Characterisation of bacteria representing a novel Nitrosomonas clade: Physiology, genomics and distribution of missing ammonia oxidizer

Abstract Members of the genus Nitrosomonas are major ammonia oxidizers that catalyse the first step of nitrification in various ecosystems. To date, six subgenus‐level clades have been identified. We have previously isolated novel ammonia oxidizers from an additional clade (unclassified cluster 1) of the genus Nitrosomonas. In this study, we report unique physiological and genomic properties of the strain PY1, compared with representative ammonia‐oxidising bacteria (AOB). The apparent half‐saturation constant for total ammonia nitrogen and maximum velocity of strain PY1 were 57.9 ± 4.8 μM NH3 + NH4 + and 18.5 ± 1.8 μmol N (mg protein)−1 h−1, respectively. Phylogenetic analysis based on genomic information revealed that strain PY1 belongs to a novel clade of the Nitrosomonas genus. Although PY1 contained genes to withstand oxidative stress, cell growth of PY1 required catalase to scavenge hydrogen peroxide. Environmental distribution analysis revealed that the novel clade containing PY1‐like sequences is predominant in oligotrophic freshwater. Taken together, the strain PY1 had a longer generation time, higher yield and required reactive oxygen species (ROS) scavengers to oxidize ammonia, compared with known AOB. These findings expand our knowledge of the ecophysiology and genomic diversity of ammonia‐oxidising Nitrosomonas.

Data concerning the physiology, genome, phylogeny and distribution of AOB have accumulated steadily over the last 130 years.Understanding AOB has considerably depended on the study of axenic cultures (Stein, 2019).However, AOB gain ATP via the uptake of ammonia as the sole energy source.The consequent production of nitrite and decreased pH cause a decrease in AOB activity (Claros et al., 2013).Once AOB are isolated as pure cultures, they are exposed to elevated nitrite concentrations and acidification.This inevitable condition often impedes long-term preservation and recovery and can result in the loss of AOB isolates (Bollmann et al., 2011;Fujitani et al., 2015).AOB, which are fastidious and recalcitrant microorganisms in the laboratory setting and have not been cultured and isolated, have attracted much research attention.Recent culture-based studies have reported a new phylogeny of AOB adapted to acidic conditions (Hayatsu et al., 2017;Picone et al., 2021), indicating that the environmental diversity of AOB remains unclear.
We previously reported the isolation of novel AOB retrieved from activated sludge, which were affiliated with an additional clade (unclassified cluster 1) of the Nitrosomonas genus at the 16S rRNA and amoA gene levels (Abe et al., 2017).The isolates were obtained by detecting clonal microcolonies in activated sludge using fluorescence in situ hybridization and sorting microcolonies based on scattering signatures.However, the ecophysiology and genomics of these isolates remain unclear.In this study, we report the physiology and genome of one of the isolates, which was categorised as a previously unrecognised clade of the Nitrosomonas genus.The abundance and environmental distribution of the new Nitrosomonas clade was investigated using an integrated microbial next-generation sequencing platform (IMNGS).

Effect of catalase on the growth of strain PY1
After isolating and identifying three AOB (strains SN1, NP1 and PY1) retrieved from activated sludge (Abe et al., 2017), we established a protocol to obtain sufficient biomass for physiological experiments and genome sequencing.The three strains were cultivated and transferred to an inorganic medium containing ammonium for 2 years.Two strains, SN1 and NP1, were lost and the ammonia oxidation activity of PY1 remained low.Even if 0.56 mM NH 4 Cl was added to the culture, only 0.1 mM nitrite was produced (Figure S1).To overcome this drawback, we added various substances, including carbohydrates, organic acids, amino acids, complex nutrients and urea, to inorganic media to enhance the ammonia-oxidising activity of PY1; however, these attempts failed (Table S1).In a previous study, adding a hydrogen peroxide (H 2 O 2 ) scavenger stimulated the ammonia oxidation of Nitrosopumilus sp.DDS1 (Kim et al., 2016).Based on this knowledge, we cultivated PY1 in ammonium-based medium supplemented with catalase at 50 or 400 U mL À1 .Consequently, PY1 showed improved ammonia oxidation activity and produced more nitrite (Figure S1 and Table S1).Unless otherwise stated, inorganic media containing catalase were utilised for strain PY1 in subsequent experiments.

