Quest for Nitrous Oxide-reducing Bacteria Present in an Anammox Biofilm Fed with Nitrous Oxide

Abstract

N₂O-reducing bacteria have been examined and harnessed to develop technologies that reduce the emission of N₂O, a greenhouse gas produced by biological nitrogen removal. Recent investigations using omics and physiological activity approaches have revealed the ecophysiologies of these bacteria during nitrogen removal. Nevertheless, their involvement in‍ ‍anammox processes remain unclear. Therefore, the present study investigated the identity, genetic potential, and activity‍ ‍of N₂O reducers in an anammox reactor. We hypothesized that N₂O is limiting for N₂O-reducing bacteria‍ ‍and an‍ ‍exogeneous N₂O supply enriches as-yet-uncultured N₂O-reducing bacteria. We conducted a 1200-day incubation of N₂O-reducing bacteria in an anammox consortium using gas-permeable membrane biofilm reactors (MBfRs), which efficiently supply N₂O in a bubbleless form directly to a biofilm grown on a gas-permeable membrane. A ¹⁵N tracer test indicated that the supply of N₂O resulted in an enriched biomass with a higher N₂O sink potential. Quantitative PCR and 16S rRNA amplicon sequencing revealed Clade II nosZ type-carrying N₂O-reducing bacteria as protagonists of N₂O sinks. Shotgun metagenomics showed the genetic potentials of the predominant Clade II nosZ-carrying bacteria, Anaerolineae and Ignavibacteria in MBfRs. Gemmatimonadota and non-anammox Planctomycetota increased their abundance in MBfRs despite their overall lower abundance. The implication of N₂O as an inhibitory compound scavenging vitamin B12, which is essential for the synthesis of methionine, suggested its limited suppressive effect on the growth of B12-dependent bacteria, including N₂O reducers. We identified Dehalococcoidia and Clostridia as predominant N₂O sinks in an anammox consortium fed exogenous N₂O because of the higher metabolic potential of vitamin B12-dependent biosynthesis.

Increasing concerns for global environmental protection against nitrogen pollution has led to stricter emission standards and the consideration of biological nitrogen removal processes (McCarty, 2018). Anammox-based systems, e.g., partial nitritation-anammox (PNA), are energy-saving and low-cost nitrogen removal processes that represent an alternative to conventional nitrification/denitrification. Since PNA reduces electricity consumption by up to 60% and lowers operation costs due to fewer requirements for aeration and external organic carbon (Gilbert et al., 2015), more than 100 full-scale plants have hitherto implemented PNA in their operation since 2014 (Lackner et al., 2014). The number of anammox-based systems, including PNA, for municipal and industrial wastewater treatment is increasing (Li et al., 2018; Al-Hazmi et al., 2023).

One of the challenges associated with anammox-based systems is the emission of nitrous oxide (N₂O). N₂O has a 273-fold higher global warming potential than CO₂ and is also known as an ozone-depleting substance (Smith et al., 2021). Partial nitritation, the first step of PNA, is a hotspot for N₂O exhaustion triggered by low oxygen, but high nitrite concentrations (Domingo-Felez and Smets, 2019). Intensive N₂O emissions potentially offset the reduction in energy consumption and, in some cases, contrarily exceed the carbon footprint required for conventional nitrification/denitrification (Fenu et al., 2019).

Aeration control is a common strategy to deter N₂O production in conventional wastewater treatment processes (Domingo-Felez and Smets, 2019; Duan et al., 2021). This precautionary strategy is based on preventing the accumulation of nitrite, the primary source of N₂O production by ammonia-oxidizing and denitrifying bacteria (Peng et al., 2017). Nevertheless, this control measure is not a panacea, resulting in N₂O emissions in some cases (Han et al., 2023). In contrast, strategies that reduce N₂O produced via biotic and abiotic pathways have been proposed to mitigate N₂O emissions. In one countermeasure strategy, gaseous N₂O in an off-gas line was fed to a bio-scrubber or biofilter, in which N₂O was converted into harmless nitrogen gas by N₂O-reducing bacteria (Frutos et al., 2016; Yoon et al., 2017; Han et al., 2023). This concept has been successfully demonstrated in lab- (Yoon et al., 2017) and pilot-scale studies (Han et al., 2023), and is a promising option to reduce N₂O emissions from wastewater treatment plants (WWTPs) (Frutos et al., 2016; Yoon et al., 2017). In both strategies, harnessing N₂O-reducing bacteria is vital to accomplish the reduction of N₂O.

Regardless of the broad range of nitrogen loads, a consistent line-up of bacteria, a.k.a. the core microbiome, has been detected (Lawson et al., 2017; Keren et al., 2020; Xiao et al., 2021). Furthermore, non-denitrifying N₂O-reducing bacteria, which are promising candidates as N₂O sinks and do not possess either or both nitrite reductase and nitric oxide reductase (Sanford et al., 2012; Shan et al., 2021), have been identified (Lawson et al., 2017). In the core microbiome, some N₂O-reducing bacteria, consisting of Clade I and Clade II types, in an anammox biomass exhibited high activities devoid of external organic carbon sources as broadly available electron donors (Suenaga et al., 2021). Based on their transcriptional activities, Anaerolineaceae (Clade II) and Burkholderiaceae (Clade I), counted as part of the core microbiome (Xiao et al., 2021), potentially play an essential role in N₂O consumption in anammox reactors (Suenaga et al., 2021). A metagenomic approach provides a more detailed understanding of the ecological and physiological functions of the core microbiome. Previous studies elucidated metabolic potentials (Speth et al., 2016; Lawson et al., 2017; Oshiki et al., 2022a) by tracking the uptake of carbon labeled with radioactive and stable isotopes (¹⁴C [Kindaichi et al., 2012] and ¹³C [Lawson et al., 2021]), indicating the interdependence between anammox and heterotrophic bacteria in carbon metabolism, e.g., vitamins, in anammox reactors (Lawson et al., 2017; Keren et al., 2020; Xiao et al., 2021). However, this physiological interaction of N₂O-reducing bacteria with carbon and nitrogen compounds remains unclear and, thus, warrants further study.

Since N₂O-reducing bacteria use N₂O as an electron acceptor, an external supply of N₂O may promote the activity of N₂O-reducing bacteria when N₂O is a limiting factor. On the other hand, N₂O potentially inhibits bacterial growth, e.g., Paracoccus denitrificans, because it reacts with vitamin B12 and deters methionine biosynthesis initiated from vitamin B12 at extracellular N₂O concentrations >2.8‍ ‍mg N L^–1 (Sullivan et al., 2013). This concentration range was observed in an up-flow column reactor (Suenaga et al., 2021) and the anaerobic regions of anammox granules (Okabe et al., 2011). This suppression may be crucial in an anammox community in which most bacteria are interdependent on amino acids and vitamins provided by anammox bacteria (Keren et al., 2020), particularly vitamin B12 (Lawson et al., 2017; Oshiki et al., 2022a). Under these conditions, the supply and retention of additional exogenous N₂O is a potential inhibitor. These controversial effects of N₂O on N₂O-reducing bacteria in an anammox community warrant thorough investigation with the goal of mitigating N₂O emissions from anammox-based systems.

Therefore, the present study attempted to identify the phylogeny of N₂O-reducing bacteria in an anammox reactor and characterize their metabolic functions based on their genotypes. Since the growth of N₂O-reducing bacteria in an anammox community is limited by the supply of N₂O (Suenaga et al., 2021), we hypothesize that an external N₂O supply leads to the dominance of fast-growing N₂O-reducing bacteria under autotrophic conditions. To verify this hypothesis, we operated bioreactors designed to supply sufficient N₂O without bubble formation via a gas-permeable membrane (Kinh et al., 2017; Suenaga et al., 2019). We examined the effects of the exogenous N₂O supply on microbial community compositions and functions by the side-by-side operation of bioreactors with or without a N₂O supply with synthetic media containing ammonia and nitrite. 16S ribosomal RNA (rRNA) gene amplicon sequencing and shotgun metagenomic sequencing were performed for this evaluation.

Materials and Methods

Reactor set-up and operation

Two membrane biofilm reactors (MBfRs) (Suenaga et al., 2019) were developed and applied for the enrichment of N₂O-reducing bacteria (Fig. S1). Each MBfR had liquid and gas compartments between which a flat-sheet silicon gas-permeable membrane was inserted. Their volume (0.26 L) and dimensions are referred to in a previous study (Kinh et al., 2017). One MBfR, Reactor 1 (w/N₂O), was supplied 5% (v/v) N₂O (base gas: N₂) at 5 kPa as a feeding gas from the gas compartment (on the bottom) to the biomass grown on a flat-sheet gas-permeable silicone membrane (L×W of 170×30‍ ‍mm with a wall thickness of 1‍ ‍mm; Rubber) (Fig. S1). The other MBfR, Reactor 2 (w/o N₂O), had the same configuration and dimensions, but was not supplied with N₂O. A biomass from an up-flow column-bed anammox reactor (Suenaga et al., 2021) was inoculated into Reactors 1 and 2.

