Anaerobic ammonium oxidation (anammox) bacteria contribute significantly to the global nitrogen cycle and play a major role in sustainable wastewater treatment. Anammox bacteria convert ammonium (NH4+) to dinitrogen gas (N2) using intracellular electron acceptors such as nitrite (NO2−) or nitric oxide (NO). However, it is still unknown whether anammox bacteria have extracellular electron transfer (EET) capability with transfer of electrons to insoluble extracellular electron acceptors. Here we show that freshwater and marine anammox bacteria couple the oxidation of NH4+ with transfer of electrons to insoluble extracellular electron acceptors such as graphene oxide or electrodes in microbial electrolysis cells. 15N-labeling experiments revealed that NH4+ was oxidized to N2 via hydroxylamine (NH2OH) as intermediate, and comparative transcriptomics analysis revealed an alternative pathway for NH4+ oxidation with electrode as electron acceptor. Complete NH4+ oxidation to N2 without accumulation of NO2− and NO3− was achieved in EET-dependent anammox. These findings are promising in the context of implementing EET-dependent anammox process for energy-efficient treatment of nitrogen.
Anaerobic ammonium oxidation (anammox) by anammox bacteria contributes up to 50% of N2 emitted into Earth’s atmosphere from the oceans1,2. Also, anammox bacteria have been extensively investigated for energy-efficient removal of NH4+ from wastewater3. Initially, anammox bacteria were assumed to be restricted to NH4+ as electron donor and NO2− or NO as electron acceptor4,5. More than a decade ago, preliminary experiments suggested that Kuenenia stuttgartiensis and Scalindua could couple the oxidation of formate to the reduction of insoluble extracellular electron acceptors such as Fe(III) or Mn(IV) oxides6,7. However, extracellular electron transfer (EET) activity and the molecular mechanism of this coupling reaction have remained unexplored to date. Further, these tests with K. stuttgartiensis and Scalindua could not discriminate between Fe(III) oxide reduction for nutritional acquisition (i.e., via siderophores) vs. respiration through EET8. Therefore, these preliminary experiments are not conclusive to determine if anammox bacteria have EET capability or not.
Although preliminary work showed that K. stuttgartiensis could not reduce Mn(IV) or Fe(III) with NH4+ as electron donor6, the possibility of anammox bacteria to oxidize NH4+ coupled to EET to other type of insoluble extracellular electron acceptors cannot be ruled out. In fact, EET (and set of genes involved with EET) is not uniformly applied to all insoluble extracellular electron acceptors; some electroactive bacteria are not able to transfer electrons to carbon-based insoluble extracellular electron acceptors such as electrodes in bioelectrochemical systems but could reduce metal oxides and vice versa9. It has been known for more than two decades that carbon-based high-molecular-weight organic materials, which are ubiquitous in terrestrial and aquatic environments and that are not involved in microbial metabolism (i.e., humic substances) can be used as an external electron acceptor for the anaerobic oxidation of compounds10. Also, it has been reported that anaerobic NH4+ oxidation linked to the microbial reduction of natural organic matter fuels nitrogen loss in marine sediments11. A literature survey of more than 100 EET-capable species indicated that there are many ecological niches for microorganisms able to perform EET12. This resonates with a recent finding where Listeria monocytogenes, a host-associated pathogen and fermentative gram-positive bacterium, was able to respire through a flavin-based EET process and behaved as an electrochemically active microorganism (i.e., able to transfer electrons from oxidized fuel (substrate) to a working electrode via EET process)13. Further, it was reported that anammox bacteria seem to have homologs of Geobacter and Shewanella multi-heme cytochromes that are responsible for EET14. These observations stimulated us to investigate whether anammox bacteria can couple NH4+ oxidation with EET to carbon-based insoluble extracellular electron acceptor and can behave as electrochemically active bacteria.
Here we report that in the absence of NO2−, phylogenetically distant anammox bacteria couple the anaerobic oxidation of NH4+ with transfer of electrons to carbon-based insoluble extracellular electron acceptors such as graphene oxide (GO) or electrodes poised at a certain potential in microbial electrolysis cells (MECs). Our results also revealed that anammox bacteria oxidized NH4+ to N2 with NH2OH as intermediate of the process. Interestingly, the electrons released from the NH4+ oxidation were transferred to the extracellular electron acceptor via a pathway that is analog to the ones present in metal-reducing organisms such as Geobacter spp. and Shewanella spp. Taken together, our results revealed the potential of anammox bacteria to use solid-state electron acceptors as the terminal electron sink and demonstrated that there is no need for NO2−, NO3− or partial nitritation for anaerobic NH4+ oxidation.
Results and discussion
Ammonium oxidation coupled with EET
To evaluate if anammox bacteria possess EET capability, we first tested whether enriched cultures of three phylogenetically and physiologically distant anammox species can couple the oxidation of NH4+ with the reduction of an insoluble extracellular electron acceptor. Cultures of Ca. Brocadia (predominantly adapted to freshwater environments) and Ca. Scalindua (predominantly adapted to marine water environments) were enriched and grown as planktonic cells in membrane bioreactors (Supplementary Fig. 1a)15. Fluorescence in situ hybridization (FISH) showed that the anammox bacteria constituted >95% of the bioreactor’s community (Supplementary Fig. 1b–g). Also, a previously enriched K. stuttgartiensis (predominantly adapted to freshwater environments) culture was used4. The anammox cells were incubated anoxically for 216 h in the presence of 15NH4+ (4 mM) and GO as a proxy for insoluble extracellular electron acceptor. No NO2− or NO3− was added to the incubations. GO particles are bigger than bacterial cells and cannot be internalized, and thus GO can only be reduced by EET16. Indeed, GO was reduced by anammox bacteria as shown by the formation of suspended reduced GO (rGO), which is black in color and insoluble (Fig. 1a)16. In contrast, abiotic controls did not form insoluble black precipitates. Reduction of GO to rGO by anammox bacteria was further confirmed by Raman spectroscopy, where the formation of the characteristic 2D and D + D′ peaks of rGO17 were detected in the vials with anammox cells (Fig. 1b), whereas no peaks were detected in the abiotic control. Further, isotope analysis of the produced N2 gas showed that anammox cells were capable of 30N2 formation (Fig. 1c). In contrast, 29N2 production was not significant in any of the tested anammox species or controls, suggesting that unlabeled NO2− or NO3− were not involved. The production of 30N2 indicated that the anammox cultures use a different mechanism for NH4+ oxidation in the presence of an insoluble extracellular electron acceptor (further explained below). Gas production was not observed in the abiotic control (Fig. 1c). To determine if anammox bacteria were still dominant after incubation with GO, we extracted and sequenced total DNA from the Brocadia and Scalindua vials at the end of the experiment. Differential coverage showed that the metagenomes were dominated by anammox bacteria (Supplementary Fig. 2a, c). Taken together, these results support that anammox bacteria have EET capability.
