A novel North Sea ammonia-oxidizing archaeon Nitrosarchaeum marinum leverages a high abundance of transport systems to grow over a wide salinity range

TL;DR

Scientists isolated a new ammonia-oxidizing archaeon from the North Sea that can survive in a wide range of salt concentrations due to many transport systems.

Contribution

The first isolation and characterization of a Nitrosarchaeum strain from a fully marine environment, revealing its adaptation to broad salinity.

Findings

Nitrosarchaeum marinum T12 can grow in salinity ranging from 1 to 60 g l−1.

The strain possesses 20 different transport systems potentially involved in osmoadaptation.

Comparative genomics showed a highly diverse genetic repertoire with many unique genes in Nitrosarchaeum.

Abstract

Ammonia-oxidizing archaea (AOA) are among the most abundant micro-organisms in the biosphere. They are crucial for the global nitrogen cycle through catalyzing the oxidation of ammonia to nitrite. The biochemistry, physiology, and mechanisms underlying the adaptation of AOA to diverse habitats are not fully understood, partly due to the lack of AOA pure laboratory cultures. In this study, we present the isolation of a novel species, Nitrosarchaeum marinum T12, the first Nitrosarchaeum strain isolated from a fully marine environment. We demonstrate that this AOA can grow over a vast salinity range (1–60 g l−1). Comparative genomics of Nitrosarchaeum and Nitrosopumilus strains revealed a highly diverse genetic repertoire, with many genes unique to single species. N. marinum T12 possesses genes for chemotaxis and motility and encodes 20 different transport systems potentially involved in osmoadaptation. This diversity of transporters might provide a key for flexibility and adaptation to growth over such a wide salinity range. Strain T12 may present an example of successful adaptation to increasing salt concentration, suggesting that these AOA might be, to a certain extent, resilient to future environmental changes.

Isolation and characterization of a novel North Sea ammonia oxidizing archaeon, Nitrosarchaeum marinum T12, reveals an adaptation to a broad salinity underpinned by high abundance of transporters.

Figures

Growth characteristics of N. marinum T12. A. Growth rates at different salinities. Salinity (expressed as g l−1) includes the weight of hydrates in magnesium and calcium salts to enable comparison with previously published studies on AOA. B. Growth rates at different ammonium concentrations. C. Growth rates at different pH (black circles: media buffered with 2.5 mM MES, white circles: media buffered with 10 mM HEPES). D. Effect of pyruvate and catalase on growth (white squares: negative control, black squares: 25 U mL−1 catalase, black circles: 10 µM sodium pyruvate, grey circles: 100 µM sodium pyruvate, white circles: 1 mM sodium pyruvate). Error bars represent the standard error of biological triplicates and may be smaller than the data point.

Maximum-likelihood phylogeny of Nitrosarchaeum and Nitrosopumilus genomes. The tree was inferred from a concatenated alignment of 136 conserved non-ribosomal markers and rooted using distantly related Nitrososphaerales genomes; outgroup branch lengths are omitted (dashed branches). Branch support is indicated by dots on internal nodes: white, <75%; gray, ≥75%; black, ≥90%. The scale bar indicates amino acid substitutions per site. Clade colors denote genera. Genome environmental origin (habitat) is shown as colored squares preceding tip labels. Black bars indicate genome completeness. Colored bars indicate gene clusters: core (present in 70–83 genomes), red; cloud (present in 2–69 genomes), purple; unique (present in a single genome), yellow.

Core, cloud, and unique genomes of Nitrosarchaeum and Nitrosopumilus species (83 genomes). The core genome was defined as gene clusters present in ≥70 genomes. The cloud genome comprises gene clusters present in 2–69 genomes, and the unique genome comprises gene clusters present in a single genome. a, Number of gene clusters in the core, cloud, and unique genomes. b, Number of gene clusters with multiple annotations and without annotation. c, Abundance of gene clusters assigned to the core, cloud, and unique genomes, stratified by COG category. d, Distribution of gene clusters encoding multicopper oxidases and plastocyanins across genomes. Detailed COG function and KOfam annotations for gene clusters are provided in Supplementary Table S5. Colors: red, core; purple, cloud. MC, undefined multicopper oxidase; NirK, NirK-like multicopper oxidase.