The MBfRs were operated in a thermostatic chamber at 30°C for 1237 days. The medium was continuously supplied, and the liquid was recirculated through ports on the side walls of the MBfRs (Fig. S1). The medium used contained (L^–1 of distilled water) 100‍ ‍mg N of NH₄⁺, 100‍ ‍mg N of NO₂^–, 540‍ ‍mg of NaHCO₃, 27‍ ‍mg of KH₂PO₄, 300‍ ‍mg of MgSO₄·7H₂O, and 180‍ ‍mg of CaCl₂·2H₂O. One milliliter of trace element solutions with compositions described elsewhere (deGraaf et al., 1996) was added to per liter of the medium. The mixed medium was continuously purged with N₂ gas to eliminate dissolved oxygen, and oxygen in the medium influent tank was maintained at a low concentration. The hydraulic retention time (HRT) was consistently set at 1 day.

A sample was taken from the influent and effluent ports (Fig. S1) and stored after filtration through a 0.45-μm membrane filter (A045A025A; Advantec). NH₄⁺, NO₂^–, and NO₃^– concentrations were measured by ion chromatography (ICS1000 and ICS90; Thermo Fisher Scientific). In addition, 1‍ ‍mL of the mixed medium was sampled into a gas vial (13‍ ‍mL) sealed with a butyl rubber stopper filled with 12‍ ‍mL of pure nitrogen gas to measure the gaseous concentration of N₂O. The dissolved concentration of N₂O was assessed by the liquid-gas equilibrium method, as previously reported (Isobe et al., 2011; Riya et al., 2012; Suenaga et al., 2021), collecting a gaseous sample for the measurement. The concentration of N₂O in the headspace was measured by gas chromatography-quadrupole mass spectrometry (GCMS-QP2010 Ultra; Shimadzu).

Batch test using a ¹⁵N-labeled tracer

A ¹⁵N tracer test was performed using the biomass collected from Reactors 1 and 2 on day 776 to evaluate gross anammox, N₂O production, and N₂O consumption activities. In the evaluation, 40‍ ‍mg N L^–1 of ¹⁵N-labeled ¹⁵NO₂^– (98% labeled; Shoko Science) and 40‍ ‍mg N L^–1 of unlabeled NH₄⁺ were used. Fifteen milliliters of the mixed medium, the same volume as that supplied to the MBfRs, was poured into a 30-mL vial. Approximately 0.3‍ ‍g of the dewatered biomass was inoculated, and ⁴⁴N₂O was added to evaluate the N₂O consumption potential. N₂O and N₂ concentrations in the headspace were measured by GCMS. Experiments were conducted in triplicate. The biomass pretreatment and measurement conditions are described in our previous study (Suenaga et al., 2021).

Since ¹⁵NO₂^– contained 98% of labeled ¹⁵N (¹⁵N fraction: F=0.98), it was assumed that ¹⁵NO₂^– was converted to ³⁰N₂ via ⁴⁶N₂O mediated by a heterotrophic denitrifying pathway and ¹⁵NO₂^– and non-labeled NH₄⁺ were converted into ²⁹N₂ by the anammox reaction. N₂O production, N₂O consumption, and anammox activities were obtained using Eqs. 2–4. In these equations, V_{obs, 44N2O}, V_{obs, 46N2O}, V_obs, _29N2, and V_{obs, 30N2} (note that V_{obs, 44N2O} is a negative value) were obtained by the linear approximation of changes in the concentrations of ⁴⁴N₂O, ⁴⁶N₂O, ²⁹N₂, and ³⁰N₂ from 4 to 15 h. Detailed calculations are described in our previous study (Suenaga et al., 2021).

$${\mathrm{N}}_2\mathrm{O}\mathrm{ }\text{production activity}=\left(V_{obs,30N2}+V_{obs,46N2O}\right)\bullet F^{-2}$$(Eq. 1)

$${\mathrm{N}}_2\mathrm{O}\mathrm{ }\text{consumption activity}=\mathrm{ }{V}_{obs,30N2}\bullet F^{-2}-V_{obs,44N2O}$$(Eq. 2)

$$\text{Anammox activity}$$$$=\mathrm{ }\left(V_{obs,29N2}-2\bullet V_{obs,30N2}\bullet F^{-1}\left(1-F\right)\right) F^{-1}$$ (Eq. 3)

Spatial localization of targeted bacteria by fluorescence in situ hybridization (FISH)

Biomass samples on day 1044 were subjected to FISH. The floc biomass was collected from Reactor 1 (w/N₂O) and Reactor 2 (w/o N₂O) (Fig. S2), followed by immediate fixation with 4% paraformaldehyde. Fixation and subsequent hybridization procedures were performed as previously described (Terada et al., 2013). The oligonucleotide probes applied and formamide percentages are shown in Table S1. A microscopic ana­lysis was performed using a confocal laser scanning microscope (LSM 900; Carl Zeiss) with a Diode Laser (488, 561, and 640‍ ‍nm) and Airyscan 2. Image processing was performed using ZEN 3.0 (blue edition) (Carl Zeiss) and Imaris 10.0 (Oxford Instruments).

Taxonomy compositions and functional gene abundance

The biomass was routinely collected for a phylogenetic ana­lysis targeting the 16S rRNA gene and the quantification of functional genes. The targeted genes were the nirK and nirS genes that encode nitrite reductase, the cnorB and qnorB genes that encode nitric oxide reductase, and the Clade I nosZ and Clade II nosZ genes that encode N₂O reductase (NOS). DNA extraction was conducted using the FastDNA Spin Kit for Soil (MP Biomedicals) according to the manufacturer’s instructions. Functional gene quantification was performed using the CFX96 Real-Time PCR Detection System (BioRad Laboratories). The corresponding primers and sequences are summarized in Table S2, and PCR conditions are described in the Supplementary Information (SI).

PCR amplification of the V4 hypervariable region of the 16S rRNA gene was conducted using the primer set 515f-806r (Table S2). PCR conditions and library preparation procedures are described in SI. The DNA libraries generated and the initial control (bacteriophage PhiX; Illumina) were sequenced with a 300-cycle MiSeq Reagent kit (version 2, Illumina) using a MiSeq DNA sequencer (Illumina) in the paired-end sequencing mode.

Sequence data were analyzed using the following bioinformatics tools: removal of the adapter and low-quality reads were conducted using BBMap/BBduk (v. 38.84) (Bushnell, 2014), and trimmed reads were merged using FLASh (version: 2.2.00) (Magoc and Salzberg, 2011). Merged reads were imported into Dada2 (Prodan et al., 2020) (version: 1.26.0). PhiX sequences were removed from imported reads with Dada2’s FilterAndTrim command, and the reminders were then clustered into amplicon sequence variant (ASV) inference (default parameters), followed by the elimination of chimeras (default parameters). Samples with yields >10000 reads in total after the elimination of chimeras were used for a downstream ana­lysis. 16S rRNA ASVs imported into Qiime2 (Bolyen et al., 2019) (q2cli: 2023.5.1) were assigned by lineages using the QIIME 2 q2-feature-classifier plugin (Bokulich et al., 2018) with the pre-trained Silva (138.1) database (Quast et al., 2013; Robeson et al., 2021). Diversity and statistical ana­lyses were conducted using R version 4.2.2 (2022-10-31) and the phyloseq package (McMurdie and Holmes, 2013). The parameters used in the ana­lysis are listed in SI and Table S3-S6.

Reconstruction and ana­lyses of the metagenome-assembled genome

Three biomass samples, two from Reactor 1 (w/N₂O) and one from Reactor 2 (w/o N₂O) were collected on day 981 for shotgun metagenomic sequencing. DNA was extracted using a phenol-chloroform method (Butler, 2012; Yasuda et al., 2020). Biomass samples were centrifuged (10000‍ ‍rpm), and TE buffer (10‍ ‍mM Tris/HCl and 10‍ ‍mM EDTA, pH=8) and 10% SDS were added to the pelleted biomass. Genomic DNA was purified by repeating DNA and protein separation using phenol, chloroform, and CTAB/NaCl solution. RNA as a contaminant in genomic DNA was decomposed by RNaseA (TaKaRa Bio). After ethanol precipitation, genomic DNA was suspended in TE buffer and stored in a freezer until used. Library preparation and sequencing were performed at Azenta Life Science. Sequencing was performed on a Novaseq (Illumina) with a 150-bp paired-end sequencing protocol, and 10 Gb of data was obtained per sample.

The metagenomic pipeline employed in the present study is shown in Fig. S3. A quality check of raw sequencing data and low-quality read trimming were performed using fastp v0.22.0 (Chen et al., 2018). Trimmed reads were assembled by Megahit (v1.2.9) (Li et al., 2015) and metaSPAdes (v3.13.1) (Nurk et al., 2017) in parallel with default parameters. Contigs and scaffolds were filtered using SeqKit (v2.2.0) (Shen et al., 2016), and those longer than 500 bp were used in subsequent ana­lyses.