Electroactivity of anammox bacteria
Electrochemical techniques provide a powerful tool to evaluate EET, where electrodes substitute for the insoluble minerals as the terminal electron acceptor13. Compared with metal oxides, the use of electrodes as the terminal electron acceptor allow us to quantify the number of externalized electrons per mol of NH4+ oxidized. Also, since the electrode is only used for bacterial respiration, then we can better assess EET activity compared with metal oxides, where we cannot differentiate between metal oxide reduction for nutritional acquisition from respiration through EET activity. Therefore, we tested if anammox bacteria interact with electrodes via EET and use them as the sole electron acceptor in MEC. One single-chamber MEC operated at eight different set potentials (from –0.1 to 0.6 V vs. standard hydrogen electrode (SHE)) using multiple working electrodes (Supplementary Fig. 1h) was initially operated under abiotic conditions with the addition of NH4+ only. No current and NH4+ removal were observed in any of the abiotic controls. Subsequently, the Ca. Brocadia culture was inoculated into the MEC and operated under optimal conditions for anammox (i.e., addition of NH4+ and NO2−). Under this scenario, NH4+ and NO2− were completely removed from the medium without any current generation (Fig. 2a). Stoichiometric ratios of consumed NO2− to consumed NH4+ (∆NO2−/∆NH4+) and produced NO3− to consumed NH4+ (∆NO3−/∆NH4+) were in the range of 1.0–1.3 and 0.12–0.18, respectively, which are close to the theoretical ratios of the anammox reaction18. These ratios indicated that anammox bacteria were responsible for NH4+ removal in the MEC. Subsequently, NO2− was gradually decreased to 0 mM leaving the electrodes as the sole electron acceptor. When the exogenous electron acceptor (i.e., NO2−) was completely removed from the feed, anammox cells began to form a biofilm on the surface of the electrodes (Supplementary Fig. 1i) and current generation coupled to NH4+ oxidation was observed in the absence of NO2− (Fig. 2a). Further, NO2− and NO3− were below the detection limit at all time points when the working electrode was used as the sole electron acceptor, suggesting that NO2− and NO3− did not play an apparent role in the process. The magnitude of the current generation was proportional to the NH4+ concentration (Fig. 2a), and maximum current density was observed at the set potential of 0.6 V vs. SHE. There was no visible biofilm growth and current generation at set potentials ≤0.2 V vs. SHE. To confirm that the electrode-dependent anaerobic oxidation of NH4+ was catalyzed by anammox bacteria, additional control experiments were conducted in chronological order in the MEC. The presence of allylthiourea (ATU), a compound that selectively inhibits aerobic NH3 oxidation by ammonia monooxygenase (AMO) in ammonia-oxidizing bacteria (AOB), ammonia-oxidizing archaea (AOA), and Comammox19, did not result in an inhibitory effect on NH4+ removal and current generation (Fig. 2a). NH4+ was not oxidized when the MEC was operated in open circuit voltage mode (OCV; the electrode is not used as an electron acceptor) (Fig. 2b), strongly suggesting an electrode-dependent NH4+ oxidation and that trace amounts of O2 if present, are not responsible for NH4+ oxidation. Addition of NO2− resulted in an immediate drop in current density with simultaneous removal of NH4+ and NO2− and formation of NO3−, in the expected stoichiometry18 (Fig. 2c). Repeated addition of NO2− resulted in the complete abolishment of the current generation, indicating that anammox bacteria were solely responsible for current production in the absence of an exogenous electron acceptor. Absence of NH4+ from the feed resulted in no current generation, and current was immediately resumed when NH4+ was added again to the feed (Fig. 2d), further supporting the role of anammox bacteria in current generation. These results also indicate that current generation was not catalyzed by electrochemically active heterotrophs, which might utilize organic carbon generated from endogenous decay processes. Autoclaving the MEC immediately stopped current generation and NH4+ removal (Fig. 2d) indicating that current generation was due to biotic reaction. Similar results were also obtained with MECs operated with Ca. Scalindua or K. stuttgartiensis cultures (Supplementary Fig. 3a, b), suggesting that they are also electrochemically active and can oxidize NH4+ using working electrodes as the electron acceptor. Taken together, these results provide strong evidence for electrode-dependent anaerobic oxidation of NH4+ by phylogenetically distant anammox bacteria. Cyclic voltammetry (CV) was used to correlate between current density and biofilm age, in cell-free filtrates (filtered reactor solution) and the developed biofilms at different time intervals. The anodes exhibited similar redox peaks with midpoint potentials (E1/2) of −0.01 ± 0.05 V vs. SHE for all three anammox species (Fig. 2e and Supplementary Fig. 3c, d). The midpoint potentials obtained in our CV analyses were in the redox windows of cytochromes involved in external electron transport in Shewanella spp. such as CymA and MtrC20. In contrast, our results differ from a previous study that reported the complete anoxic conversion of NH4+ to N2 at oxidative potentials of 0.73 ± 0.06 V vs. SHE in a nitrifying bioelectrochemical system21. This difference in the redox potentials suggest different pathways of anoxic NH4+ oxidation. No redox peaks were observed for the cell-free solution, indicating that soluble mediators are not involved in EET. Also, the addition of exogenous riboflavin, which is a common soluble mediator involved in flavin-based EET process in gram-positive and gram-negative bacteria13,22, did not invoke changes in current density. Thus, the CV analysis corroborated that the electrode biofilms were responsible for current generation through direct EET mechanism.
The mole of electrons transferred to the electrode per mole of NH4+ oxidized to N2 (Supplementary Table 1) was stoichiometrically close to Eq. 1. Also, electron balance calculations showed that coulombic efficiency (CE) was 87.8 ± 3.2% for all NH4+ concentrations and anammox cultures tested in the experiments with electrodes as the sole electron acceptor (Supplementary Table 1).
To determine if cathodic reaction (i.e., hydrogen evolution reaction) has an effect on electrode-dependent anaerobic NH4+ oxidation, additional experiments with Ca. Brocadia were conducted by operating single and double-chamber MECs in parallel (at 0.6 V vs. SHE applied potential). However, there was no significant difference in NH4+ oxidation and current production between the different reactor configurations (Supplementary Fig. 4), suggesting no influence of cathodic reaction (i.e., H2 recycling) on the process. This was further supported by electron balance and CE calculations (Supplementary Table 1). In addition, NH4+ oxidation and current production were not affected by the addition of Penicillin G (Supplementary Fig. 4), a compound that has inhibitory effects in some heterotrophs, but it does not have any observable short-term effects on anammox activity23,24. Similar results were obtained with Ca. Scalindua and K. stuttgartiensis (data not shown). As pointed above, one of the limitations of Penicillin G is that it does not arrest the activity of all heterotrophs. Despite this limitation, the role of heterotrophs in current production was excluded because of the other experimental controls conducted in this study. Since no exogenous organic carbon was added to the MEC reactors, the only source of organics for heterotrophic organisms was through endogenous decay. However, there was no current generation in the absence of NH4+ (Fig. 2d), suggesting the lack of involvement of heterotrophic electroactive bacteria in the process.
Scanning electron microscopy (SEM) confirmed biofilm formation on the electrodes’ surface for the three tested anammox bacteria (Supplementary Fig. 5). The biofilm cell density of MECs inoculated with Ca. Brocadia was higher at 0.6 V vs. SHE (Supplementary Fig. 5e, f) compared with other set potentials, and no biofilm was observed at set potentials ≤0.2 V vs. SHE (Supplementary Fig. 5a). These observations correlate very well with the obtained current profiles at different set potentials (Fig. 2a). Cell appendages between cells and the electrode were not observed. Cell appearance was very similar to reported SEM images of anammox cells23.
FISH with anammox-specific probes (Fig. 2f) and metagenomics of DNA extracted from the biofilm on the working electrodes of MECs showed that anammox were the most abundant bacteria in the biofilm community (Supplementary Fig. 2b, d). Similarly, AOB were not detected, which further supports the lack of ATU inhibition on NH4+ removal and current generation. By differential coverage and sequence composition-based binning25, it was possible to extract high-quality genomes of Brocadia and Scalindua species from the electrodes (Supplementary Fig. 2b, d). Based on the differences in the genome content, average amino acid identity (AAI) ≤ 95% compared with reported anammox genomes to date, and evolutionary divergence in phylogenomics analysis (Supplementary Fig. 6) we propose a tentative name for Ca. Brocadia present in our MECs: Candidatus Brocadia electricigens (etymology: L. adj. electricigens; electricity generator).
Molecular mechanism of EET-dependent anammox process
After confirming through bioelectrochemical analyses that anammox bacteria are electrochemically active, isotope labeling experiments were carried out to better understand how NH4+ is converted to N2 by anammox bacteria in EET-dependent anammox process. Complete oxidation of NH4+ to N2 was demonstrated by incubating the MECs with 15NH4+ (4 mM) and 14NO2− (1 mM). Consistent with expected anammox activity, anammox bacteria consumed first the 14NO2− resulting in the accumulation of 29N2 in the headspace of the MECs. Interestingly, after depletion of available 14NO2−, a steady increase of 30N2 was observed with slower activity rates compared with the typical anammox process (Fig. 3a, Supplementary Table 2). These results confirm the GO experiments where 30N2 was detected when the three anammox species were incubated with 15NH4+ (Fig. 1c). Gas production was not observed in the abiotic control incubations. In the current model of the anammox reaction (Eq. 2)4, NH4+ is converted to N2 with NO2− as the terminal electron acceptor. This is a process in which first, NO2− is reduced to nitric oxide (NO, Eq. 3) and subsequently condensed with ammonia (NH3) to produce hydrazine (N2H4, Eq. 4), which is finally oxidized to N2 (Eq. 5). The four low-potential electrons released during N2H4 oxidation fuel the reduction reactions (Eqs. 3 and 4), and are proposed to build up the membrane potential and establish a proton-motive force across the anammoxosome membrane driving the ATP synthesis.
