreactor were identical to those reported in Dai et al. (2007). Bacterial quantification was performed using FISH following the method detailed by Meyer et al. (2006). The following probes were used: PAO462, PAO641 and PAO846 to target Accumulibacter (Crocetti et al., 2000); GAOQ989 (Crocetti et al., 2000) and GB_G2 (Kong et al., 2002) to target Competibacter; DF1MIX (Wong et al., 2004) and DF2MIX (Meyer et al., 2006) to target Defluviicoccus vanus-related organisms; and SBR9-1a (Beer et al., 2004) to target Sphingomonas-related organisms.


Biomass was acquired from the bioreactors at the end of the aerobic period and diluted with HEPES buffer (pH 7.5, final concentration 1 g L−1). Experiments were carried out in triplicate, with 100 mL of biomass suspension at 22–24°C and pH 8.0 (±0.02). Oxygen was excluded by sparging with O2-free N2 for at least 10 min before and throughout the experiment. Incubations were undertaken in the presence of CCCP (100 μM) and N-N′-dicyclohexylcarboiimide (DCCD; 50 μM) and compared with a control without an inhibitor. DCCD inhibits via covalent modification and thus the biomass was preincubated with DCCD for 30 min before acetate addition, to allow the reaction to take place (Hermolin & Filingame, 1989). Samples for total protein were taken at the beginning of each incubation in order to standardize for variation in biomass concentration. Samples (5 mL) for acetate and phosphate concentrations were taken periodically, filtered with a sterile 0.22 μm filter (Millipore) and frozen for analysis. For the incubations where polyhydroxyalkanoates was analysed (requiring 50 mL volumes), the test volume was increased to 500 mL.

Chemical analyses

The total protein of each biomass dilution was determined using the Bicinchoninic Acid Protein Determination Kit (Sigma) using bovine serum albumin (Sigma) as the standard. Orthophosphate (PO4-P) was analysed using a Lachet Quik-Chem8000 flow injection analyser. Acetate and propionate concentrations were measured using HPLC with a Hewlett Packard X-87H 300 mm × 7.8 mm, BioRad Aminex ion exclusion HPLC column operated at 65°C. Polyhydroxyalkanoates concentrations were determined using the methods described in Oehmen et al. (2005c). Substrate release and uptake rates were determined by linear regression, and each value presented is an average (±SD; n=3) of the rates determined in three separate incubations. The student t-test (two-tail) was used to test significance.

Results and discussion

Characterization of the enrichment cultures

In the absence of pure cultures of these organisms, these extraordinarily highly enriched cultures were used to approximate pure cultures. The PAO-enriched biomass (Fig. 1a) was obtained using a novel substrate switching feed method and contained 95%Accumulibacter. This represents the most highly enriched biomass currently reported (Lu et al., 2006). The known GAOs –Competibacter, D. vanus-related organisms and Sphingomonas-related organisms – were not detected. The GAO-enriched biomass (Fig. 1b) contained 98%Competibacter, and no Accumulibacter or Sphingomonas-related organisms. A few D. vanus-related organisms were observed in each microscopic field of view (<1%). Furthermore, in anaerobic batch tests with the GAO-enriched biomass, no phosphate release was measured after 30 min of acetate uptake under anaerobic conditions, demonstrating that no PAOs were present. These data confirmed that Accumulibacter and Competibacter were the dominant organisms in the two biomasses studied.

Figure 1.

 Typical FISH images of the PAO- and GAO-enriched biomass. (a) Accumulibacter appear magenta (PAOMIX and EUBMIX) and (b) Competibacter appear magenta (GAOMIX and EUBMIX). Other Bacteria (EUBMIX) appear blue in both images. Bar=10 μm.

Acetate uptake in Accumulibacter and Competibacter is sensitive to uncouplers

In the presence of CCCP, the acetate uptake rate decreased by 90% and 100% (Table 1) in Accumulibacter- and Competibacter- enrichments respectively, demonstrating that acetate uptake by both organisms was sensitive to uncoupling. These data are consistent with a role for PMF in acetate transport. The rate of anaerobic propionate uptake by Accumulibacter was also inhibited by 80% in the presence of CCCP (Table 1). Similar tests were not made with Competibacter as its propionate uptake under anaerobic conditions is very slow (Oehmen et al., 2005b).