To analyze the genomic profile of the nosZ gene in metagenome samples, gene predictions in filtered contigs and scaffolds were performed by Prodigal (v. 2.6.3) (Hyatt et al., 2010). Predicted genes from all biomass samples were integrated to be non-redundant by Cd-hit (v. 4.8.1) (Li and Godzik, 2006) with ‘-c 0.95‍ ‍-aS 0.9 -g 1’. Function assignments were conducted using EggNOG mapper (v2.1.9) (Cantalapiedra et al., 2021) with the diamond mode and InterProScan (v. 5.65–97.0) (Jones et al., 2014; Blum et al., 2021). Quality-trimmed reads were mapped onto the assembled contigs using BWA-MEM (v. 2.2.1) (Vasimuddin et al., 2019) with default parameters. The resulting sequence alignment mapping files were sorted using SAMtools (v. 1.13) (Danecek et al., 2021), and the read coverage of each contig was calculated using featureCounts (v. 2.0.3) (Liao et al., 2013). Read coverage was further normalized by the total number of mapped reads in each sample, yielding reads per kilobase per million mapped reads (RPKM) values.

To reconstruct the metagenome-assembled genome (MAG), filtered contigs and scaffolds were grouped into primary-bins with MaxBin (v2.2.6) (Wu et al., 2016), MetaBAT (v2.12.1) (Kang et al., 2019), and CONCOCT (v 1.0.0) (Wu et al., 2016) with default parameters. The bins were consolidated into final bins by DASTools (v1.1.4) (Sieber et al., 2018) using default parameters.

Redundant bins generated in parallel were de-replicated by dRep (v3.3.1) (Olm et al., 2017) with ‘-l 50000 -pa 0.90 -sa 0.99 -comp 50 -con 25 -nc 0.1’. CheckM2 (v1.0.1) (Chklovski et al., 2022 CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning. bioRxiv: https://doi.org/10.1101/2022.07.11.499243) was used to check the quality of bins, and bins with completeness <70% and contamination >5% were excluded. The presence of 16S/23S/5S rRNA was checked with barrnap (v0.9) (Seemann, 2018), and the numbers and types of tRNAs were counted using tRNAscan-SE (v 2.0.9) (Chan et al., 2021). Taxonomy assignment was conducted using GTDB-Tk (v2.1.1) (Chaumeil et al., 2022) with GTDB r207 (Parks et al., 2022). DFAST (v1.2.14) (Tanizawa et al., 2018) was used for gene predictions and functional annotations. Additional annotations for function assignments were conducted using EggNOG mapper (v2.1.9) (Cantalapiedra et al., 2021) with a diamond mode and KofamScan (v 1.3.0) (Aramaki et al., 2020). Annotation results were merged based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthology (KO) and mapped to the KEGG database (Kanehisa and Goto, 2000) using KEGG Decoder (v1.3) (Graham et al., 2018). The relative abundance of bins in each sample was calculated using CoverM (v0.6.1). Since two biomass samples were collected from Reactor 1 (w/N₂O), the average relative abundance of the two samples was used.

Data availability

Raw 16S rRNA gene amplicon and metagenomic sequencing data are available in the DNA Data Bank of Japan (DDBJ) nucleotide sequence database under accession numbers DRA016674 and DRA016675, respectively. Assembled and annotated MAGs were deposited in the DDBJ nucleotide sequence database with the accession numbers shown in Table S7.

Results

Reactor operation and nitrogen removal performance

Similar changes were observed in effluent NH₄⁺, NO₂^–, and NO₃^– concentrations in Reactors 1 (w/N₂O) and 2 (w/o N₂O) (Fig. S4 and S5). In the first 90 days of the incubation, approximately 50‍ ‍mg N L^–1 of NH₄⁺ and NO₂^– remained in the effluent of both reactors. NH₄⁺ and NO₂^– concentrations in the effluent then decreased in both reactors, reaching average NO₂^– and NH₄⁺ concentrations of 18.2 and 20.7‍ ‍mg‍ ‍N‍ ‍L^–1, respectively, from day 100 to 260 in Reactor 1 and 9.76 and 25.1‍ ‍mg N L^–1, respectively, in Reactor 2. After day 295, NH₄⁺ and NO₂^– concentrations in the effluent further decreased in Reactors 1 and 2. However, NH₄⁺ and NO₂^– accumulated in the effluent from day 310 to 405 (Fig. S4b and S5b), possibly due to the lack of maintenance during the lockdown period of COVID-19. After day 426, NO₂^– and NH₄⁺ concentrations again decreased (16.5 and 12.6‍ ‍mg N L^–1, respectively, in Reactor 1 and 18.3 and 13.8‍ ‍mg N L^–1, respectively, in Reactor 2 on average). Influent and effluent pH were 7.90±0.21 and 8.17±0.41, respectively, in Reactor 1 and 7.92±0.31 and 8.13±0.30, respectively, in Reactor 2. The exogenous supply of N₂O markedly affected dissolved N₂O concentrations in the bulk liquid. The dissolved concentration of N₂O in the bulk liquid was 3.96±2.67‍ ‍mg N L^–1 on average in Reactor 1 with a maximum of 8.95‍ ‍mg N L^–1 (day 805) and a minimum of 1.31‍ ‍mg N L^–1 (day 864) (Fig. S6). The dissolved concentration of N₂O was markedly lower in Reactor 2 (Fig. S7), with an average of 0.67±0.38‍ ‍mg N L^–1, maximum of 1.72‍ ‍mg N L^–1 (day 776), and minimum of 0.34‍ ‍mg N L^–1 (day 1091). Since the total concentration of nitrogen in the influent was 200‍ ‍mg N L^–1, the N₂O conversion ratio over total nitrogen in Reactor 2 was 0.34% under the assumption of marginal N₂O exhaustion to the gaseous layer due to the absence of the headspace, which was similar to that in an anammox column reactor (Suenaga et al., 2021). The stoichiometric ratios of NO₂^– consumption and NO₃^– production over NH₄⁺ consumption in both reactors were similar to those of enriched anammox bacteria (Strous et al., 1998) (Fig. S4c and S5c), suggesting that an anammox reaction was responsible for biological nitrogen removal.

Regardless of the presence or absence of a N₂O supply to the reactors, their biomasses were mostly aggregated, consisting of granules or flocs. Although both reactors were constantly stirred by liquid recirculation, some of the biomass sedimented onto the gas-permeable membrane, while the remainder continued to be suspended in the bulk liquid or formed a biofilm on the reactor sidewall or gas-permeable membrane surface (Fig. S2a). However, the biofilm mass of these parts was insufficient to extract DNA (Fig. S2b) for routine functional gene quantification or a phylogenetic ana­lysis. Therefore, samples for DNA extraction were collected from the aggregated biomass in a suspension.

Intrinsic activity test

Biomass-specific anammox rates were 0.90±0.22‍ ‍mg N [g‍ ‍MLVSS]^–1 h^–1 for Reactor 1 (w/N₂O) and 0.97±0.28‍ ‍mg N [g MLVSS]^–1‍ ‍h^–1 for Reactor 2 (w/o N₂O), with no significant difference (P=0.735) (Fig. 1). Similar N₂O production rates were attained: 0.037±0.014‍ ‍mg N [g MLVSS]^–1 h^–1 for Reactor 1 (w/N₂O) and 0.039±0.0028‍ ‍mg N [g MLVSS]^–1‍ ‍h^–1 for Reactor 2 (w/o N₂O). N₂O consumption rates were 0.21±0.064‍ ‍mg N [g MLVSS]^–1 h^–1 for Reactor 1 (w/o N₂O) and 0.14±0.014‍ ‍mg N [g MLVSS]^–1 h^–1 for Reactor 2 (w/o N₂O). Although no significant differences were observed in N₂O production (P=0.930) or consumption (P=0.452), an increase was noted in the N₂O consumption activity of the biomass fed N₂O.

Fig. 1.

Biomass-specific (a) N₂O production and N₂O consumption, (b) anammox activities calculated from the ¹⁵N tracer test. Experiments were conducted in triplicate for reproducibility. Bars and error bars represent mean values and standard deviations, respectively. In the statistical ana­lysis, Welch’s t-test was performed (P>0.05).

Quantification of functional gene dynamics

The dynamics of functional gene abundance are shown in Fig. 2. The continuous external supply of N₂O allowed for the distinct dynamics of denitrifying genes. Gene copies of nirK, nirS, cnorB, and qnorB increased to day 280. Gene copies of nirK and nirS remained constant in Reactor 1 (w/o N₂O), while those in Reactor 2 (w/o N₂O) decreased from those in the inoculum. nirK gene abundance was similar in both reactors, whereas that of the nirS counterpart was significantly higher in Reactor 1 than in Reactor 2 after day 280, except on days 718 and 810 (P<0.05). The continuous external supply of N₂O increased nosZ gene abundance. Despite the presence or absence of the external N₂O supply, the Clade II nosZ gene was one order of magnitude higher than the Clade I nosZ gene. In contrast, Clade II nosZ gene copies in Reactor 2 (w/o N₂O) were lower on most sampling dates than those on day 0, which was opposite to that attained in Reactor 1 (w/N₂O). Clade II nosZ gene copies were significantly higher in Reactor 1 than in Reactor 2 after day 280 (P<0.05).