Physiological characteristics
Physiological experiments were conducted to determine the optimum temperature (Figure 1A), ammonium concentration (Figure 1B), growth rate (Figure 1C) and kinetics (Figure 1D).Strain PY1 was cultivated at temperatures ranging from 4 C to 46 C, and nitrite was produced only at temperatures ranging from 23 C to 28 C with a total consumption of 0.71 mM ammonium for 2-3 weeks (Figure 1A).Compared to other ammonia oxidizers, including Nitrosomonas species, strain PY1 required a relatively narrow temperature range (23 C-28 C) to oxidize ammonia.However, the optimal temperature for the PY1 strain was typical of ammonia oxidizers (Table 1).
To investigate tolerance to ammonium concentrations, PY1 was cultivated with different initial ammonium concentrations.During the 32 days of incubation, 0.71 mM and 1.4 mM ammonium were completely consumed, and equivalent nitrite was produced (Figure 1B).However, an initial ammonium concentration exceeding 2.8 mM delayed nitrite production, resulting in prolonged lag time.An initial ammonium concentration of 5.7 or 7.1 mM completely prevented nitrite production by the strain PY1.The maximum ammonium tolerance of PY1 was considerably lower than that reported for other AOB (Table 1).
The growth of PY1 during ammonia oxidation was evaluated using qPCR with 16S rRNA gene primers.After 14 days of incubation, cell density increased with the stoichiometric oxidation of ammonia to nitrite (Figure 1C).The specific growth rate calculated by fitting the cell growth curve to an exponential equation, was 0.23 day À1 (Figure S2).The maximum specific growth rate was 0.49 ± 0.20 day À1 .The generation time was 3.0 days, which was longer than that of most ammonia oxidizers, except for Nitrosospira sp.AF isolated from acidic soil (Table 1).The estimated growth yield of PY1 was estimated to be 33.5 cells pmol À1 .
The apparent half-saturation constant (K m (app) ) for total ammonia nitrogen (NH 3 and NH 4 + ) and maximum velocity (V max ) were calculated by manually fitting the data obtained to the Michaelis-Menten equation (Figure 1D, Figure S3).Based on experiments performed in three biological replicates, the K m (app) and V max of strain PY1 were 57.9 ± 4.8 μM NH 3 + NH 4 + (3.3 μM NH 3 ) and 18.5 ± 1.8 μmol N (mg protein) À1 h À1 , respectively.The K m (app) of strain PY1 was within the range of Nitrosomonas cluster 6a AOB, whereas the V max of strain PY1 was lower than that of other AOB (Table 1).Considering these physiological properties, PY1 appears to be a typical AOB adapted to oligotrophic environments.

Morphology
Scanning electron microscopy (SEM) revealed that PY1 cells were rod-shaped, similar to previously isolated strains (Koops et al., 1991).As strain PY1 was isolated by sorting pure microcolonies from activated sludge samples (Abe et al., 2017), this strain may form cell aggregates (Figure 2A).The width and length of the rod-shaped cells ranged from 0.9 to 1.6 μm and 0.5 to 0.8 μm, respectively (Figure 2B).Ultrathin sections of the cells were observed using transmission electron microscopy (TEM).As observed in other strains of the genus Nitrosomonas (Koops et al., 2006), PY1 had multilayer intracytoplasmic membranes (Figure 2C).The maximum ammonium tolerance concentration was considered to be the minimum initial ammonium concentration in the absence of nitrite production or bacterial growth.§ This value was the minimum generation time that was calculated by maximum specific growth rate.