In the MEC experiments with Ca. Brocadia using multiple working electrodes as sole electron acceptors, we observed the production of NH2OH followed by a transient accumulation of N2H4 (Supplementary Fig. 7). No inhibitory effect was observed in incubations with 2-phenyl-4,4,5,5,-tetramethylimidazoline-1-oxyl-3-oxide (PTIO) (Supplementary Fig. 8), a NO scavenger4. Therefore, we hypothesized that NH2OH, and not NO, is an intermediate of the electrode-dependent anammox process. To investigate whether NH2OH could be produced directly from NH4+ in electrode-dependent anammox process, MECs were incubated with 15NH4+ (4 mM) and 14NH2OH (2 mM). The isotopic composition of the reactors revealed that unlabeled 14NH2OH was used as a pool substrate, and we detected newly synthetized 15NH2OH from 15NH4+ oxidation (Fig. 3b). Similarly, a previous study showed that NH2OH was the major intermediate of anoxic NH4+ oxidation performed by electroactive nitrifying microorganisms21. Even though Vilajeliu et al.21 observed the same intermediate, the difference in the community composition and midpoint redox potentials, suggest different pathways of microbial-driven anoxic NH4+ oxidation to NH2OH. It is known that NO and NH2OH, the known intermediates in the anammox process, are strong competitive inhibitors of the N2H4 oxidation activity by the hydrazine dehydrogenase (HDH)26. However, oxidation of N2H4 (Supplementary Fig. 7) and detection of 30N2 (Fig. 3a) in our experiments, suggest that even though there might be some inhibition caused by the NH2OH, the HDH is still active. Also, comparative transcriptomics analysis of the electrode’s biofilm revealed that the HDH was one of the most upregulated genes when the electrode was used as the electron acceptor instead of NO2− (Supplementary discussion). Incubations with 15NH4+ (4 mM) in 10% deuterium oxide (D2O) showed accumulation of 15NH2OD, which suggests that in order to oxidize the NH4+ to NH2OH, the different anammox bacteria use OH− ions generated from water (Fig. 3c). Abiotic incubations did not show any production of NH2OH or NH2OD. Based on these results we propose the following reactions for EET-dependent anammox process:
The complete NH4+ oxidation to N2 coupled with reproducible current production can only be explained by electron transfer from the anammoxosome compartment (energetic central of anammox cells and where the NH4+ is oxidized) to the electrode. In order to identify the possible pathways involved in NH4+ oxidation and electron flow through compartments (anammoxosome) and membranes (cytoplasm and periplasm) in EET-dependent anammox process (electrode poised at 0.6 V vs. SHE as electron acceptor) vs. typical anammox process (i.e., nitrite used as electron acceptor), we conducted a genome-centric comparative transcriptomics analysis (Supplementary discussion). Even though comparative transcriptomics analysis is a useful approach for exploring and detecting genes not previously known to play a role in adaptive responses to environmental changes, the levels of mRNA are not directly proportional to the expression level of the protein they encode, and therefore it is difficult to predict protein function and activity from quantitative transcriptome data. Accordingly, the results presented below from the comparative transcriptomics analysis should be read as hypothetical. In the anammoxosome compartment, the genes encoding for ammonium transporter (AmtB), hydroxylamine oxidoreductase and HDH were the most upregulated in response to the electrode as the electron acceptor (Supplementary Table 8). This observation agrees with the NH4+ removal and oxidation to N2 observed in the MECs and isotope labeling experiments (Figs. 2a and 3a). The genes encoding for NO and NO2− reductases (nir genes) and their redox couples were significantly downregulated when the electrode was used as the electron acceptor (Supplementary Table 8). This is expected as NO2− was not added in the electrode-dependent anammox process. Also, this supports the hypothesis that NO is not an intermediate of the electrode-dependent anammox process and that there was no effect of PTIO when NO2− was replaced by the electrode as electron acceptor (Supplementary Fig. 8). Isotope labeling experiments revealed that NH2OH was the intermediate in EET-dependent anammox process and NO was not detected throughout the experiment (Fig. 3b), suggesting that the production of NH2OH was not through NO reduction. This was further supported by the observation that the electron transfer module (ETM) and its redox partner whose function is to provide electrons to the hydrazine synthase (HZS) for NO reduction to NH2OH were downregulated (Supplementary Table 6). Interestingly, our analysis revealed that the electrons released from the N2H4 oxidation (Eq. 8) are transferred to the electrode via an EET pathway that is analog to the ones present in metal-reducing organisms such as Geobacter spp. and Shewanella spp. (Supplementary Fig. 10, Supplementary discussion). Highly expressed cytoplasmic electron carriers such as NADH and ferredoxins can be oxidized at the cytoplasmic membrane by the NADH dehydrogenase (NADH-DH) and/or formate dehydrogenase to directly reduce the menaquinone pool inside the cytoplasmic membrane (Supplementary Table 3). An upregulated protein similar to CymA (tetraheme c-type cytochrome) in Shewanella would then oxidize the reduced menaquinones, delivering electrons to highly upregulated periplasmic cytochromes shuttles and to a porin-cytochrome complex that spans the outer membrane (Supplementary Fig. 10, Supplementary Table 3). From this complex, electrons could be directly accepted by the insoluble extracellular electron acceptor. Taken together, the results from the comparative transcriptomics analysis suggest an alternative pathway for NH4+ oxidation coupled to EET when the working electrode is used as electron acceptor compared with NO2− as the electron acceptor.
In conclusion, our study provides the first experimental evidence that phylogenetically and physiologically distant anammox bacteria have EET capability and can couple the oxidation of NH4+ with the transfer of electrons to carbon-based insoluble extracellular electron acceptors. The prevalence of EET-based respiration has been demonstrated using bioelectrochemical systems for both gram-positive and gram-negative bacteria13,27. However, compared with reported EET-capable bacteria, to externalize electrons anammox bacteria have to overcome an additional electron transfer barrier: the anammoxosome compartment. Electrochemically active bacteria are typically found in environments devoid of oxygen or other soluble electron acceptors27. Our results show a novel process of anaerobic ammonium oxidation coupled to EET-based respiration of carbon-based insoluble extracellular electron acceptor by both freshwater and marine anammox bacteria and suggest that this process may also occur in natural anoxic environments where soluble electron acceptors are not available. In environments such as anoxic sediments, microbial metabolism is limited by the diffusive supply of electron acceptors28. Nitrogen loss by anammox and denitrification are expected to be limited by the diffusive flux of NO2− or NO3−, and/or by O2 diffusion that can be used by aerobic NH4+ oxidizers. However, 30N2 production has been observed in 15NH4+ incubations of sediments at depths below the penetration depth of NO2−, NO3− and O2 in marine11,28,29,30 and freshwater31 environments. These observations cannot be explained by conventional denitrification or anammox process. Interestingly, this phenomenon was observed in sediments rich in metal oxides or natural organic matter such as humic substances11,31. EET-dependent anaerobic ammonium oxidation may play a role in such environments devoid of soluble electron acceptors such as NO2−, NO3−, and O2. In natural environments, the NH4+ used by anammox bacteria is derived from heterotrophic pathways of degradation of organic matter32, and therefore anammox cannot operate independently from the mineralization of organic matter33. Ubiquitous high-molecular-weight organic compounds such as humic substances are known to act as terminal electron acceptors in anaerobic microbial respiration10,34. Also, humic substances may serve as redox mediators between electron donors and poorly soluble metal oxides minerals in soils and sediments mediating dissimilatory metal oxide reduction34. Therefore, humic substances present in natural organic matter may be a good candidate in natural environments of carbon-based electron acceptor for EET-based anaerobic ammonium oxidation, or as redox mediator between anammox and iron oxides. These results offer a new perspective of a key player involved in the biogeochemical nitrogen cycle, which previously was believed to rely strictly on soluble electron acceptors for NH4+ oxidation. The fact that anammox bacteria can perform NH4+ oxidation coupled with EET, suggest that this process may have implications in the global nitrogen cycle by contributing to the nitrogen loss in environments where soluble electron acceptors are unavailable. Therefore, a better understanding of EET processes contributes to our understanding of the cycles that occur on our planet27. Also, compared with conventional anammox process, EET-dependent anammox process achieved complete removal of NH4+ (at low and high concentrations) to nitrogen gas with no accumulation of NO2− or NO3− or the production of the greenhouse gas N2O. In conventional anammox process (i.e., when NO2− is used as the electron acceptor), NO3− is generated as a result of the oxidation of NO2− by anammox bacteria. Consequently, the effluent from conventional anammox process for wastewater treatment requires further polishing by nitrate reducing organisms before discharge. In contrast, as shown by our results, since NO2− was not added in the EET-dependent anammox process, no production of NO3− was detected. Given the fact that EET-dependent anammox process can occur at very low voltages (0.3–0.6 V vs. SHE), the process can be powered by renewable energy sources such as wind or solar35. Also, the energy released from NH4+ oxidation can be captured in the form of hydrogen gas. These findings have important implications in energy-efficient treatment of N-rich wastewater using bioelectrochemical systems. Future studies should focus on evaluating the EET-dependent process using anode materials with more conductive surface area such as porous micro-channeled electrodes in order to improve the NH4+ removal rates.