Pangenomic comparison of Nitrosarchaeum and Nitrosopumilus genomes. a, Enrichment of gene clusters in genomes from similar habitats (marine, brackish/transition, and freshwater; adjusted q-value < 0.05; Table S6). b, Number of gene clusters with multiple COG annotations, a single COG annotation, or no COG annotation. c, Abundance of gene clusters with a single COG category annotation, summarized by COG category.

Gene clusters linked to chemotaxis and the archaeal secretion–pili system. Gene clusters encoding chemotaxis system (Che system) proteins are shown in green, gene clusters oncoding Secretion−pili system are shown in purple. ‘Potential Che’ (putative chemotaxis-associated proteins) includes gene clusters likely belonging to the Che system, including methyl-accepting chemotaxis proteins (MCPs) and CheY-like and CheB-like response regulators. The group ‘Other’ includes gene clusters encoding additional histidine kinases (HKs) and response regulators (RR) that were linked to chemotaxis based on COG or KOfam annotations. T4SS, type IV secretion system. Colored squares to the right of accession numbers denote habitat.

Gene clusters linked to osmoprotection. Gene clusters encoding proteins potentially involved in the biosynthesis of compatible solutes are colored brown (GlBe, glycine betaine; Spmd, spermidine). Gene clusters encoding transporters/channels for ions or compatible solutes are colored blue (MSC, mechanosensitive channels; SSS, sodium: solute symporter family). Colored squares to the right of accession numbers denote habitat.

Full text

Introduction

Ammonia-oxidizing archaea (AOA) are among the most ubiquitous micro-organisms on Earth. Since their discovery about 20 years ago (Venter et al. 2004, Treusch et al. 2005, Könneke et al. 2005), their diversity and abundance have been extensively studied, showing that they account for up to 40% of all microbial plankton in the ocean (Karner et al. 2001, Francis et al. 2005, Alves et al. 2018). They perform a key function of the global nitrogen cycle by catalyzing the first step of nitrification, the aerobic oxidation of ammonia to nitrite. Their metabolic activity was also linked to substantial production of the greenhouse gas nitrous oxide (Stieglmeier et al. 2014, Trimmer et al. 2016, Löscher et al. 2012, Jung et al. 2014). In addition to nitrogen cycling, AOA also significantly contribute to the marine carbon cycle. They incorporate carbon via a highly energy-efficient fixation pathway (Könneke et al. 2014) and some AOA strains also synthesize methylphosphonic acid, which acts as a precursor for methane production in aerobic ocean waters (Metcalf et al. 2012).

Besides AOA, nitrification can also be performed by ammonia-oxidizing bacteria (AOB) (Kowalchuk et al. 2001) and by complete ammonia oxidizers (comammox) (van Kessel et al. 2015, Daims et al. 2015). All groups of ammonia oxidizers have been shown to co-exist in many environments, but AOA often dominate in oligotrophic habitats such as ocean waters (Wuchter et al. 2006). Substrate affinity for ammonia was proposed as a key factor differentiating the niches of AOA and AOB (Martens-Habbena et al. 2009). This hypothesis was recently challenged but still seems applicable for AOA of the order Nitrosopumilales with their very small cell sizes (Jung et al. 2022).

Despite many efforts, the pathway and biochemistry of ammonia oxidation in AOA are still not fully understood. In contrast to the AOB, which possess an iron-based respiratory chain, AOA lack cytochrome c and rely on copper proteins for their energy metabolism (Walker et al. 2010). It is known that hydroxylamine and nitric oxide (NO) are intermediates of the archaeal pathway (Vajrala et al. 2013, Martens-Habbena et al. 2015), but the enzymology of their oxidation remains unresolved. Candidate proteins have been proposed for individual steps but face the problem of not being conserved across the entire diversity of AOA (Lehtovirta-Morley 2018 and references therein).