Table 1.   Calculated rates for the VFA uptake batch tests in the presence of CCCP (±SD, n=3)
Biomass and VFA typeControl (μmol−1 min−1 mg−1 protein)CCCP (μmol−1 min−1 mg−1 protein)Inhibition (%)
Accumulibacter– acetate0.101 (± 0.006)0.010 (± 0.009)90
Competibacter– acetate0.047 (± 0.002)0.000 (± 0.000)100
Accumulibacter– propionate0.186 (± 0.018)0.037 (± 0.002)80

PMF generation under anaerobic conditions by Competibacter

In the absence of respiratory electron transport, several mechanisms have been described for the generation of a PMF (Konings et al., 1994). Some organisms export the products of fermentative processes in symport with charged ions (usually protons), leading to the generation of a PMF, for example, lactic acid bacteria. This source of PMF is excluded in Accumulibacter and Competibacter as the products of their anaerobic metabolism of volatile fatty acids are internally stored polyhydroxyalkanoatess. A PMF can be produced by the export of protons through the F1F0-ATPase, at the expense of ATP (Konings et al., 1994). Both PAOs and GAOs produce ATP from stored polymers and this could be consumed by the F1F0-ATPase, to generate the PMF necessary for secondary transport.

In the incubations in the presence of DCCD an inhibitor of the F1F0-ATPase, acetate uptake by Competibacter was strongly reduced (Table 2). This suggests that the F1F0-ATPase is a key source of the PMF in Competibacter. A similar study using DCCD to inhibit ATPase activity in a GAO-enriched biomass found that acetate uptake was not inhibited (Liu et al., 1994). DCCD inhibits the ATPase by covalent alteration of the F0 channel and this reaction is not instantaneous. Hermolin & Fillingame (1989) demonstrated that the activity of the E. coli ATPase was only 70% reduced after 10 min and required 30 min for a 95% reduction in activity. A preincubation in the presence of DCCD before the addition of acetate is not detailed in the methods of Liu et al. (1994) and the small reported reduction in the acetate uptake rate in the presence of DCCD may have been due to insufficient time allowed for the inhibition to occur. Alternatively, the GAOs enriched by Liu et al. (1994) were not identified and might not have been Competibacter. Indeed, it would be interesting to perform these experiments on other known GAOs (Defluviicoccus-related and Sphingomonas-related) to determine whether they use the same mechanism for anaerobic VFA uptake.

Table 2.   Calculated rates for the acetate uptake batch tests in the presence of DCCD (± SD, n=3)
BiomassControl (μmol−1 min−1 mg−1 protein)DCCD (μmol−1 min−1 mg−1 protein)Inhibition (%)
Accumulibacter0.101 (± 0.006)0.100 (± 0.011)1
Competibacter0.047 (± 0.002)0.009 (± 0.015)81

GAOs balance their reducing equivalents during anaerobic acetate uptake by shunting a fraction of the carbon produced by glycolysis through the reductive TCA cycle producing propionyl-CoA (Zeng et al., 2003). Flux through this pathway should also generate a PMF as the reduction of fumarate to succinate by membrane-bound fumarate reductase is linked to translocation of a proton (Nicholls & Ferguson, 2002). The metabolic model for acetate uptake by GAOs at pH 7 (Zeng et al., 2003) predicts the consumption of 0.05 mol NADH by fumarate reductase and consumption of 0.06 mol ATP as energy for acetate uptake. The present results show that the energy for acetate uptake is derived from the PMF and both these processes therefore contribute to the energy for acetate uptake. As fumarate reductase transports 1 mol H+ mol−1 NADH and the ATPase transports 3 mol H+ mol−1 ATP, these processes can be converted to 0.05 and 0.18 mol H+, respectively. Therefore, based on these assumptions, fumarate reductase would contribute c. 22% of the PMF at pH 7.

PMF generation under anaerobic conditions by Accumulibacter

Acetate uptake by Accumulibacter after reaction with DCCD was also investigated (Table 2). In contrast to Competibacter, the rate of acetate uptake was not significantly reduced (P>0.13), which suggests that the F1F0-ATPase is not a key source of the PMF for Accumulibacter.

Likewise, the release of phosphate was unaffected by inhibition of the F1F0-ATPase. The ratio of phosphate release rate to acetate uptake rate was 0.47±0.08 P-mol C-mol−1 (n=3) compared with 0.46±0.07 P-mol C-mol−1 (n=3) in the uninhibited control. The efflux of inorganic phosphate by bacteria occurs through the inorganic phosphate transport system, Pit, in symport with a proton and thus generates a PMF (van Veen, 1997); hence, the release of phosphate through the Pit system appears to be a primary source of PMF for this organism under anaerobic conditions.

ATP generation under anaerobic conditions from the PMF

Polyphosphate is hypothesized to be a source of energy for PAOs under anaerobic conditions (Smolders et al., 1994) but the mechanism by which ATP is produced has not been investigated experimentally for Accumulibacter. A mechanism has also been proposed (Martin et al., 2006) for the direct production of ATP from polyphosphate by the reverse action of polyphosphate kinase (PPK) or by the combined action of polyphosphate : AMP phosphotransferase and adenylate kinase. However, McMahon et al. (2002) cloned and overexpressed the ppk gene from Accumulibacter and found that this enzyme had minimal ATP production activity from polyphosphate. An indirect mode of ATP production by the F1F0-ATPase, driven by a PMF generated by phosphate release through the Pit system, has been described in Acinetobacter (van Veen et al., 1994). Accumulibacter could also use this mechanism to produce ATP from the PMF generated during inorganic phosphate release.