Genome overview
The high-quality draft genome of PY1, comprising four scaffolds, was reconstructed using paired-end sequences and a mate-pair sequence.The four scaffold sizes were 2,710,791 bp (scaffold 1), 58,489 bp (scaffold 2), 58,786 bp (scaffold 3) and 26,708 bp (scaffold 4).The total size of the genome was 2,854,774 bp.The G + C content was 42.7%.The genome contained 2612 predicted protein-coding DNA sequences (CDS), a coding ratio of 86.6%, two rRNA genes (16S and 23S), 37 tRNA genes and one CRISPR gene (Table S2).The range of the average nucleotide identity (ANI) between strain PY1 and other AOB isolates was 76.7%-82.9%(Table 2).Considering that the species cut-off point for ANI is 95% (Konstantinidis & Tiedje, 2005), strain PY1 was considered a novel AOB, at least at the species level.The range of the average amino acid identity (AAI) between strain PY1 and other AOB isolates was 60.2%-70.8%(Table 2).In addition, a genome tree revealed that PY1 is phylogenetically separated from the well-known AOB cluster (Figure 3).Previously, we reported that the identity between strain PY1 and its closest species was 95% and 88% for 16S rRNA and amoA genes, respectively, and categorised the phylogeny of strain PY1 into a novel clade within the genus Nitrosomonas (Abe et al., 2017, Figure S4).This assignment was validated using the results of the genome tree constructed in this study.

Nitrogen metabolism
The PY1 genome contained three copies of the gene cluster encoding ammonia monooxygenase (amoCAB), followed by two copies of the amoED operon.The genes likely encoding the copper transport and residence proteins copC and copD were detected downstream of one of these amoCABED operons.The genome of strain PY1 contains two copies of the singleton amoC followed by one copy of amoE.PY1 potentially encodes three complete operons for hydroxylamine dehydrogenase (haoAB-cysAB), copper-containing nitrite reductase (nirK), nitric oxide reductase (norCBQD) and nitrosocyanin (ncyA) (Figure 4, Table S3).Recently, nirK and ncyA have been suggested as candidates for nitric oxide oxidoreductase (NOO), a newly proposed enzyme in the pathway for energy conservation by AOB (Caranto & Lancaster, 2017;Zorz et al., 2018).The copy numbers of the amoCAB, singleton amoC and haoAB-cysAB operons were almost the same as those of cluster 6a AOB (Nitrosomonas sp.Is79A3 and Nitrosomonas sp.AL212, N. ureae) (Table 2).
Although the genomes of some AOB contain an ammonia transporter, none were detected in the PY1 genome, which is similar to the N. eutropha C91 genome (Stein et al., 2007) (Table 2).Strain PY1 and other AOB (Hayatsu et al., 2017;Offre et al., 2014) may incorporate ammonia via passive diffusion.These properties could be disadvantageous for PY1, hindering its adaptation to environments with variable ammonium/ammonia concentrations.Furthermore, strain PY1 was more sensitive to higher ammonium concentrations than the other AOB (Table 1).The lack of control of ammonium concentration within the cytoplasm might be responsible for the difficulty in cultivating PY1 in a laboratory setting.Additionally, PY1 has a potential formate/nitrite transporter (focA), which transfers nitrite to the cytoplasm and ferredoxin-nitrite reductase (nirA), which reduces nitrite to ammonia (Figure 4).Strain PY1 contains a urea transporter that transfers urea to the cytoplasm, urease (ureCBA) and urease accessory protein (ureDEFG) (Koper et al., 2004) (Figure 4, Table S5).However, the addition of urea to the inorganic medium did not affect the growth of PY1 in the absence of catalase (Table S1).To determine whether strain PY1 can use urea as an energy source, we conducted an experiment on urea utilisation by PY1.In the presence of catalase, we found that urea was hydrolysed to ammonium, which was converted into nitrite.In the absence of catalase, urea was hydrolysed to ammonium, which accumulated without nitrite production (Figure S5).Therefore, PY1 can only use urea as an energy source in the presence of reactive oxygen species (ROS) scavengers.In many AOB, the two-electron reduction in two molecules of NO to N 2 O is performed by two classes of cytochrome c NOR (cNOR), encoded by the norCBQD operon, and the alternative NO reductase (sNOR) encoded by the norSY-SenC operon (Kits et al., 2019).The PY1 genome contained only one copy of the gene for cNOR but no sNOR gene, similar to Nitrosomonas sp.AL212 and N. cryotolerans (Rice et al., 2017;Suwa et al., 2011).AOB species within clusters 7 and 8, which are adapted to eutrophic environments, tended to have two types of NOR.In contrast, AOB species within cluster 6a and the PY1 cluster, which are adapted to oligotrophic environments, had only cNOR or neither NOR.Therefore, the NOR setup might be specific to each AOB cluster.Considering that ammonia concentration is considerably low in oligotrophic environments where does not accumulate n-oxides such as nitric oxide and hydroxylamine, NOR is not always required for oligotrophic AOB (Kozlowski et al., 2016c).However, nitric oxide and hydroxylamine frequently accumulate in wastewater treatment plants (WWTPs) and in laboratory cultures of activated sludge (Kampschreur et al., 2008;Wang et al., 2016;Yang & Alleman, 1992).Considering that PY1 and AL212 were isolated from activated sludge, it might be essential to have multiple mechanisms for oxidising the intermediate nitrogen metabolites of these species.In addition, PY1 harbours two copies of cytochrome P460 gene (cytL).CytL has been shown to oxidise nitric oxide and hydroxylamine to nitrite (Stein, 2011) and/or to oxidise hydroxylamine to nitrous oxide (Caranto et al., 2016).More recently, Nitrosomonas.sp AL212 CytL was biochemically characterised (Smith et al., 2019;Smith & Lancaster, 2018).Considering that the expression of CytL was confirmed during aerobic growth in some AOB, but not all tested AOB (Zorz et al., 2018), it remains unknown whether CytL in strain PY1 is functional.