Fig. 2.

Time courses of gene densities by quantitative PCR (n=3): (a) nirS, (b) nirK, (c) cnorB, (d) qnorB, (e) Clade I nosZ, and (f) Clade II nosZ genes normalized by DNA weight.

Taxonomy compositions

The top 6 taxa at the phylum level during the entire incubation period are shown in Fig. 3. In both reactors, the phyla Planctomycetota and Chloroflexota accounted for more than 50% at all periods, except on day 246. The relative abundance of anammox bacteria (class Brocadiales) was higher in Reactor 2 than in Reactor 1. The relative abundance of anammox bacteria in Reactor 2 (w/o N₂O) was on average 24.8% for the period before day 385 and 51.2% after. Within the class Brocadiales, Candidatus Jettenia was the dominant bacterial species, regardless of the supply of N₂O.

Fig. 3.

Evolution of microbial community compositions by 16S rRNA gene amplicons in Reactors 1 (w/N₂O supply) and 2 (w/o N₂O supply). The top six phyla and genera during the overall incubation are listed.

Chloroflexota was the dominant phylum among non-anammox bacteria, with Anaerolineae (Reactor 1: 6.3–48.1% of the total, Reactor 2: 5.6–45.6%) and Ardenticatenaceae (Reactor 1: 1.9–48.1%, Reactor 2: 5.6–45.6%) as the predominant classes. Other members were‍ ‍Ignavibacteria, Betaproteobacteria/Burkholderiales, Fimbriimonadia, Gammaproteobacteria/Xanthomonadales, and Gammaproteobacteria/Phycisphaerae. The relative abundance of these non-anammox bacteria, potentially regarded as denitrifiers, was higher in Reactor 1 than in Reactor 2. The relative abundance of non-anammox bacteria changed over time, accounting for an average of approximately 60% and reaching a maximum of 77% in Reactor 1 after day 426, when nitrogen removal performance stabilized.

The bacterial community compositions of the biofilm adhering to the membrane and the aggregate biomass deposited onto the membrane-bound biofilm taken from Reactor 1 on day 199 (Fig. S2) are shown in Fig. S8. The major taxa constituting the community were analogous; however, their compositions differed. Anaerolineae was the most dominant biofilm component, of which the ASV assigned to the genus SBR1031 had a higher relative abundance in the biofilm (32.6%>9.8% in the aggregate biomass). On the other hand, the relative abundance of the family Brocadiaceae was lower in the biofilm (3.5%) than in the aggregate biomass (15.7%).

Activity ana­lysis with FISH imaging

FISH ana­lyses showed sufficient fluorescence in samples taken from Reactor 1 (w/N₂O) (Fig. 4a, b, c, d, e, f, and g). Consistent with amplicon sequencing results, anammox and Chloroflexota were abundant. The spatial distribution of these bacteria in the floc did not markedly differ between the two reactors (data not shown). The mesh-like structure of the phylum Chloroflexota, possessing a filamentous cell morphology, was distributed over the floc, particularly in the floc exterior (Fig. 4a, b, c, d, and e). In contrast, anammox bacteria were present inside the floc. Bacteria other than Chloroflexota and anammox bacteria were detected near filamentous cells.

Fig. 4.

Spatial distributions of anammox and Chloroflexi by fluorescence in situ hybridization, taken by confocal laser scanning microscopy under 630× magnification. Anammox and Chloroflexi were hybridized by probes of Amx 368 and CFX1228+GNSB941 with fluorochromes of Cy5 and Cy3, counterstained by the EUB338 mix with the fluorochrome of FITC. Cy5 is shown in blue, Cy3 in red, and FITC in green. Panels a, b, c, d, e, and f: Reactor 1 (with N₂O supply). Panel g: Reactor 2 (without N₂O supply). In Panels a, b, c, d, and e, the number in the upper right in each image indicates the distance from the aggregate surface.

Analyses of MAGs

Of the reconstructed bins, after quality screening (completeness ≥70%, contamination <5%), 58 MAGs were acquired (Fig. 5 and Table S8). MAGs detected with higher‍ ‍relative abundance (>1.5-fold) in Reactor 1 than in Reactor 2 were assigned to Vicinamibacteria (bin 57), Chthonomonadetes (bin 19), Anaerolineae (bin 60), Dehalococcoidia (bin 23), Clostridia (bin 30), Gemmatimonadetes (bins 55 and 53), Phycisphaerae (bin 22), Burkholderiales (bin 52), Rhodocyclales (bin 45), Nevskiales (bin 37), and Xanthomonadales (bin 59). Abundant MAGs‍ ‍in Reactor 1 did not necessarily harbor either of the‍ ‍nosZ genes. It is important to note that all of the retrieved MAGs harboring the nosZ gene did not possess a complete set of denitrifying genes, suggesting that they were non-denitrifying genotypes.

Fig. 5.

Pathway coverage of high quality (completeness ≥70%, contamination <5%) metagenome-assembled genomes. The dot indicates metabolic pathway completeness calculated using the KEGG Decoder based on KEGG Orthology (KO) results assigned by DFAST, EggNOG mapper, and KofamKOALA.

MAGs (bins 48, 49, 50, and 51) were affiliated within Ignavibacteriaceae, a taxon containing bacteria frequently detected in anammox reactors that has been suggested to function as a N₂O sink (Lawson et al., 2017) harboring the Clade II nosZ gene and dissimilatory nitrate reduction to ammonium (DNRA) gene (nirBD and/or nrfAH). In addition, bins 48 and 51 carried the narGHI genes encoding nitrate reductase. The MAG assigned to Bacteroidia (bin 79), which has potential as a N₂O sink, harbored the Clade II nosZ gene and possessed the norBC and narGHI genes.

Anaerolineae, one of the most dominant non-anammox bacterial genera, was annotated with the narGHI gene, nirS and nirK genes, and a group of genes related to DNRA. Dehalococcoidia harbored the nitrate reductase nar gene,‍ ‍the nir gene, and the Clade II nosZ gene. Gemmatimonadetes (bins 55 and 53), showing a highly transcribed Clade II nosZ gene in anammox reactors as previously reported (Suenaga et al., 2021), contained the nirK gene encoding nitrite reductase, the narGHI gene, and genes‍ ‍involved in DNRA. The MAGs of Planctomycetota (bins 77, 58, and 47) were also annotated with the Clade II nosZ gene. The MAGs abundantly detected in Reactor 1 (w/N₂O) did not necessarily harbor any nosZ (i.e., bins 19, 60, 6, 45, 37, and 59). The abundance of the nosZ-coding sequences collected from the assembled contigs revealed that the higher RPKM values derived from Clade II nosZ (174.2 RPKM in Reactor 1 and 164.0 RPKM in Reactor 2) than from Clade I nosZ (56.5 RPKM in Reactor 1 and 44.5 RPKM in Reactor 2) were retrieved (Fig. S9).

Genes involved in the metabolism and recycling of peptidoglycans involved in carbon degradation and alpha-amylase and beta-glucosidase, which play a role in the metabolism of polysaccharides, a constitute of extracellular polymeric substances (EPS), were annotated in Armatimonadota (bins 61, 19, and 81), Ignavibacteria, Gemmatimonadetes, Myxococcota, and Planctomycetota. Chloroflexota MAGs differed in their genotypic patterns associated with carbon degradation at the phylum level. For‍ ‍example, the Anaerolineae MAG (bin 60) possessed the‍ ‍genes encoding enzymes for peptidoglycan degradation,‍ ‍alpha-amylase and beta-glucosidase, whereas the Dehalococcoidia MAG (bin 80) carried an incomplete set of functional genes encoding enzymes for peptidoglycan degradation (Fig. 5).

The genes involved in cobalamin biosynthesis (Corrin ring biosynthesis and adenylation and nucleotide loop assembly) were well annotated in Dehalococcoidia, Clostridia (bin 30), and Brocadiae MAGs, suggesting the presence of important vitamin B12 producers in the anammox community. The coverage of these MAGs was Dehalococcoidia (bin 85: 0.08%, bin 23: 1.73%) and Clostridia (bin 30: 0.23%). Burkholderiales (bins 66, 72, and 78), Anaerolineae, and Nitrospiraceae (bin 67) MAGs harbored the genes responsible for a vitamin B12 transporter and nucleotide loop assembly for Cobalamin biosynthesis (3) with high annotation (>60%).