Materials and methods
Enrichment and cultivation of anammox bacteria
Biomass from upflow column reactors (XK 50/60 Column, GE Healthcare, UK) with Ca. Brocadia and Ca. Scalindua were harvested and used as inoculum. Ca. Brocadia and Ca. Scalindua planktonic cells were enriched in two bioreactors (BioFlo®115, New Brunswick, USA) equipped with a microfiltration (average pore size 0.1 µm) hollow fiber membrane module (Zena-membrane, Czech Republic) (Supplementary Fig. 1a). Operating conditions of the membrane bioreactors (MBRs) were described previously15. The MBRs were operated at pH 7.5–8.0 and 35 ± 1 °C for Brocadia and room temperature (20–25 C) for Scalindua. The culture liquid in the MBRs was continuously mixed with a metal propeller at a stirring speed of 150 rpm and purged with 95% Ar–5% CO2 at a flow rate of 10 mL min−1 to maintain anaerobic conditions. The inorganic synthetic medium was fed continuously to the reactors at a rate of ~5 L d−1 and hydraulic retention time was maintained at one day. The synthetic medium was prepared by adding the following constituents; NH4+ (2.5–10) mM, NO2− (2.5–12) mM, CaCl2 100 mg L–1, MgSO4 300 mg L–1, KH2PO4 30 mg L–1, KHCO3 500 mg L–1, and trace element solutions36. In the case of Ca. Scalindua culture, the synthetic medium was prepared using non-sterilized Red Sea water. Samples for microbial community characterization were taken from the MBRs for FISH and metagenomics analysis (see FISH and DNA extraction, metagenome library preparation, sequencing and sequence processing and analysis sections below). A previously enriched K. stuttgartiensis culture was also used for the experiments4.
Incubation of anammox bacteria with NH4 + and graphene oxide
To test whether anammox bacteria have EET capability, the three enriched anammox cultures were incubated in serum vials for 216 h with 15NH4+ and graphene oxide (GO) as a proxy for insoluble extracellular electron acceptor. Standard anaerobic techniques were employed in the batch incubation experiments. All the procedures were performed in the anaerobic chamber (Coy Laboratory Products; Grass Lake Charter Township, MI, USA). Anoxic buffers and solutions were prepared by repeatedly vacuuming and purging helium gas (>99.99%) before experiments. Biomass from the MBRs was centrifuged, washed twice, and suspended in inorganic medium containing 2 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, pH 7.8) prior to inoculation into the vials. The same composition of the inorganic medium used in the MBRs was supplied to the vials. The cell suspension was dispensed into 100 mL glass serum vials, which were sealed with butyl rubber stoppers and aluminum caps. Biomass concentration in the vials ranged from 0.1 to 0.9 mg–protein mL–1. The headspace of the serum vials was replaced by repeatedly vacuuming and purging with pure (>99.99%) helium gas. Positive pressure (50–75 kPa) was added to the headspace to prevent unintentional contamination with ambient air during the incubation and gas sampling. Prior the addition of 15NH4+, the vials were pre-incubated overnight at room temperature (~25 °C) to remove any trace amounts of substrates and oxygen. The activity test was initiated by adding 4 mM of 15NH4Cl (Cambridge Isotope Laboratories) and GO to a final concentration of 200 mg L–1 using a gas–tight syringe (VICI; Baton Rouge, LA, USA). No NO2− or NO3− was added to the incubations. The vials were incubated in triplicates at 30 °C for Ca. Brocadia and K. stuttgartiensis cultures and at room temperature (~25 °C) for vials with Ca. Scalindua. Vials without biomass were also prepared as abiotic controls. The concentrations of 28N2, 29N2, and 30N2 gas were determined by gas chromatography mass spectrometry (GC/MS) analysis37. Fifty microliters of headspace gas was collected using a gas-tight syringe (VICI; Baton Rouge, LA, USA) and immediately injected into a GC (Agilent 7890A system equipped with a CP-7348 PoraBond Q column) combined with 5975C quadrupole inert MS (Agilent Technologies; Santa Clara, CA, USA), and mass to charge (m/z) = 28, 29, and 30 was monitored. Standard calibration curve of N2 gas was prepared with 30N2 standard gas (>98% purity) (Cambridge Isotope Laboratories; Tewksbury, MA, USA). At the end of the batch incubations, DNA was extracted and sequenced for metagenomics analysis (see DNA extraction, metagenome library preparation, sequencing and sequence processing, and analysis section below). To confirm the reduction of the GO, the samples were centrifuged and subjected to dehydration process with absolute ethanol. Samples were maintained in a desiccator until Raman spectroscopy analysis. Raman spectroscopy (StellarNet Inc) was performed with the following settings: Laser 473 nm, acquisition time 20 s, accumulation 5 and objective 50×.
To evaluate if anammox bacteria (Ca. Brocadia and Ca. Scalindua) are electrochemically active, single-chamber multiple working electrode glass reactors with 500 mL working volume were operated in microbial electrolysis cell (MEC) mode. The working electrodes (anodes) were graphite rods of 8 cm length (7.5 cm inside the reactor) and 0.5 cm in diameter. Platinum mesh was used as counter electrode (cathode) and Ag/AgCl as reference electrode (Bioanalytical Systems, Inc.). A schematic representation of the multiple working electrode MEC is presented in Supplementary Fig. 1h. The multiple working electrodes were operated at a set potential of –0.1, 0, 0.1, 0.2, 0.3, 0.4, 0.5, and 0.6 V vs. SHE. The amperometric current was monitored continuously using a VMP3 potentiostat (BioLogic Science Instruments, USA), with measurements every 60 s and analyzed using EC-lab V 10.02 software. To evaluate if K. stuttgartiensis is electrochemically active, experiments were conducted in single-chamber MECs (300 mL working volume) with carbon cloth working electrode (0.6 V vs. SHE). The reactors and working and counter electrodes were sterilized by autoclaving prior to the start of the experiments. The reference electrodes were sterilized by soaking in 3 M NaCl overnight and rinsing with sterile medium. After the reactors were assembled, epoxy glue was used to seal every opening in the reactor to avoid leakage. Gas bags (0.1L Cali -5 -Bond. Calibrate, Inc.) were connected to the MECs to collect any gas generated. The gas composition in the gas bags was analyzed using a gas chromatograph (SRI 8610C gas chromatograph, SRI Instruments).