Carbon fixation
AOB potentially possess two types of ribulose-1,-5-bisphosphate carboxylase/oxygenase (RuBisCO), type IA and type IC.The affinity for carbon dioxide is known to be higher for type IA than type IC (Badger & Bek, 2008).We observed that strain PY1 formed contained type IA RuBisCO, three copies of carbonic anhydrase and no carboxysome genes (Figure 4, Table S3).AOB affiliated with cluster 6a have one copy of each type of RuBisCO and thus a more flexible ability to fix carbon dioxide (Bollmann et al., 2013).In contrast, strain PY1 has only type IA RuBisCO.Similar to strain PY1, cluster 7 AOB possesses only the type IA RuBisCO, whereas some but not all cluster 7 AOB encode carboxysome genes (Sedlacek et al., 2019).Therefore, there seems to be no correlation between RuBisCO and carboxysome genes.Strain PY1 might  Kozlowski et al., 2016b, Kozlowski et al., 2016cThandar et al., 2016Stein et al., 2007Chain et al., 2003Norton et al., 2008 Abbreviations: +, positive; À, negative; N/A, not applicable.† Average nucleotide identity (ANI) and average amino acids identity (AAI) of PY1 strain with other AOB pure strains.require a more stringent carbon dioxide concentration than AOB in cluster 6a (Table 2).