Discussion

Our reactor design with a flat-sheet gas-permeable membrane allowed a bubbleless N₂O supply to the biomasses in the MBfRs. Continuous incubations by the MBfRs successfully provided environments with high (Reactor 1) and low (Reactor 2) concentrations of N₂O (Fig. S6 and S7). The concentration of N₂O in Reactor 1 (3.96‍ ‍mg N L^–1 on average, i.e., 141‍ ‍μM) was more than one order of magnitude higher than the apparent half-saturation constant for N₂O (K_m,_N2O) of 9.50±3.0‍ ‍μM (0.266±0.083‍ ‍mg N L^–1) of an enriched anammox biomass (Suenaga et al., 2021). Therefore, this ensures an environment in which fast-growing N₂O reducers with a low affinity for N₂O, but a high N₂O consumption rate (Andrews and Harris, 1986; Yin et al., 2022) preferentially grow. This study demonstrated that the application of an MBfR concept (Nerenberg, 2016) to enrich highly efficient N₂O-reducing bacteria in an anammox biomass secures a stable environment with high or low concentrations of N₂O with the biomass. The maintenance of a different N₂O level for a long period was essential in the present study, while we previously conducted a short-term (within 1 day) biomass exposure to a high N₂O concentration (Suenaga et al., 2021).

By using the MBfR concept, the present study investigated whether an external N₂O supply led to the dominance of fast-growing N₂O-reducing bacteria. According to the microbial community compositions in the two MBfRs with and without the external N₂O supply (Fig. 3), slight changes in microbial community compositions were attained. This result refutes our hypothesis that an external N₂O supply leads to the dominance of fast-growing N₂O-reducing bacteria. Nevertheless, the biomass from Reactor 1 showed a higher N₂O consumption rate than that from Reactor 2 (Fig. 1). The quantitative values of Clade I and Clade II nosZ genes indicated that the supply of N₂O as an external electron acceptor was advantageous to the bacterial community harboring nosZ. This effect was significantly pronounced for Clade II nosZ (Fig. 2f). Although the contribution of the external N₂O supply to microbial community changes may be limited (Fig. 3), the N₂O supply was instrumental in the increases observed in Clade II nosZ gene abundance (Fig. 2f) and intrinsic N₂O consumption activities (Fig. 1).

An autotrophic environment with a high nitrogen concentration likely favors Clade II nosZ N₂O-reducing bacteria. This notion was supported by the present results showing that Clade II nosZ N₂O-reducing bacteria were consistently one order of magnitude and three-fold more abundant than the Clade I type in MBfRs regardless of the supply of N₂O by qPCR (Fig. 2e and 2f) and the metagenomic ana­lysis (Fig. S9), respectively. This result may be attributed to an inoculum being highly enriched under autotrophic conditions (Suenaga et al., 2021). Previous studies on N₂O-reducing bacteria harboring either Clade I or Clade II nosZ genes indicated that Clade II nosZ bacteria exhibited higher affinities (Yoon et al., 2016; Suenaga et al., 2019), which was not consistent with our results, where even a high concentration of N₂O allowed a higher abundance of Clade II nosZ bacteria. Clade II nosZ bacteria were found to be dominant during an incubation with a N₂O supply (Suenaga et al., 2019). These findings and the present results indicate that the concentration of N₂O during the enrichment of N₂O-reducing bacteria is not the sole factor affecting the predominance of the clade of the nosZ gene.

Anaerolineae and Ignavibacteria, which were dominant among non-anammox bacteria in the present study (Fig. 3), may be critical N₂O reducers in anammox systems. They reportedly utilize endogenous organic matter when fed an organic-free medium, and are also the most prevalent when fermentable organic matter is supplied at low C/N ratios (Xiao et al., 2021). Ignavibacteria MAGs (bins 48, 49, 50, and 51) and Anaerolineae MAG (bin 80) were annotated with metabolic pathways regarding soluble microbial products (SMP) by KEGG Decoder (Ni et al., 2012), including EPS composed of polysaccharides and proteins secreted by anammox bacteria (Hou et al., 2015; Ali et al., 2018). The annotation count by KEGG Decoder accounted for over 65% of the total metabolic pathways. In addition, these associated genes were functionally assigned to carbon degradation to utilize cell debris (Fig. 5). Therefore, these taxa were likely to be N₂O consumers using SMP produced by anammox and others in the anammox reactor.

SBR1031 belonging to the class Anaerolineae showed a higher relative abundance in the biofilm grown on membrane surfaces (Fig. S8) with presumably high N₂O concentrations. The results of the metagenomic ana­lysis suggest that this genus contains bacteria harboring the Clade II nosZ gene (Bovio-Winkler et al., 2023), which may have a promising N₂O consumption capacity. Therefore, the class Anaerolineae has potential as a N₂O sink in the anammox system that is not susceptible to high N₂O concentrations. In contrast, a lower abundance of the family Brocadiaceae in the biofilm than the deposited aggregates on the biofilm outer surface (Fig. S8) suggested that a high N₂O concentration condition was unfavorable for anammox growth.

The phylum Chloroflexota coexisting with anammox bacteria was detected at the periphery of the biomass, surrounding anammox cell aggregates (Fig. 4). This spatial arrangement did not appear to change with or without the N₂O supply, which is consistent with previous findings (Wong et al., 2023). In addition, filamentous Chloroflexota has been suggested to function as a junction for biomass aggregates and degrade EPS secreted by anammox bacteria, which are responsible for decomposing EPS to SMP available to other heterotrophic bacteria (Wong et al., 2023). Our Anaerolineae MAG (bin 60) was also well annotated with pathways (alpha-amylase and beta-glucosidase) involved in the degradation of polysaccharide chains (Oshiki et al., 2022a) (Fig. 5). Our FISH ana­lysis also showed fluorescence derived from neither Planctomycetota nor Chloroflexota in the vicinity of Chloroflexota (Fig. 4), and its spatial coordination was similar to that observed in a previous study (Wong et al., 2023). Therefore, the preference and availability of recalcitrant organic matter were deemed necessary for elucidating the composition of the non-anammox bacterial population. This unexplored population may be one of the reasons why N₂O-reducing bacteria with faster N₂O consumption rates were more dominant than N₂O-producing bacteria.

Although the phyla Gemmatimonadota and non-anammox Planctomycetota were not predominant in the microbial community, they were frequently detected taxa in Reactor 1 (Fig. 5), suggesting their involvement in another N₂O sink. These phyla exhibit high transcription activities of the nosZ gene in anammox processes (Park et al., 2017; Suenaga et al., 2021). The higher coverage of MAGs belonging to these strains in Reactor 1 suggested that the physiological traits matched the environment in the MBfR with external N₂O supply, alluding to N₂O reducers favorable in environments with high N₂O concentrations. The production of N₂O during denitrification, often found in anammox processes as a minor microbial reaction among nitrogen transformation, may be attributed to a N₂O production rate via the reduction of NO₃^– and NO₂^– to N₂O, overwhelming the N₂O consumption rate. A Gemmatimonadetes MAG (bin 53), harboring narGHI and Clade II nosZ genes, possessed the metabolic potentials of NO₃^– and N₂O reduction (Fig. 5), potentially consuming N₂O depending on N₂O concentrations according to the biokinetics in previous study (Oshiki et al., 2022b). Moreover, non-anammox Planctomycetes (bins 77 and 47) possessed two denitrification-related genes, the napAB gene encoding nitrate reductase and the Clade II nosZ gene (Fig. 5). When bacteria carry the nap and nos genes, they preferentially use N₂O as an electron acceptor more than those with nar and nos genes (Gao et al., 2021; Oba et al., 2022), suggesting that non-anammox planctomycetes harboring nosZ genes are preferential N₂O sinks.

Theoretically, there has been a concern about the loss of nitrogen removal performance due to reduced microbial activity caused by N₂O, which has yet to be examined in detail. Excessive N₂O in Reactor 1 may react with cob(I)alamin (vitamin B12) as previously reported (Drummond and Matthews, 1994), thereby inhibiting the function of the cobalamin-dependent enzyme (Shelton et al., 2019). The N₂O level in Reactor 1 (3.96‍ ‍mg N L^–1) exceeded the concentration (>2.8‍ ‍mg N L^–1) at which N₂O reacts with MetH methionine synthase, the most widely used cobalamin-dependent enzyme, thereby retarding the cell growth of P. denitrificans (Sullivan et al., 2013). To ameliorate the inhibition of N₂O, P. denitrificans and Dehalococcoides mccartyi either activate vitamin B12-independent MetE methionine synthase or maintain its activity by an endogenous and exogenous vitamin B12 supply (Sullivan et al., 2013; Yin et al., 2019). Bacteroidia and Anaerolineae (bins 79 and 60), predominant taxa based on 16S rRNA gene amplicon sequencing (Fig. 3), did not possess a metE gene encoding vitamin B12-independent MetE methionine synthase (Fig. 5). Bacteria possessing only vitamin B12-dependent enzymes may be disadvantageous for survival under excessive N₂O conditions. Nevertheless, their abundance did not decrease in the presence of excessive N₂O in Reactor 1 (Fig. 3). The results of the FISH ana­lysis also indicated that the dominance of Chloroflexota was sustained (Fig. 4). Therefore, the anammox community in Reactor 1 appeared to have a bypassing mechanism by which vitamin B12 was relayed from its producers to acceptors. Further studies are needed to elucidate the underlying mechanisms.