AOA belong to the class Nitrososphaeria within phylum Thermoproteota (formerly Thaumarchaeota) and five main lineages at the order-level are currently described: Nitrososphaerales, Nitrosopumilales, ‘Ca. Nitrosotaleales’, ‘Ca. Nitrosocaldales’ (Alves et al. 2018) and ‘Ca. Nitrosomirales’ (Zheng et al. 2024). Whereas Nitrososphaerales often dominate in neutral soils (Gubry-Rangin et al. 2011), ‘Ca. Nitrosotaleales’ seem adapted to acid environments (Lehtovirta-Morley et al. 2016, Herbold et al. 2017) and ‘Ca. Nitrosocaldales’ are found in habitats with high temperatures, typically hot springs (Daebeler et al. 2018, Abby et al. 2018). Nitrosopumilales include, besides the marine Nitrosopumilus and Nitrosopelagicus, the two genera Nitrosotenuis and Nitrosarchaeum, which have often been found in freshwater and brackish estuarine environments (Alves et al. 2018, Ren and Wang 2022). Research on the mechanisms and physiology behind the successful adaptation to such a diverse range of habitats is still in its infancy.

Global sea-level rise and the increasing frequency and intensity of hurricanes associated with climate change have tremendous effects on coastal areas, leading to the introduction of salt into coastal ecosystems (Cloern et al. 2016, Colombani et al. 2016). In addition, anthropogenic activities such as soil irrigation for agriculture, mining and the use of salt as deicer for roadways contribute to the salinization of rivers (Cañedo-Argüelles et al. 2013). Salinization of ecosystems is currently occurring at an unprecedented geographical scale and rate, with largely unknown impacts on ecology, ecosystem processes and dynamics (Herbert et al. 2015). Micro-organisms are key players in global nutrient cycling and knowledge of their response to the increasing salt concentrations is urgently needed. Furthermore, the extensive use of ammonia-based fertilizers has profoundly affected the global nitrogen cycle and microbial nitrifiers play a central role in the transformations and loss of nitrogen from the environment. AOA are the dominant nitrifiers in many ecosystems, including marine and coastal regions. Nevertheless, not much is known about their physiology and genetic potential related to osmoadaptation.

Despite their vast abundance and importance to the environment, the number of AOA isolated in pure cultures remains limited. In this work, we describe the isolation of an AOA belonging to genus Nitrosarchaeum from North Sea water. Strain T12 represents the second pure culture of the genus Nitrosarchaeum so far and is the first representative of this genus obtained from a habitat with marine salt concentrations. We demonstrate that this micro-organism can grow over a wider salinity range than closely related species. To gain further insights into its adaptation strategies, we performed comparative genomics of Nitrosarchaeum and Nitrosopumilus.

Experimental procedures

Sampling site and enrichment setup

Dutch coastal North Sea water was collected in February 2008 at high tide at the jetty of the Royal Netherlands Institute for Sea Research on the island of Texel (53°00′25’’N, 4°78′27’’E) and stored at 4°C. Sea water was filtered to concentrate it, remove large particles and passed through a 0.45 µm pore size filter. 200 ml of the filtered, concentrated North Sea water was used as inoculum and transferred to a sterile screw-cap glass bottle, supplemented with 5 mM NaHCO_3_ and 180 µM NH_4_Cl from sterile 1 M stock solutions and incubated at 22 ± 2°C, pH 7.8 ± 0.2 in the dark. This initial pre-enrichment batch culture was maintained for almost 35 months, replenishing medium and substrate whenever necessary. Hereafter, the culture was upscaled in a bioreactor culture and final isolation performed in batch culture derived from bioreactor material. Further details on the enrichment strategy and process are described in the Supplemental Material.