Oxidative stress
Strain PY1, as well as other AOB, have superoxide dismutase, catalase, rubrerythrin (rdx) and alkyl hydroperoxide reductase (alpC) genes (Figure 4, Table S3) but lack the gene encoding transaldolase (tal), which likely detoxifies ROS and functions in a non-oxidative reaction of the pentose phosphate pathway (Table S3).Ammonia oxidation by Nitrosopumilus spp.DDS1, whose genome lacks sequences encoding catalase, was stimulated by adding α-ketoglutarate to remove H 2 O 2 from the inorganic medium (Kim et al., 2016).Comparative proteomics has revealed that AOB in the exponential growth phase express more rdx and alpC than the catalase gene to defend against oxidative stress (Zorz et al., 2018).Even if strain PY1 has multiple genes to defend against oxidative stress, we added catalase to the inorganic medium to enhance its growth.Therefore, strain PY1 is the first AOB that requires a ROS scavenger to enhance its ammonia- The genome tree based on the amino acid sequences of 120 single-copy marker genes shared between representative Nitrosomonas and Nitrosospira genomes using the randomised accelerated maximum likelihood (RAxML) software (the JTT and the Gamma models with rapid 100-times bootstrapping).Numbers at branch nodes indicate all bootstrap values (%).oxidising activity under normal growth conditions.Considering that the expression of some genes that defend against oxidative stress depends on the growth phase and/or culture conditions (Zorz et al., 2018), the catalase gene and other ROS genes of strain PY1 might be inactive under normal growth conditions.Previous physiological and transcriptomic analyses of co-culture experiments using AOB and NOB suggested that Nitrobacter winogradkyi helps reduce oxidative stress and improves the growth of Nitrosomonas sp.Is79 (Sedlacek et al., 2016).Similarly, physiological and proteomic approaches have demonstrated that co-culture with marine AOA and heterotrophs decreases hydroperoxide in the medium and supports AOA cell growth (Bayer et al., 2019).Therefore, co-culture of strain PY1 with other bacteria, such as NOB and/or heterotrophs, might effectively stimulate ammonia oxidation in strain PY1.
Environmental abundance and distribution of Nitrosomonas sp.PY1-like sequences Nitrosomonas sp.PY1-like sequences were present in 934 different publicly available sequence read archive (SRA) samples deposited in GenBank (from 259,144 total SRA runs analysed with IMNGS) (Figure 5).The number of SRA samples, including representative AOB, is shown in Table S4.The ratios of the amplicon dataset, including Nitrosomonas sp.PY1-like sequences, in the total amplicon data set were 30.0% in activated sludge and 7.28% in freshwater samples.Although the amplicon data set containing the Nitrosomonas sp.PY1-like sequences were also detected in soil and marine samples, the percentage was considerably lower.The relative abundances of Nitrosomonas sp.PY1-like sequences to total 16S rRNA sequences in each amplicon data set were 0.0003%-2%.In the four examples where the PY1 cluster comprised more than 2% of the total community, three freshwater samples (9.2%, 6.6% and 5.9%) contained data on chlorine-free drinking water (Run accession numbers: ERR653163, ERR653166 and ERR653157, respectively).The other sample (2.1%) contained data on the soil in an olive orchard (ERR527336).Interestingly, high abundances of the PY1 cluster were detected in the three freshwater samples.We analysed the distribution of the other AOB clusters using IMNGS (Figure S6).In the activated sludge, the relative abundance of cluster 7 AOB sequences was 1%-10% whereas those of cluster PY1 and cluster 6a AOB sequences were less than 1%.In freshwater samples, the relative abundance of cluster 7 AOB sequences was less than 0.1%, whereas some relative abundances of PY1 cluster and cluster 6a AOB sequences were greater than 0.1%.Based on the comparative analysis using IMNGS, the PY1 cluster might be the predominant AOB in oligotrophic freshwater, although a distinct niche of differentiation between the PY1 cluster and cluster 6a AOB was not found.
Challenges in isolating previously uncultured AOB and insights into their growth mechanism AOB often have long generation times and are sensitive to environmental factors (Stein, 2019).In our previous study, we isolated three Nitrosomonas strains by segregating clonal microcolonies or Nitrosomonas cells from microbial consortia in activated sludge (Abe et al., 2017).However, it was difficult to maintain pure cultures of these cells.In this study, the addition of catalase, rather than any other organic compound, was a more effective way to increase the PY1 biomass.Considering that pyruvate and α-ketoglutarate did not affect the ammonia-oxidising activity of PY1 (Table S1), the supply of some organic compounds could have no specific effect on AOB strains.Previous studies have revealed that the removal of hydroperoxide by the addition of catalase (Kim et al., 2016) or in the presence or absence of associated heterotrophic contaminants (Bayer et al., 2019) enhances the ammonia oxidation activity of AOA.Based on these findings, removing hydroperoxide could effectively improve the isolation of nitrifiers and the production of sufficient biomass in pure culture.However, the effect of oxidative stress on nitrifier cells remains unclear.The molecular mechanisms underlying resistance to oxidative stress can be elucidated using genetic recombination techniques.Except for some nitrifiers (Bock et al., 1983;Klein et al., 2022;Koops et al., 1991;Sorokin et al., 2012), obtaining pure colonies of nitrifiers and/or recombinant nitrifier strains on solid media remains challenging.Medium composition and cultivation protocols that reduce oxidative stress might contribute to the isolation of novel nitrifiers and provide insights into growth mechanisms.