The inorganic medium composition in the MECs was the same as the one supplied in the MBRs (see enrichment and cultivation of anammox bacteria section above), with variations in the NH4+ and/or NO2− concentration. After preparation, the inorganic medium was boiled, sparged with N2:CO2 (80:20) gas mix for 30 min to remove any dissolved oxygen and finally autoclaved. The autoclaved medium was cooled down to room temperature inside the anaerobic chamber (Coy Laboratory, USA). Prior to the experiments, KHCO3 was weighed in the anaerobic chamber and dissolved in the medium. The reactors were operated in fed-batch mode at 30 °C for Ca. Brocadia and K. stuttgartiensis cultures and at room temperature (~25 °C) for Ca. Scalindua. The medium in the MECs was gently mixed with a magnetic stirrer throughout the course of the experiments. The pH of the MECs was not controlled but was at all times between 7.0 and 7.5. To exclude the effect of abiotic (i.e., non-Faradaic) current, initial operation of the reactors was done without any biomass addition. After biomass inoculation, the MECs were operated with set potentials and optimal conditions for the anammox reaction (i.e., addition of NH4+ and NO2−). Afterward, NO2− was gradually decreased to 0 mM leaving the working electrodes as the sole electron acceptor. To confirm that the electrode-dependent anaerobic oxidation of NH4+ was catalyzed by anammox bacteria, additional control experiments were conducted in chronological order including addition of allylthiourea (ATU), operation in open circuit voltage mode (i.e., anodes were not connected to the potentiostat; electrode is not used as electron acceptor), addition of nitrite, operation without addition of NH4+ and then with addition of NH4+, and autoclaving. ATU was added to a final concentration of 100 μM to evaluate the contribution of nitrifiers to the process19. Biomass from a nitrifying reactor was incubated in triplicate vials with 100 μM of ATU and was used as a positive control for the inhibitory effect of ATU. Throughout the reactor operation, the concentrations of NH4+, NO2−, and NO3− were determined as described below (see “Analytical methods” section). All experiments were done in triplicate MECs, unless mentioned otherwise.
CV at a scan rate of 1 mV s–1 was performed for the anodic biofilms at different time intervals following initial inoculation to determine their redox behavior. Scans ranged from −0.6 to 0.6 V vs. SHE at pH 7.0 and 25 °C. Current was normalized to the geometric anode surface area. To determine the presence of extracellular secreted redox mediators by anodic communities, CVs were performed with cell-free filtrates (filtered using a 0.2 μm pore diameter filter) collected from the reactors and placed in separate sterile electrochemical cells. Also, experiments were conducted to evaluate the effect of adding riboflavin, which is a common soluble mediator involved in EET in gram-positive and gram-negative bacteria13,22. Riboflavin was added to the mature anammox biofilm to a final concentration of 250 nM22.
To test if cathodic reaction (i.e., hydrogen evolution reaction) has an effect on electrode-dependent anaerobic ammonium oxidation, experiments were also conducted in double-chamber MECs (Supplementary Fig. 1k) with a single carbon cloth working electrode (0.6 V vs. SHE). The anode and cathode chambers in double-chamber MECs were separated by a proton-exchange Nafion membrane. Also, to exclude the effect of heterotrophic activity on the current generation, 500 mg L−1 of penicillin G (Sigma-Aldrich, St. Louis, MO)8 was added in the last batch cycle to inhibit heterotrophs.
To determine the role of NO in the electrode-dependent anammox metabolism, single-chamber MECs were incubated with 4 mM NH4+ and 100 μM of 2-phenyl-4,4,5,5,-tetramethylimidazoline-1-oxyl-3-oxide (PTIO), a NO scavenger. MECs with 4 mM NH4+ and without PTIO addition were run in parallel as the negative control. PTIO inhibits K. stuttgartiensis activity when NO is an intermediate of the anammox reaction4, therefore vials with K. stuttgartiensis were used as positive control of the effect of PTIO. Liquid samples were taken every day and filtered using a 0.2 μm filter and subjected to determination of NH4+ concentration as described below (see “Analytical methods” section).
For isotopic and comparative transcriptomics analysis experiments, single-chamber MECs (Adams & Chittenden Scientific Glass, USA) with a single carbon cloth working electrode (0.6 V vs. SHE) and 300 mL working volume were used (Supplementary Fig. 1k).
15N tracer batch experiments in MECs
To elucidate the molecular mechanism of electrode-dependent anaerobic ammonium oxidation by different anammox bacteria, isotopic labeling experiments were conducted in single-chamber MECs operated at set potential of 0.6 V vs. SHE. All batch incubation experiments were performed in triplicate MECs. MEC incubations without biomass for the 15N tracer batch experiments were also prepared to exclude any possibility of an abiotic reaction. Standard anaerobic techniques were employed in the batch incubation experiments. All the procedures were performed in the anaerobic chamber (Coy Laboratory Products; Grass Lake Charter Township, MI, USA). Anoxic buffers and solutions were prepared by repeatedly vacuuming and purging helium gas (>99.99%) before the experiments. The purity of 15N-labeled compounds was greater than 99%. The headspace of the MECs was replaced by repeatedly vacuuming and purging with pure (>99.99%) helium gas. Positive pressure (50–75 kPa) was added to the headspace to prevent unintentional contamination with ambient air during the incubation and gas sampling. Oxidation of NH4+ to N2 was demonstrated by incubating the MECs with 15NH4Cl (Cambridge Isotope Laboratories, 4 mM) and 14NO2− (1 mM). The MECs were incubated for 144 h at 30 °C for Ca. Brocadia and K. stuttgartiensis cultures, and at room temperature (~25 °C) for Ca. Scalindua. The concentrations of 28N2, 29N2, 30N2, 14NO, 15NO, 28N2O, 29N2O, and 30N2O gas were determined by GC/MS37. Fifty microliters of headspace gas was collected using a gas-tight syringe (VICI; Baton Rouge, LA, USA) and immediately injected into a GC (Agilent 7890A system equipped with a CP-7348 PoraBond Q column) combined with 5975C quadrupole inert MS (Agilent Technologies; Santa Clara, CA, USA). Standard calibration curve of N2 gas was prepared with 30N2 standard gas (>98% purity) (Cambridge Isotope Laboratories; Tewksbury, MA, USA).
To investigate whether hydroxylamine (NH2OH) could be produced directly from NH4+ in electrode-dependent anaerobic ammonium oxidation by anammox bacteria, single-chamber MECs were incubated with 15NH4Cl (4 mM, Cambridge Isotope Laboratories) and an unlabeled pool of 14NH2OH (2 mM) for 144 h. Liquid samples were taken every day and filtered using a 0.2 μm filter and subjected to determination of 15NH2OH and 14NH2OH. NH2OH was determined by GC/MS analysis after derivatization using acetone38. Briefly, 100 µl of liquid sample was mixed with 4 µl of acetone, and 2 µl of the derivatized sample was injected to a GC (Agilent 7890 A system equipped with a CP-7348 PoraBond Q column) combined with 5975 C quadrupole inert MS (Agilent Technologies; Santa Clara, CA, USA) in splitless mode. NH2OH was derivatized to acetoxime (C3H7NO), and the molecular ion peaks were detected at mass to charge (m/z) = 73 and 74 for 14NH2OH and 15NH2OH, respectively. Twenty-five micrometers of 14NH2OH and 15NH2OH were used as standards. To determine the source of the oxygen used in the electrode-dependent NH4+ oxidation to NH2OH, MECs were incubated with 15NH4Cl (4 mM, Cambridge Isotope Laboratories) in the presence of 10% D2O for 144 h. Stable isotopes of NH2OH were determined by GC/MS analysis after derivatization using acetone as described above.
Activity and electron balance calculations
Activities of specific anammox (29N2) with nitrite as the preferred electron acceptor and electrode-dependent anammox (30N2) with working electrode (0.6 V vs. SHE) as sole electron acceptor were calculated based on the changes in gas concentrations in single-chamber MEC batch incubations. The activity was normalized against the protein content of the biofilm on the electrodes. Protein content was measured as described below (see “Analytical methods” section).
The moles of electrons recovered as current per mole of NH4+ oxidized were calculated using
where I is the current (A) obtained from the chronoamperometry, dt (s) is the time interval over which data were collected, NH4+ is the moles of NH4+ consumed during the experiment, and F = 96,485 C mol–1 is Faraday’s constant. CE was calculated using
where nCE Theo(NH4+) is the theoretical number of moles of electrons (in our case it is three moles of electrons) recovered as current per mole of NH4+ oxidized.