Previous studies reported that de novo cobamide biosynthesis consists of approximately 30 steps, which may be broadly divided into the following steps: tetrapyrrole precursor biosynthesis, (aerobic/anaerobic) corrin ring biosynthesis and adenylation, and nucleotide loop assembly (Shelton et al., 2019; Lu et al., 2020; Balabanova et al., 2021). Furthermore, among an anammox community, only Brocadia sp. synthesizes cobalamin (Lawson et al., 2017; Keren et‍ ‍al., 2020). In the present study, only anammox Planctomycetes (bin 17) had complete de novo biosynthetic pathways, which was consistent with previous findings.

The results obtained herein demonstrated that Dehalococcoidia (bins 23 and 85) and Clostridia (bin 30) carried a nucleotide loop assembly pathway following the cobalamin precursor synthesis pathway. Some bacteria have been shown to partially possess de novo cobalamin biosynthesis pathways (Shelton et al., 2019; Lu et al., 2020). As shown in physiological studies (Schipp et al., 2013; Shelton et al., 2019), some Dehalococcoidia and Clostridia synthesize cobamides from salvaged precursors or change the cobamide structure for their growth. The results of our metagenomic ana­lysis suggest that Clade II nosZ N₂O-reducing bacteria, such as Dehalococcoidia and Clostridia, retain many genes associated with the cobalamin synthesis pathway. Putatively, they are prone to endogenously synthesize cobalamin and/or exogenously receive cobalamin supplied by Brocadia as a survival strategy. The present results imply the strong contribution of anammox bacteria as a de novo cobalamin producer to relay cobalamin to coexisting bacteria missing its production functions, thereby restricting bacteria from cobalamin inhibition caused by N₂O in the anammox community.

Conclusion

In the present study, we operated MBfRs that were externally supplied with N₂O for 1200 days for the continuous incubation of bacteria consuming N₂O. We demonstrated the long-term maintenance of a high abundance of Clade II nosZ and an increased capacity for N₂O consumption in the community enriched by an exogenous N₂O supply. On the other hand, slight differences were attained in the dominant taxa in the MBfRs with or without an external N₂O supply, which indicated that supplying N₂O as an additional electron acceptor did not necessarily result in the dominance of fast-growing N₂O-reducing bacteria. Therefore, other factors, particularly electron donors, limited the growth rate of heterotrophic N₂O-reducing bacteria. Despite growth-limiting conditions for N₂O-reducing bacteria, the metagenomic ana­lysis revealed significant functions to acquire organic matter and vitamins that serve as electron donors. Anaerolineae and Ignavibacteria, regarded as parts of the core microbiome and dominant in the MBfR supplied with N₂O, have a broad availability of organic matter and survive at high N₂O concentrations. Dehalococcoidia and Clostridia have partial vitamin B12 production pathways, likely carrying a survival strategy in excessive N₂O environments. Moreover, we found that non-anammox Planctomycetes appeared to contribute to N₂O consumption despite their low abundance in the anammox community. Since the concentration of N₂O in the MBfR system supplied with N₂O was markedly higher than that often observed in anammox reactors, there have been concerns that the anammox community may lose its activity due to the inhibition of N₂O. Nevertheless, the N₂O-fed reactor operation underpinned the anammox community, exerting a stable N₂O sink at excessively high N₂O concentrations. Our in-depth ana­lysis of microbial community compositions and functions provide insights on the as-yet-unknown functions of non-denitrifying Clade II nosZ bacteria. These ecophysiological descriptions will facilitate the utilization of N₂O-reducing bacteria inhabiting engineered systems that emit N₂O, such as anammox-based processes, and may be expanded to agricultural fields requiring N₂O mitigation. The may lead to the development of countermeasures against N₂O emissions, e.g., bioaugmentation by inoculating biomasses enriching N₂O reducers.

Citation

Oba, K., Suenaga, T., Yasuda, S., Kuroiwa, M., Hori, T., Lackner, S., and Terada, A. (2024) Quest for Nitrous Oxide-reducing Bacteria Present in an Anammox Biofilm Fed with Nitrous Oxide. Microbes Environ 39: ME23106.

https://doi.org/10.1264/jsme2.ME23106

Acknowledgements

We thank the late Ms. Kanako Mori for her experimental support. This research was funded by Grants-in-Aid for Scientific Research (Grant no. 20H04362 and 23H03565), Fostering Joint International Research (20KK0243) from the Japan Society for the Promotion of Science (JSPS), and the Kurita Water and Environment Foundation (22T012).