CONCLUSION
We report the isolation, morphology, physiology, kinetics and genome of strain PY1, which belongs to the unclassified cluster 1 of Nitrosomonas and was isolated from activated sludge present in a wastewater treatment plant.Ammonia oxidation by PY1 was enhanced by the addition of catalase to an inorganic medium containing ammonia.Compared with the known Nitrosomonas strains, strain PY1 had a longer generation time, higher yield and required more ROS scavengers.Although the morphology and genome of strain PY1 were similar to those of the Nitrosomonas genus, phylogenetic analysis, including genomic information, revealed that this strain belongs to a novel clade of the Nitrosomonas genus.These findings broaden our understanding of ammonia-oxidising Nitrosomonas, an ecologically important group in the biogeochemical nitrogen cycle and artificial nitrogen removal processes.

F
I G U R E 1 Physiological characteristics of strain PY1.(A) Temperature dependence of nitrite production.The black, orange and purple circles completely overlapped because nitrite production was not observed at 4 C, 37 C and 46 C. (B) Influence of initial ammonium concentration on ammonia oxidation activity.The orange and purple circles completely overlapped because nitrite production was not observed at initial NH 4 Cl concentrations of 5.7 mM and 7.1 mM.(C) Changes in ammonium and nitrite concentrations as well as cell density estimated by the 16S rRNA gene copies of Betaproteobacterial AOB using the qPCR analysis.White bars on days 4 and 6 were used to calculate the maximum specific growth rate.(D) Representative data showing the ammonia uptake activity.The curve was fitted to the Michaelis-Menten equation: Error bars depicting the standard deviations of triplicate biological measurements are shown; they are not shown if the bars are smaller than the symbols.T A B L E 1 Physiologic characteristics of Nitrosomonas sp.PY1 and other selected members of AOB.When the generation time was not described, they were calculated using the following fomula: generation time = Ln(2)/μ, μ: specific growth rate.When the unit of generation time were hours (h), it was divided by 24 and converted to days (d).Superscripted alphabets show values determined in previous studies.Abbreviation: ND, not determined.*Thegeneration time were applied from the values estimated from the exponential growth curve.† Km(app) (μM NH3) values were calculated by the formula described inAnthonisen et al. (1976), when Km(app) (μM NH3+ NH4+) values were only indicated in the previous study.‡

F
I G U R E 2 Electron micrographs of strain PY1.(A) Cell aggregation and (B) individual cells observed by scanning electron microscopy.(C) Ultrathin section of a cell observed using transmission electron microscopy.IM, intracytoplasmic membranes.TA B L E 2 Genome features of Nitrosomonas sp.PY1 and some AOB species.

F
I G U R E 5 Global distribution of Nitrosomonas sp.PY1-like sequences.Operational taxonomic units (OTUs) of the 16S rRNA gene were obtained from amplicon studies deposited as sequence read archive (SRA) runs.All runs with OTUs >97% identity to cultured Nitrosomonas sp.PY1 16S rRNA gene sequences were selected using an integrated microbial NGS platform (IMNGS).(A) The percent of SRA runs with Nitrosomonas sp.PY1-like OTUs within respective environments are shown with the number of SRA runs with Nitrosomonas sp.PY1-like OTUs displayed above each bar.(B) The relative abundance of all Nitrosomonas sp.PY1-like OTUs are plotted in each environment.