All samples were filtered through a 0.2 µm pore-size syringe filters (Pall corporation) prior to chemical analysis. NH4+ concentration was determined photometrically using the indophenol method39 (lower detection limit = 5 μM). Absorbance at a wavelength of 600 nm was determined using multi-label plate readers (SpectraMax Plus 384; Molecular Devices, CA, USA). NO2− concentration was determined by the naphthylethylenediamine method39 (lower detection limit = 5 μM). Samples were mixed with 4.9 mM naphthylethylenediamine solution, and the absorbance was measured at a wavelength of 540 nm. NO3− concentration was measured by HACH kits (HACH, CO, USA; lower detection limit = 0.01 mg l−1 NO3−-N). User’s guide was followed for these kits and concentrations were measured by spectrophotometer (D5000, HACH, CO, USA). Concentrations of NH2OH and hydrazine (N2H4) were determined colorimetrically as previously described40. For NH2OH, liquid samples were mixed with 8-quinolinol solution (0.48% (w/v) trichloroacetic acid, 0.2% (w/v) 8-hydroxyquinoline and 0.2 M Na2CO3), and heated at 100 °C for 1 min. After cooling down for 15 min, absorbance was measured at 705 nm41. N2H4 was derivatized with 2% (w/v) p-dimethylaminobenzaldehyde and absorbance at 460 nm was measured42. The concentration of biomass on the working electrodes was determined as protein concentration using the DC Protein Assay Kit (Bio-Rad, Tokyo, Japan) according to manufacturer’s instructions. Bovine serum albumin was used as the protein standard.
Fluorescence in situ hybridization (FISH)
The microbial community in the MBRs and the spatial distribution of anammox cells on the surface of the graphite rod electrodes was examined by FISH after 30 days of reactor operation. The graphite rod electrodes were cut in the anaerobic chamber with a sterilized tube cutter (Chemglass Life Sciences, US). The electrode samples were fixed with 4% (v/v) paraformaldehyde (PFA), followed by 10 nm cryosectioning at −30 °C (Leica CM3050 S Cryostat). FISH with rRNA-targeted oligonucleotide probes was performed as described elsewhere43 using the EUB338 probe mix composed of equimolar EUB338 I, EUB338 II, and EUB 338 III44,45 for the detection of bacteria and probes AMX820 or SCA1309 for anammox46,47. Cells were counterstained with 1 μg ml−1 DAPI (4′,6-diamidino-2-phenylindole) solution. Fluorescence micrographs were recorded by using a Leica SP7 confocal laser scanning microscope. To determine the relative abundance of anammox bacteria by quantitative FISH, 20 confocal images of FISH probe signals were taken at random locations in each well and analyzed by using the digital image analysis DAIME software as described elsewhere48.
Scanning electron microscopy
The graphite rod electrodes were cut in the anaerobic chamber with sterilized tube cutter (Chemglass Life Sciences, US). The electrode samples were soaked in 2% glutaraldehyde solution containing phosphate buffer (50 mM, pH 7.0) and stored at 4 °C. Sample processing and scanning electron microscopy (SEM) was performed as described elsewhere49. Samples from the carbon-cloth electrodes were punched out using a 4.8 mm Ø biopsy punch and placed into a 200 μm cavity of a type A platelet (6 mm diameter; 0.1–0.2 mm depth, Leica Microsystems) and closed with the flat side of a type B platelet (6 mm diameter, 300 μm depth). Platelet sandwiches were cryo-fixed by high-pressure freezing (Leica HPM 100; Leica Microsystems, Vienna, Austria) and stored in liquid nitrogen until use. For Hexamethyldisalizane (HMDS) embedding, frozen samples were freeze-substituted in anhydrous methanol containing 2% osmium tetroxide, 0.2% uranyl acetate, and 1% H2O50. The substitution followed several intervals: cells were kept at −90 °C for 47 h; brought to −60 °C at 2 °C per hour and kept at −60 °C for 8 h; brought to −30 °C at 2 °C per hour and kept at −30 °C for 8 h in a freeze-substitution unit (AFS; Leica Microsystems, Vienna, Austria). To remove fixatives the samples were washed four times for 30 min in the AFS device at −30 °C with anhydrous methanol and subsequently infiltrated with HMDS by incubating two times for 15 min with 50% HMDS in anhydrous methanol followed by two times 15 min 100% HMDS. After blotting and air-drying the electrode samples were mounted on specimen stubs using conductive carbon tape and sputter-coated with gold-palladium before imaging in a JEOL JSM-6335F SEM, operating at 3 kV.
Metagenomics sequencing and analysis
Biomass from the vials of the GO experiment was harvested by centrifugation (4000 × g, 42 °C) at the end of the batch incubations. Biofilm samples from the electrodes were collected after 30 days of reactor operation with the working electrode as the sole electron acceptor. The biomass pellet and the electrode samples were suspended in Sodium Phosphate Buffer in the Lysing Matrix E 2 mL tubes (MP Biomedicals, Tokyo, Japan). After 2 min of physical disruption by bead beating (Mini-beadbeaterTM, Biospec products), the DNA was extracted using the Fast DNA spin kit for soil (MP Biomedicals, Tokyo, Japan) according to the manufacturer’s instructions. The DNA was quantified using Qubit (Thermo Fisher Scientific, USA) and fragmented to ~550 bp using a Covaris M220 with microTUBE AFA Fiber screw tubes and the settings: duty factor 20%, peak/displayed power 50 W, cycles/burst 200, duration 45 s, and temperature 20 °C. The fragmented DNA was used for metagenome preparation using the NEB Next Ultra II DNA library preparation kit. The DNA library was paired-end sequenced (2 × 301 bp) on a Hiseq 2500 system (Illumina, USA).
Raw reads obtained in the FASTQ format were processed for quality filtering using Cutadapt package v. 1.1051 with a minimum phred score of 20 and a minimum length of 150 bp. The trimmed reads were assembled using SPAdes v. 3.7.152. The reads were mapped back to the assembly using minimap253 (v. 2.5) to generate coverage files for metagenomic binning. These files were converted to the sequence alignment/map (SAM) format using samtools54. Open reading frames (ORFs) were predicted in the assembled scaffolds using Prodigal55. A set of 117 hidden Markov models (HMMs) of essential single-copy genes were searched against the ORFs using HMMER3 (http://hmmer.janelia.org/) with default settings, with the exception that option (-cut_tc) was used56. Identified proteins were taxonomically classified using BLASTP against the RefSeq protein database with a maximum e-value cut-off of 10−5. MEGAN was used to extract class-level taxonomic assignments from the BLAST output57. The script network.pl (http://madsalbertsen.github.io/mmgenome/) was used to obtain paired-end read connections between scaffolds. 16S rRNA genes were identified using BLAST58 (v. 2.2.28+, and the 16S rRNA fragments were classified using SINA59 (v. 1.2.11) with default settings except min identity adjusted to 0.80. Additional supporting data for binning was generated according to the description in the mmgenome package60 (v. 0.7.1.). Genome binning was carried out in R61 (v. 3.3.4) using the R-studio environment. Individual genome bins were extracted using the multimetagenome principles25 implemented in the mmgenome R package61 (v. 0.7.1). Completeness and contamination of bins were assessed using coverage plots through the mmgenome R package and by the use of CheckM62 based on the occurrence of a set of single-copy marker genes63. Genome bins were refined manually as described in the mmgenome package and the final bins were annotated using PROKKA64 (v. 1.12-beta).
Extracted bins and reported anammox genomes were used for phylogenetic analysis. Reported anammox genomes were downloaded from the NCBI GenBank. HMM profiles for 139 single-copy core genes63 were concatenated using anvi’o platform65. Phylogenetic trees with estimated branch support values were constructed from these concatenated alignments using MEGA766 with Neighbor Joining, Maximum-likelihood and UPGMA methods.