References

•  Al-Hazmi,  H.E.,  Lu,  X.,  Grubba,  D.,  Majtacz,  J.,  Badawi,  M., and  Mąkinia,  J. (2023) Sustainable nitrogen removal in anammox-mediated systems: Microbial metabolic pathways, operational conditions and mathematical modelling. Sci Total Environ 868: 161633.
•  Ali,  M.,  Shaw,  D.R.,  Zhang,  L.,  Haroon,  M.F.,  Narita,  Y.,  Emwas,  A.H., et al. (2018) Aggregation ability of three phylogenetically distant anammox bacterial species. Water Res 143: 10–18.
• Andrews JH, Harris RF. r- and K-Selection and Microbial Ecology. In: Marshall KC, editor. Advances in Microbial Ecology. Boston, MA: Springer US; 1986. p. 99-147.
•  Aramaki,  T.,  Blanc-Mathieu,  R.,  Endo,  H.,  Ohkubo,  K.,  Kanehisa,  M.,  Goto,  S., and  Ogata,  H. (2020) KofamKOALA: KEGG Ortholog assignment based on profile HMM and adaptive score threshold. Bioinformatics 36: 2251–2252.
•  Balabanova,  L.,  Averianova,  L.,  Marchenok,  M.,  Son,  O., and  Tekutyeva,  L. (2021) Microbial and genetic resources for cobalamin (vitamin B12) biosynthesis: From ecosystems to industrial biotechnology. Int J Mol Sci 22: 4522.
•  Blum,  M.,  Chang,  H.Y.,  Chuguransky,  S.,  Grego,  T.,  Kandasaamy,  S.,  Mitchell,  A., et al. (2021) The InterPro protein families and domains database: 20 years on. Nucleic Acids Res 49: D344–D354.
•  Bokulich,  N.A.,  Kaehler,  B.D.,  Rideout,  J.R.,  Dillon,  M.,  Bolyen,  E.,  Knight,  R., et al. (2018) Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier plugin. Microbiome 6: 90.
•  Bolyen,  E.,  Rideout,  J.R.,  Dillon,  M.R.,  Bokulich,  N.A.,  Abnet,  C.C.,  Al-Ghalith,  G.A., et al. (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol 37: 852–857.
•  Bovio-Winkler,  P.,  Guerrero,  L.D.,  Erijman,  L.,  Oyarzua,  P.,  Suarez-Ojeda,  M.E.,  Cabezas,  A., and  Etchebehere,  C. (2023) Genome-centric metagenomic insights into the role of Chloroflexi in anammox, activated sludge and methanogenic reactors. BMC Microbiol 23: 45.
• Bushnell, B. (2014) BBTools software package. URL http://sourceforge.net/projects/bbmap
• Butler JM. Chapter 2 - DNA Extraction Methods. In: Butler JM, editor. Advanced Topics in Forensic DNA Typing: Methodology. San Diego: Academic Press; 2012. p. 29–47.
•  Cantalapiedra,  C.P.,  Hernández-Plaza,  A.,  Letunic,  I.,  Bork,  P., and  Huerta-Cepas,  J. (2021) eggNOG-mapper v2: Functional annotation, orthology assignments, and domain prediction at the metagenomic scale. Mol Biol Evol 38: 5825–5829.
•  Chan,  P.P.,  Lin,  B.Y.,  Mak,  A.J., and  Lowe,  T.M. (2021) tRNAscan-SE 2.0: improved detection and functional classification of transfer RNA genes. Nucleic Acids Res 49: 9077–9096.
•  Chaumeil,  P.-A.,  Mussig,  A.J.,  Hugenholtz,  P., and  Parks,  D.H. (2022) GTDB-Tk v2: memory friendly classification with the genome taxonomy database. Bioinformatics 38: 5315–5316.
•  Chen,  S.F.,  Zhou,  Y.Q.,  Chen,  Y.R., and  Gu,  J. (2018) fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34: 884–890.
•  Danecek,  P.,  Bonfield,  J.K.,  Liddle,  J.,  Marshall,  J.,  Ohan,  V.,  Pollard,  M.O., et al. (2021) Twelve years of SAMtools and BCFtools. GigaScience 10: giab008.
•  deGraaf,  A.A.V.,  deBruijn,  P.,  Robertson,  L.A.,  Jetten,  M.S.M., and  Kuenen,  J.G. (1996) Autotrophic growth of anaerobic ammonium-oxidizing micro-organisms in a fluidized bed reactor. Microbiology (Reading) 142: 2187–2196.
•  Domingo-Felez,  C., and  Smets,  B.F. (2019) Regulation of key N(2)O production mechanisms during biological water treatment. Curr Opin Biotechnol 57: 119–126.
•  Drummond,  J.T., and  Matthews,  R.G. (1994) Nitrous oxide degradation by cobalamin-dependent methionine synthase: characterization of the reactants and products in the inactivation reaction. Biochemistry 33: 3732–3741.
•  Duan,  H.,  Zhao,  Y.,  Koch,  K.,  Wells,  G.F.,  Zheng,  M.,  Yuan,  Z., and  Ye,  L. (2021) Insights into nitrous oxide mitigation strategies in wastewater treatment and challenges for wider implementation. Environ Sci Technol 55: 7208–7224.
•  Fenu,  A.,  Smolders,  S.,  De Gussem,  K., and  Weemaes,  M. (2019) Conflicting carbon footprint and energy saving in a side-stream Anammox Process. Biochem Eng J 151: 107336.
•  Frutos,  O.D.,  Quijano,  G.,  Pérez,  R., and  Muñoz,  R. (2016) Simultaneous biological nitrous oxide abatement and wastewater treatment in a denitrifying off-gas bioscrubber. Chem Eng J 288: 28–37.
•  Gao,  Y.,  Mania,  D.,  Mousavi,  S.A.,  Lycus,  P.,  Arntzen,  M.O.,  Woliy,  K., et al. (2021) Competition for electrons favours N2 O reduction in denitrifying Bradyrhizobium isolates. Environ Microbiol 23: 2244–2259.
•  Gilbert,  E.M.,  Agrawal,  S.,  Schwartz,  T.,  Horn,  H., and  Lackner,  S. (2015) Comparing different reactor configurations for Partial Nitritation/Anammox at low temperatures. Water Res 81: 92–100.
•  Graham,  E.D.,  Heidelberg,  J.F., and  Tully,  B.J. (2018) Potential for primary productivity in a globally-distributed bacterial phototroph. ISME J 12: 1861–1866.
•  Han,  H.,  Kim,  D.D.,  Song,  M.J.,  Yun,  T.,  Yoon,  H.,  Lee,  H.W., et al. (2023) Biotrickling filtration for the reduction of N(2)O emitted during wastewater treatment: results from a long-term in situ pilot-scale testing. Environ Sci Technol 57: 3883–3892.
•  Hou,  X.,  Liu,  S., and  Zhang,  Z. (2015) Role of extracellular polymeric substance in determining the high aggregation ability of anammox sludge. Water Res 75: 51–62.
•  Hyatt,  D.,  Chen,  G.L.,  Locascio,  P.F.,  Land,  M.L.,  Larimer,  F.W., and  Hauser,  L.J. (2010) Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinf 11: 119.
•  Isobe,  K.,  Suwa,  Y.,  Ikutani,  J.,  Kuroiwa,  M.,  Makita,  T.,  Takebayashi,  Y., et al. (2011) Analytical techniques for quantifying (15)N/(14)N of nitrate, nitrite, total dissolved nitrogen and ammonium in environmental samples using a gas chromatograph equipped with a quadrupole mass spectrometer. Microbes Environ 26: 46–53.
•  Jones,  P.,  Binns,  D.,  Chang,  H.Y.,  Fraser,  M.,  Li,  W.,  McAnulla,  C., et al. (2014) InterProScan 5: genome-scale protein function classification. Bioinformatics 30: 1236–1240.
•  Kanehisa,  M., and  Goto,  S. (2000) KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res 28: 27–30.
•  Kang,  D.D.,  Li,  F.,  Kirton,  E.,  Thomas,  A.,  Egan,  R.,  An,  H., and  Wang,  Z. (2019) MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ 7: e7359.
•  Keren,  R.,  Lawrence,  J.E.,  Zhuang,  W.,  Jenkins,  D.,  Banfield,  J.F.,  Alvarez-Cohen,  L., et al. (2020) Increased replication of dissimilatory nitrate-reducing bacteria leads to decreased anammox bioreactor performance. Microbiome 8: 7.
•  Kindaichi,  T.,  Yuri,  S.,  Ozaki,  N., and  Ohashi,  A. (2012) Ecophysiological role and function of uncultured Chloroflexi in an anammox reactor. Water Sci Technol 66: 2556–2561.
•  Kinh,  C.T.,  Suenaga,  T.,  Hori,  T.,  Riya,  S.,  Hosomi,  M.,  Smets,  B.F., and  Terada,  A. (2017) Counter-diffusion biofilms have lower N(2)O emissions than co-diffusion biofilms during simultaneous nitrification and denitrification: Insights from depth-profile ana­lysis. Water Res 124: 363–371.
•  Lackner,  S.,  Gilbert,  E.M.,  Vlaeminck,  S.E.,  Joss,  A.,  Horn,  H., and  van Loosdrecht,  M.C. (2014) Full-scale partial nitritation/anammox experiences--an application survey. Water Res 55: 292–303.
•  Lawson,  C.E.,  Wu,  S.,  Bhattacharjee,  A.S.,  Hamilton,  J.J.,  McMahon,  K.D.,  Goel,  R., and  Noguera,  D.R. (2017) Metabolic network ana­lysis reveals microbial community interactions in anammox granules. Nat Commun 8: 12.
•  Lawson,  C.E.,  Nuijten,  G.H.L.,  de Graaf,  R.M.,  Jacobson,  T.B.,  Pabst,  M.,  Stevenson,  D.M., et al. (2021) Autotrophic and mixotrophic metabolism of an anammox bacterium revealed by in vivo (13)C and (2)H metabolic network mapping. ISME J 15: 673–687.
•  Li,  D.,  Liu,  C.M.,  Luo,  R.,  Sadakane,  K., and  Lam,  T.W. (2015) MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics 31: 1674–1676.
•  Li,  J.,  Li,  J.,  Gao,  R.,  Wang,  M.,  Yang,  L.,  Wang,  X., et al. (2018) A critical review of one-stage anammox processes for treating industrial wastewater: Optimization strategies based on key functional microorganisms. Bioresour Technol 265: 498–505.
•  Li,  W., and  Godzik,  A. (2006) Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22: 1658–1659.
•  Liao,  Y.,  Smyth,  G.K., and  Shi,  W. (2013) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30: 923–930.
•  Lu,  X.,  Heal,  K.R.,  Ingalls,  A.E.,  Doxey,  A.C., and  Neufeld,  J.D. (2020) Metagenomic and chemical characterization of soil cobalamin production. ISME J 14: 53–66.
•  Magoc,  T., and  Salzberg,  S.L. (2011) FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27: 2957–2963.