When the ATPase was inhibited, the total polyhydroxyalkanoates production increased from 7.33±0.22 C-mol mg−1 protein (n=3) to 8.26±0.62 C-mol mg−1 protein (n=3) and the molar fraction of 3-hydroxyvalerate monomers in the polyhydroxyalkanoates increased from 0.7±2.0% to 9.1±3.5% (Table 3). This indicates that, with the ATPase inhibited, glycogen consumption had increased and the extra reducing equivalents were being consumed through the production of propionyl-CoA and its reduction to PHV. Such an increase in flux through glycolysis is also observed in GAOs, when they are required to consume extra glycogen to meet ATP demands (Zeng et al., 2003). This suggests that Accumulibacter, at least in part, meet ATP demands through the consumption of the PMF generated by inorganic phosphate release. While this does not preclude the direct generation of ATP from polyphosphate, it would be unusual that Accumulibacter meet the increased demand for ATP by an increase in glycolysis if they can directly produce ATP from polyphosphate.

Table 3.   Percent C-mol fraction of 3-hydroxybutyrate (3HB) and 3-hydroxyvalerate (3HV) in the PHA (± SD, n=3)
 PHA production (C-mol mg−1)3HB fraction (%)3HV fraction (%)
PAO model (Smolders et al., 1994) 1000
Accumulibacter– uninhibited7.33 ± 0.2299 (± 2.1)0.7 (± 2.0)
Accumulibacter– ATPase inhibited8.26 ± 0.6291 (± 3.6)9.1 (± 3.5)
GAO model (Zeng et al., 2003) 7525

Integrated models for membrane transport by Accumulibacter and Competibacter under anaerobic conditions

Accumulibacter generate a PMF by the Pit system through the efflux of inorganic phosphate generated by the hydrolysis of polyphosphate (Fig. 2a). Competibacter do not accumulate polyphosphate; they derive their PMF through the combined action of proton export through the ATPase at the cost of ATP generated by substrate-level phosphorylation in glycolysis and the reductive TCA cycle (Fig. 2b).

Figure 2.

 (a) Diagram of the proposed mechanism of acetate uptake by Accumulibacter. Acetate uptake through the acetate permease, ActP, is primarily driven by a PMF generated by the efflux of P through the Pit transporter. (b) The proposed mechanism of acetate uptake by Competibacter. Acetate uptake is primarily driven by a PMF generated by the ATPase and fumarate reductase. The ATPase consumes ATP generated by glycolysis and fumarate reductase catalyses a step in the conversion of phosphoenolpyruvate (PEP) to propionyl-CoA (marked with *). PHA, polyhydroxyalkanoate.

Filipe et al. (2001) found that an increased pH during anaerobic VFA uptake caused PAOs to release more phosphate but did not affect the consumption of glycogen, while GAOs consumed more glycogen at the higher pH. The current study proposes mechanisms that are consistent with these data.

The normal state for a cell at neutral pH is to maintain the intracellular pH slightly alkaline compared with the extracellular pH. The ΔpH thus generated is a considerable fraction of the PMF (Nicholls & Ferguson, 2002). Therefore, when the extracellular pH increases, the ΔpH decreases and the cells efflux more protons to maintain the ΔpH component of the PMF. Bond et al. (1999) demonstrated that a phosphate release of the same magnitude as that induced by acetate uptake could be induced by uncoupling the pH across the membrane using weak acids. Based on these results, it was proposed that the phosphate release was a form of intracellular pH regulation by PAOs. The present model for Accumulibacter supports these findings but proposes that phosphate release maintains intracellular pH as a part of PMF homeostasis of which the ΔpH regulation is a considerable fraction at near-neutral extracellular pH.

Previous models have predicted that PAOs generate PMF through the membrane-bound transhydrogenase (Pramanik et al., 1999) and the metabolic model proposed from the recent Accumulibacter genome sequence predicted PMF generation by the H+-translocating pyrophosphatase (Martin et al., 2006). The present experimental data do not support either of these hypotheses as both these enzymes are also inhibited by DCCD (Nyren et al., 1991; Glavas et al., 1993). These findings underline the need to experimentally validate the activity of the potential systems suggested by models and genomic data.

This study used highly-enriched cultures to gain insights into a unique metabolism but the utility of this approach is limited. The models presented here for Accumulibacter and Competibacter do not specify a metabolic capability over the ability to store glycogen, polyhydroxyalkanoates and, for Accumulibacter, polyphosphate; these are metabolic capabilities that are widespread among bacteria. Continued attempts should be made to isolate these organisms into pure culture to enable a fundamental analysis of this unique metabolism.


This work was funded by the Environmental Biotechnology Cooperative Research Centre, established and supported under the Australian Government's Cooperative Research Centres Program.