Comparative transcriptomics analysis
Comparative transcriptomic analysis was conducted to compare the metabolic pathway of NH4+ oxidation and electron flow when working electrode is used as the electron acceptor vs. NO2− as the electron acceptor. Samples for comparative transcriptomic analysis were taken from mature electrode’s biofilm of duplicate single-chamber MECs with NO2− as the sole electron acceptor and after switching to set potential growth (0.6 V vs. SHE, electrode as the electron acceptor). Biofilm samples were collected from carbon cloth electrodes with sterilized scissors in the anaerobic chamber. Samples were stored in RNAlater™ Stabilization Solution (Invitrogen™) until further processing. Total RNA was extracted from the samples using PowerBiofilm RNA Isolation kit (QiAGEN) according to the manufacturer’s instructions. The RNA concentration of all samples was measured in duplicate using the Qubit BR RNA assay. The RNA quality and integrity were confirmed for selected samples using TapeStation with RNA ScreenTape (Agilent Technologies). The samples were depleted of rRNA using the Ribo-zero Magnetic kit (Illumina Inc.) according to manufacturer’s instructions. Any potential residual DNA was removed using the DNase MAX kit (MoBio Laboratories Inc.) according to the manufacturer’s instructions. After rRNA depletion and DNase treatment the samples were cleaned and concentrated using the RNeasy MinElute Cleanup kit (QIAGEN) and successful rRNA removal was confirmed using TapeStation HS RNA Screentapes (Agilent Technologies). The samples were prepared for sequencing using the TruSeq Stranded Total RNA kit (Illimina Inc.) according to the manufacturer’s instructions. Library concentrations were measured using Qubit HS DNA assay and library size was estimated using TapeStation D1000 ScreenTapes (Agilent Technologies). The samples were pooled in equimolar concentrations and sequenced on an Illumina HiSeq2500 using a 1 × 50 bp Rapid Run (Illumina Inc).
Raw sequence reads in fastq format were trimmed using USEARCH67 v10.0.2132, -fastq_filter with the settings -fastq_minlen 45 -fastq_truncqual 20. The trimmed transcriptome reads were also depleted of rRNA using BBDuk68 with the SILVA database as reference database69. The reads were mapped to the predicted protein coding genes generated from Prokka64 v1.12 using minimap253 v2.8-r672, both for the total metagenome and each extracted genome bin. Reads with a sequence identity below 0.98 were discarded from the analysis. The count table was imported to R61, processed and normalized using the DESeq2 workflow70 and then visualized using ggplot2. Analyses of overall sample similarity were done using normalized counts (log transformed), through vegan71 and DESeq270 packages (Supplementary Fig. 9). Differentially expressed genes were evaluated for the presence of N-terminal signal sequences, transmembrane spanning helices (TMH) and subcellular localization using SignalP 5.072, TMHMM 2.0 software and PSORTb 3.0.273 respectively. Differentially expressed genes that appeared annotated as ‘hypothetical’ were reconsidered for a putative function employing BLAST searches (i.e., BLASTP, CD-search, SmartBLAST), MOTIF search, COG, and PFAM databases, as well as by applying the HHpred homology detection and structure prediction program (MPI Bioinformatics Toolkit).
Statistics and reproducibility
The number of replicates is detailed in the subsections for each specific experiment and was mostly determined by the amount of biomass available for the different cultures. In all experiments, three biological replicates were used, unless mentioned otherwise. No statistical methods were used to predetermine the sample size. The experiments were not randomized, and the investigators were not blinded to allocation during experiments and outcome assessment. Statistical analyses were carried out in R61 v. 3.3.4 using the R-studio environment.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Raw sequencing reads of Illumina HiSeq of metagenomics and metatranscriptomics data associated with this project can be found at the NCBI under BioProject PRJNA517785. Annotated GenBank files for the anammox genomes extracted in this study can be found under the accession numbers SHMS00000000 and SHMT00000000. The genome binning and the comparative transcriptomics analysis are entirely reproducible using the R files available on https://github.com/DarioRShaw/Electro-anammox. Also, complete datasets generated in the differential expression analysis are available in the online version of the paper.
Devol, A. H. Denitrification, anammox, and N2 production in marine sediments. Ann. Rev. Mar. Sci. 7, 403–423 (2015).
Lam, P. & Kuypers, M. M. M. Microbial nitrogen cycling processes in oxygen minimum zones. Ann. Rev. Mar. Sci. 3, 317–345 (2010).
Kartal, B., Kuenen, J. G. & van Loosdrecht, M. C. M. Sewage treatment with anammox. Science 328, 702–703 (2010).
Kartal, B. et al. Molecular mechanism of anaerobic ammonium oxidation. Nature 479, 127–130 (2011).
Hu, Z., Wessels, H. J. C. T., van Alen, T., Jetten, M. S. M. & Kartal, B. Nitric oxide-dependent anaerobic ammonium oxidation. Nat. Commun. 10, 1244 (2019).
Strous, M. et al. Deciphering the evolution and metabolism of an anammox bacterium from a community genome. Nature 440, 790–794 (2006).
Vossenberg, J. et al. Enrichment and characterization of marine anammox bacteria associated with global nitrogen gas production. Environ. Microbiol 10, 3120–3129 (2008).
Jiménez, F., Chan, C. H. & Bond, D. R. Identification of different putative outer membrane electron conduits necessary for Fe(III) Citrate, Fe(III) Oxide, Mn(IV) oxide, or electrode reduction by Geobacter sulfurreducens. J. Bacteriol. 200, e00347–18 (2018).
Rabaey, K., Angenent, L., Schröder, U. & Keller, J. Bioelectrochemical Systems: From Extracellular Electron Transfer to Biotechnological Application (IWA Publishing, 2009).
Lovley, D. R., Coates, J. D., Blunt-Harris, E. L., Phillips, E. J. P. & Woodward, J. C. Humic substances as electron acceptors for microbial respiration. Nature 382, 445–448 (1996).
Rios-Del Toro, E. E., Valenzuela, E. I., Ramírez, J. E., López-Lozano, N. E. & Cervantes, F. J. anaerobic ammonium oxidation linked to microbial reduction of natural organic matter in marine sediments. Environ. Sci. Technol. Lett. 5, 571–577 (2018).
Koch, C. & Harnisch, F. Is there a specific ecological niche for electroactive microorganisms? ChemElectroChem 3, 1282–1295 (2016).
Light, S. H. et al. A flavin-based extracellular electron transfer mechanism in diverse Gram-positive bacteria. Nature 562, 140–144 (2018).
Ferousi, C. et al. Iron assimilation and utilization in anaerobic ammonium oxidizing bacteria. Curr. Opin. Chem. Biol. 37, 129–136 (2017).
Oshiki, M., Awata, T., Kindaichi, T., Satoh, H. & Okabe, S. Cultivation of planktonic anaerobic ammonium oxidation (anammox) bacteria using membrane bioreactor. Microbes Environ. 28, 436–443 (2013).
Salas, E. C., Sun, Z., Lüttge, A. & Tour, J. M. Reduction of graphene oxide via bacterial respiration. ACS Nano 4, 4852–4856 (2010).
Kalathil, S. et al. Bioinspired synthesis of reduced graphene oxide-wrapped geobacter sulfurreducens as a hybrid electrocatalyst for efficient oxygen evolution reaction. Chem. Mater. 31, 3686–3693 (2019).
Lotti, T., Kleerebezem, R., Lubello, C. & van Loosdrecht, M. C. M. Physiological and kinetic characterization of a suspended cell anammox culture. Water Res. 60, 1–14 (2014).
Kessel, M. A. H. J. Van et al. Complete nitrification by a single microorganism. Nature 528, 555–559 (2015).
Kracke, F., Vassilev, I. & O. Krömer, J. Microbial electron transport and energy conservation—the foundation for optimizing bioelectrochemical systems. Front. Microbiol. 6, 1–18 (2015).
Vilajeliu-Pons, A. et al. Microbial electricity driven anoxic ammonium removal. Water Res. 130, 168–175 (2018).
Marsili, E. et al. Shewanella secretes flavins that mediate extracellular electron transfer. PNAS 105, 6–11 (2008).
Oshiki, M. et al. Nitrate-dependent ferrous iron oxidation by anaerobic ammonium oxidation (Anammox) Bacteria. Appl. Environ. Microbiol. 79, 4087–4093 (2013).
Hu, Z., van Alen, T., Jetten, M. S. M. & Kartal, B. Lysozyme and penicillin inhibit the growth of anaerobic ammonium-oxidizing planctomycetes. Appl. Environ. Microbiol. 79, 7763–7769 (2013).
Albertsen, M. et al. Genome sequences of rare, uncultured bacteria obtained by differential coverage binning of multiple metagenomes. Nat. Biotechnol. 31, 533–538 (2013).
Maalcke, W. J. et al. Characterization of anammox hydrazine dehydrogenase, a Key -producing enzyme in the global nitrogen cycle. J. Biol. Chem. 291, 17077–17092 (2016).