•  McCarty,  P.L. (2018) What is the best biological process for nitrogen removal: When and why? Environ Sci Technol 52: 3835–3841.
•  McMurdie,  P.J., and  Holmes,  S. (2013) phyloseq: an R package for reproducible interactive ana­lysis and graphics of microbiome census data. PLoS One 8: e61217.
•  Nerenberg,  R. (2016) The membrane-biofilm reactor (MBfR) as a counter-diffusional biofilm process. Curr Opin Biotechnol 38: 131–136.
•  Ni,  B.J.,  Ruscalleda,  M., and  Smets,  B.F. (2012) Evaluation on the microbial interactions of anaerobic ammonium oxidizers and heterotrophs in Anammox biofilm. Water Res 46: 4645–4652.
•  Nurk,  S.,  Meleshko,  D.,  Korobeynikov,  A., and  Pevzner,  P.A. (2017) metaSPAdes: a new versatile metagenomic assembler. Genome Res 27: 824–834.
•  Oba,  K.,  Suenaga,  T.,  Kuroiwa,  M.,  Riya,  S., and  Terada,  A. (2022) Exploring the functions of efficient canonical denitrifying bacteria as N₂O sinks: Implications from (15)N tracer and transcriptome ana­lyses. Environ Sci Technol 56: 11694–11706.
•  Okabe,  S.,  Oshiki,  M.,  Takahashi,  Y., and  Satoh,  H. (2011) N₂O emission from a partial nitrification-anammox process and identification of a key biological process of N₂O emission from anammox granules. Water Res 45: 6461–6470.
•  Olm,  M.R.,  Brown,  C.T.,  Brooks,  B., and  Banfield,  J.F. (2017) dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J 11: 2864–2868.
•  Oshiki,  M.,  Takaki,  Y.,  Hirai,  M.,  Nunoura,  T.,  Kamigaito,  A., and  Okabe,  S. (2022a) Metagenomic ana­lysis of five phylogenetically distant anammox bacterial enrichment cultures. Microbes Environ 37: ME22017.
•  Oshiki,  M.,  Toyama,  Y.,  Suenaga,  T.,  Terada,  A.,  Kasahara,  Y.,  Yamaguchi,  T., and  Araki,  N. (2022b) N(2)O reduction by gemmatimonas aurantiaca and potential involvement of gemmatimonadetes bacteria in N(2)O reduction in agricultural soils. Microbes Environ 37: ME21090.
•  Park,  D.,  Kim,  H., and  Yoon,  S. (2017) Nitrous oxide reduction by an obligate aerobic bacterium, Gemmatimonas aurantiaca strain T-27. Appl Environ Microbiol 83: e00502-17.
•  Parks,  D.H.,  Chuvochina,  M.,  Rinke,  C.,  Mussig,  A.J.,  Chaumeil,  P.A., and  Hugenholtz,  P. (2022) GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy. Nucleic Acids Res 50: D785–D794.
•  Peng,  L.,  Carvajal-Arroyo,  J.M.,  Seuntjens,  D.,  Prat,  D.,  Colica,  G.,  Pintucci,  C., and  Vlaeminck,  S.E. (2017) Smart operation of nitritation/denitritation virtually abolishes nitrous oxide emission during treatment of co-digested pig slurry centrate. Water Res 127: 1–10.
•  Prodan,  A.,  Tremaroli,  V.,  Brolin,  H.,  Zwinderman,  A.H.,  Nieuwdorp,  M., and  Levin,  E. (2020) Comparing bioinformatic pipelines for microbial 16S rRNA amplicon sequencing. PLoS One 15: e0227434.
•  Quast,  C.,  Pruesse,  E.,  Yilmaz,  P.,  Gerken,  J.,  Schweer,  T.,  Yarza,  P., et al. (2013) The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41: D590–596.
•  Riya,  S.,  Zhou,  S.,  Watanabe,  Y.,  Sagehashi,  M.,  Terada,  A., and  Hosomi,  M. (2012) CH4 and N₂O emissions from different varieties of forage rice (Oryza sativa L.) treating liquid cattle waste. Sci Total Environ 419: 178–186.
•  Robeson,  M.S., 2nd,  O’Rourke,  D.R.,  Kaehler,  B.D.,  Ziemski,  M.,  Dillon,  M.R.,  Foster,  J.T., and  Bokulich,  N.A. (2021) RESCRIPt: Reproducible sequence taxonomy reference database management. PLoS Comput Biol 17: e1009581.
•  Sanford,  R.A.,  Wagner,  D.D.,  Wu,  Q.,  Chee-Sanford,  J.C.,  Thomas,  S.H.,  Cruz-García,  C., et al. (2012) Unexpected nondenitrifier nitrous oxide reductase gene diversity and abundance in soils. Proc Natl Acad Sci U S A 109: 19709–19714.
•  Schipp,  C.J.,  Marco-Urrea,  E.,  Kublik,  A.,  Seifert,  J., and  Adrian,  L. (2013) Organic cofactors in the metabolism of Dehalococcoides mccartyi strains. Philos Trans R Soc Lond B Biol Sci 368: 20120321.
• Seemann, T. (2018) barrnap 0.9: rapid ribosomal RNA prediction. URL: https://github.com/tseemann/barrnap
•  Shan,  J.,  Sanford,  R.A.,  Chee-Sanford,  J.,  Ooi,  S.K.,  Loffler,  F.E.,  Konstantinidis,  K.T., and  Yang,  W.H. (2021) Beyond denitrification: The role of microbial diversity in controlling nitrous oxide reduction and soil nitrous oxide emissions. Glob Chang Biol 27: 2669–2683.
•  Shelton,  A.N.,  Seth,  E.C.,  Mok,  K.C.,  Han,  A.W.,  Jackson,  S.N.,  Haft,  D.R., and  Taga,  M.E. (2019) Uneven distribution of cobamide biosynthesis and dependence in bacteria predicted by comparative genomics. ISME J 13: 789–804.
•  Shen,  W.,  Le,  S.,  Li,  Y., and  Hu,  F. (2016) SeqKit: A cross-platform and ultrafast toolkit for FASTA/Q file manipulation. PLoS One 11: e0163962.
•  Sieber,  C.M.K.,  Probst,  A.J.,  Sharrar,  A.,  Thomas,  B.C.,  Hess,  M.,  Tringe,  S.G., and  Banfield,  J.F. (2018) Recovery of genomes from metagenomes via a dereplication, aggregation and scoring strategy. Nat Microbiol 3: 836–843.
• Smith, C., Nicholls, Z.R.J., Armour, K., Collins, W., Forster, P., Meinshausen, M., et al. (2021) The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity Supplementary Material. In Climate Change 2021. The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Masson-Delmotte, V., Zhai, P., Pirani, A., Connors, S.L., Péan, C., Berger, S., et al., (eds). URL: https://www.ipcc.ch/
•  Speth,  D.R.,  In’t Zandt,  M.H.,  Guerrero-Cruz,  S.,  Dutilh,  B.E., and  Jetten,  M.S. (2016) Genome-based microbial ecology of anammox granules in a full-scale wastewater treatment system. Nat Commun 7: 11172.
•  Strous,  M.,  Heijnen,  J.J.,  Kuenen,  J.G., and  Jetten,  M.S.M. (1998) The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Appl Microbiol Biotechnol 50: 589–596.
•  Suenaga,  T.,  Hori,  T.,  Riya,  S.,  Hosomi,  M.,  Smets,  B.F., and  Terada,  A. (2019) Enrichment, isolation, and characterization of high-affinity N₂O-reducing bacteria in a gas-permeable membrane reactor. Environ Sci Technol 53: 12101–12112.
•  Suenaga,  T.,  Ota,  T.,  Oba,  K.,  Usui,  K.,  Sako,  T.,  Hori,  T., et al. (2021) Combination of (15)N tracer and microbial ana­lyses discloses N₂O sink potential of the anammox community. Environ Sci Technol 55: 9231–9242.
•  Sullivan,  M.J.,  Gates,  A.J.,  Appia-Ayme,  C.,  Rowley,  G., and  Richardson,  D.J. (2013) Copper control of bacterial nitrous oxide emission and its impact on vitamin B12-dependent metabolism. Proc Natl Acad Sci U S A 110: 19926–19931.
•  Tanizawa,  Y.,  Fujisawa,  T., and  Nakamura,  Y. (2018) DFAST: a flexible prokaryotic genome annotation pipeline for faster genome publication. Bioinformatics 34: 1037–1039.
•  Terada,  A.,  Sugawara,  S.,  Yamamoto,  T.,  Zhou,  S.,  Koba,  K., and  Hosomi,  M. (2013) Physiological characteristics of predominant ammonia-oxidizing bacteria enriched from bioreactors with different influent supply regimes. Biochem Eng J 79: 153–161.
• Vasimuddin M, Misra S, Li H, Aluru S, editors. Efficient Architecture-Aware Acceleration of BWA-MEM for Multicore Systems. 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS); 2019 20-24 May 2019.
•  Wong,  L.L.,  Lu,  Y.,  Ho,  J.C.S.,  Mugunthan,  S.,  Law,  Y.,  Conway,  P., et al. (2023) Surface-layer protein is a public-good matrix exopolymer for microbial community organisation in environmental anammox biofilms. ISME J 17: 803–812.
•  Wu,  Y.W.,  Simmons,  B.A., and  Singer,  S.W. (2016) MaxBin 2.0: an automated binning algorithm to recover genomes from multiple metagenomic datasets. Bioinformatics 32: 605–607.
•  Xiao,  R.,  Ni,  B.J.,  Liu,  S., and  Lu,  H. (2021) Impacts of organics on the microbial ecology of wastewater anammox processes: Recent advances and meta-ana­lysis. Water Res 191: 116817.
•  Yasuda,  S.,  Suenaga,  T., and  Terada,  A. (2020) Complete genome sequence of methylosinus sp. strain C49, a methane-oxidizing bacterium harboring phaABC genes for polyhydroxyalkanoate synthesis. Microbiol Resour Announc 9: e00113-20.
•  Yin,  Q.,  Sun,  Y.,  Li,  B.,  Feng,  Z., and  Wu,  G. (2022) The r/K selection theory and its application in biological wastewater treatment processes. Sci Total Environ 824: 153836.
•  Yin,  Y.,  Yan,  J.,  Chen,  G.,  Murdoch,  F.K.,  Pfisterer,  N., and  Loffler,  F.E. (2019) Nitrous oxide is a potent inhibitor of bacterial reductive dechlorination. Environ Sci Technol 53: 692–701.
•  Yoon,  H.,  Song,  M.J., and  Yoon,  S. (2017) Design and feasibility ana­lysis of a self-sustaining biofiltration system for removal of low concentration N(2)O emitted from wastewater treatment plants. Environ Sci Technol 51: 10736–10745.
•  Yoon,  S.,  Nissen,  S.,  Park,  D.,  Sanford,  R.A., and  Loffler,  F.E. (2016) Nitrous oxide reduction kinetics distinguish bacteria harboring clade I NosZ from those harboring clade II NosZ. Appl Environ Microbiol 82: 3793–3800.