Kumar, A. et al. The ins and outs of microorganism–electrode electron transfer reactions. Nat. Rev. Chem. 1, 0024 (2017).
Prokopenko, M. G. et al. Nitrogen losses in anoxic marine sediments driven by Thioploca-anammox bacterial consortia. Nature 500, 194–198 (2013).
Chong, L. S., Prokopenko, M. G., Berelson, W. M., Townsend-Small, A. & McManus, J. Nitrogen cycling within suboxic and anoxic sediments from the continental margin of Western North America. Mar. Chem. 128–129, 13–25 (2012).
Garcias-Bonet, N. et al. High denitrification and anaerobic ammonium oxidation contributes to net nitrogen loss in a seagrass ecosystem in the central Red Sea. Biogeosciences 15, 7333–7346 (2018).
Crowe, S. A. et al. Novel anammox bacteria and nitrogen loss from Lake Superior. Sci. Rep. 7, 1–7 (2017).
D’Hondt, S., Pockalny, R., Fulfer, V. M. & Spivack, A. J. Subseafloor life and its biogeochemical impacts. Nat. Commun. 10, 1–13 (2019).
Trimmer, M., Engström, P. & Thamdrup, B. Stark contrast in denitrification and anammox across the deep norwegian trench in the skagerrak. Appl. Environ. Microbiol. 79, 7381–7389 (2013).
Klüpfel, L., Piepenbrock, A., Kappler, A. & Sander, M. Humic substances as fully regenerable electron acceptors in recurrently anoxic environments. Nat. Geosci. 7, 195–200 (2014).
Katuri, K. P. et al. Dual-function electrocatalytic and macroporous hollow-fiber cathode for converting waste streams to valuable resources using microbial electrochemical systems. Adv. Mater. 30, 1707072 (2018).
van de Graaf, A. A. et al. Anaerobic oxidation of ammonium is a biologically mediated process. Appl Env. Micro. 61, 1246–1251 (1995).
Ali, M. et al. Physiological characterization of anaerobic ammonium oxidizing bacterium ‘Candidatus Jettenia caeni’. Environ. Microbiol. 17, 2172–2189 (2015).
Peng, S. X. et al. Gas chromatographic-mass spectrometric analysis of hydroxylamine for monitoring the metabolic hydrolysis of metalloprotease inhibitors in rat and human liver microsomes. J. Chromatogr. B Biomed. Sci. Appl 724, 181–187 (1999).
E.W., Baird, R.B., Eaton, A. D. & Clesceri, L. S. Standard Methods for the Examination of Water and Wastewater (American Public Health Association, American Water Works Association and Water Environment Federation, 2012).
Oshiki, M., Ali, M., Shinyako-hata, K., Satoh, H. & Okabe, S. Hydroxylamine-dependent anaerobic ammonium oxidation (Anammox) by “Candidatus Brocadia sinica”. Env. Microbiol 18, 3133–3143 (2016).
Frear, D. S. & Burrell, R. C. Spectrophotometric method for determining hydroxylamine reductase activity in higher plants. Anal. Chem. 27, 1664–1665 (1955).
Watt, G. W. & Chrisp, J. D. Spectrophotometric method for determination of hydrazine. Anal. Chem. 24, 2006–2008 (1952).
Nielsen, P.H., Daims, H. & Lemmer, H. FISH Handbook for Biological Wastewater Treatment (IWA Publishing, 2009).
Daims, H., Brühl, A., Amann, R., Schleifer, K. H. & Wagner, M. 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, 434–444 (1999).
Amann, R. I. et al. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol 56, 1919–1925 (1990).
Schmid, M. et al. Molecular evidence for genus level diversity of bacteria capable of catalyzing anaerobic ammonium oxidation. Syst. Appl. Microbiol. 23, 93–106 (2000).
Schmid, M. et al. Candidatus “Scalindua brodae”, sp. nov., Candidatus “Scalindua wagneri”, sp. nov., two new species of anaerobic ammonium oxidizing bacteria. Syst. Appl. Microbiol. 26, 529–538 (2003).
Daims, H. Use of fluorescence in situ hybridization and the daime image analysis program for the cultivation-independent quantification of microorganisms in environmental and medical samples. Cold Spring Harb. Protoc. 4, 1–8 (2009).
Alqahtani, M. F. et al. Porous hollow fiber nickel electrodes for effective supply and reduction of carbon dioxide to methane through microbial electrosynthesis. Adv. Funct. Mater. 28, 1–8 (2018).
Walther, P. & Ziegler, A. Freeze substitution of high-pressure frozen samples: the visibility of biological membranes is improved when the substitution. J. Microsc. 208, 3–10 (2002).
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 17, 10–12 (2011).
Bankevich, A. et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. 19, 455–477 (2012).
Li, H. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 34, 3094–3100 (2018).
Li, H. et al. The sequence alignment/map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinform. 11, 119 (2010).
Dupont, C. L. et al. Genomic insights to SAR86, an abundant and uncultivated marine bacterial lineage. ISME J. 6, 1186–1199 (2012).
Tamura, K. et al. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, 2731–2739 (2011).
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
Pruesse, E., Peplies, J. & Glöckner, F. O. SINA: accurate high-throughput multiple sequence alignment of ribosomal RNA genes. Bioinformatics 28, 1823–1829 (2012).
Karst, S. M. S. M., Kirkegaard, R. H. & Albertsen, M. Mmgenome: A Toolbox for Reproducible Genome Extraction From Metagenomes. https://www.biorxiv.org/content/10.1101/059121v1.full (2016).
R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria, 2018).
Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P. & Tyson, G. W. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 25, 1043–1055 (2015).
Campbell, B. J., Yu, L., Heidelberg, J. F. & Kirchman, D. L. Activity of abundant and rare bacteria in a coastal ocean. Proc. Natl Acad. Sci. USA 108, 12776–12781 (2011).
Seemann, T. Prokka: rapid prokaryotic genome annotation. Bioinformatics 30, 2068–2069 (2014).
Eren, A. M. et al. Anvi’o: an advanced analysis and visualization platform for ‘omics data. PeerJ 3(e1319), 1–29 (2015).
Kumar, S., Stecher, G. & Tamura, K. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874 (2016).
Edgar, R. C. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461 (2010).
Bushnell, B. BBMap: A Fast, Accurate, Splice-aware Aligner Ernest Orlando Lawrence (Berkeley National Laboratory, 2014).
Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41, D590–D596 (2013).
Love, M. I., Anders, S. & Hu, W. Differential Analysis of Count Data—The DESeq2 Package (Colorado State University, 2016).
Oksanen, J. et al. The vegan package. Commun. Ecol. Packag. 10, 631–637 (2007).
Juan, J. et al. SignalP 5.0 improves signal peptide predictions using deep neural networks. Nat. Biotechnol. 37, 420–423 (2019).
Yu, N. Y. et al. PSORTb 3.0: improved protein subcellular localization prediction with refined localization subcategories and predictive capabilities for all prokaryotes. Bioinformatics 26, 1608–1615 (2010).
This work was supported by Center Competitive Funding Program (FCC/1/1971-33-01) to P.E.S. from King Abdullah University of Science and Technology (KAUST). M.S.M.J. was supported by ERC AG 232937 and 339880 and SIAM OCW/NWO 024002002.
The authors declare no competing interests.
Peer review information Nature Communications thanks the anonymous reviewers for their contributions to the peer review of this work. Peer review reports are available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Shaw, D.R., Ali, M., Katuri, K.P. et al. Extracellular electron transfer-dependent anaerobic oxidation of ammonium by anammox bacteria. Nat Commun 11, 2058 (2020). https://doi.org/10.1038/s41467-020-16016-y
This article is cited by
Redox controls on anaerobic ammonium oxidation coupled to reduction of natural organic matter in paddy ecosystems
Biology and Fertility of Soils (2023)
A general approach to explore prokaryotic protein glycosylation reveals the unique surface layer modulation of an anammox bacterium
The ISME Journal (2022)
Controlling anammox speciation and biofilm attachment strategy using N-biotransformation intermediates and organic carbon levels
Scientific Reports (2022)
Frontiers of Environmental Science & Engineering (2022)