Isolation of an archaeal ammonia-oxidizing pure culture from the enrichment

SCMU medium, a modified version of synthetic saline (SS) medium (Supplemental Methods), was used for the isolation of strain T12 from the enrichment culture. The SCMU salts solution contained 26 g l^−1^ NaCl, 5 g l^−1^ MgCl_2_·6H_2_O, 5 g l^−1^ MgSO_4_·7H_2_O, and 0.75 g l^−1^ CaCl_2_·2H_2_O in deionized water. After the SCMU salts solution had been sterilized by autoclaving (15 min at 121°C, 15 kPa), the following media constituents were added aseptically (L^−1^): 1 ml vitamin solution, 1 ml 7.5 mM FeNaEDTA solution, 1 ml trace element solution (Supplemental Methods), 0.29 ml 1 M KH_2_PO_4_, 2 ml 1 M NaHCO_3_, 50 μL NH_4_Cl, 10 ml 1 M HEPES, and 1 ml phenol red (0.05 g l^−1^). The enrichment culture containing strain T12 was grown in 30 ml polystyrene plastic tubes (Greiner Bio-One) containing 10 ml SCMU medium at 21°C in static conditions in the dark. The growth of AOA was monitored by nitrite accumulation (Supplemental Material). A pure culture of AOA was obtained by adding antibiotics to eliminate bacterial growth, while monitoring purity via plating and screening with PCR as detailed in the Supplemental Material.

Growth experiments of strain T12

Growth experiments were performed in 30 ml polystyrene tubes (Greiner Bio-One) in triplicate. Unless otherwise stated, growth experiments were conducted in the presence of 200 μM NH_4_Cl, 1 mM sodium pyruvate (added to stimulate pure culture growth (Tourna et al. 2011; Kim et al. 2016) and 10 mM HEPES at the salinity of 20 g l^−1^ [corresponding to 13.88 g l^−1^ (0.24 M) NaCl, 2.85 g l^−1^ (14.0 mM) MgCl_2_·6H_2_O, 2.85 g l^−1^ (11.6 mM) MgSO_4_·7H_2_O, and 0.43 g l^−1^ (2.9 mM) CaCl_2_·2H_2_O]. To initiate growth experiments, strain T12 cells were grown in 1 L Duran bottles, harvested during mid-exponential growth phase, washed in HEPES-buffered SCMU salt solution (at 20 g l^−1^ salts) and inoculated into the tubes. Nitrite concentration was monitored every 2–3 days for up to 23 days. For the salinity range experiments, the salinities from 0.1 to 60 g l^−1^ salts were used [this value includes hydrates of magnesium and calcium salts (e.g. Elling et al. 2017)]. To allow comparison between strain T12 and previously published strains of AOA, the salinity values (g l^−1^) are expressed with the hydrates throughout the manuscript unless otherwise stated. The ratio of NaCl, MgCl_2_, MgSO_4_, and CaCl_2_ was kept constant in all treatments. For the pH range experiment, experiments at the lower end of the pH range (4.5–7.0) were performed using 2.5 mM MES [2-(N-morpholino)ethanesulfonic acid] and at the higher end of the range (7.0–9.0) using 10 mM HEPES.

DNA extraction

DNA was extracted using a CTAB (cetyltrimethylammonium bromide) extraction buffer and SDS lysis-based method adapted from the previously published protocol (Zhou et al. 1996). Briefly, reactor samples were lysed in the presence of CTAB, lysozyme, RNAse A, SDS, and proteinase K, DNA extracted using phenol-chloroform-based approach and precipitated using isopropanol, before resuspending the pellet in dH_2_O. Details of the method are described in the Supplemental Material.

Metagenomic sequencing

An aliquot of 1 µg of high-molecular-weight DNA from the reactor culture collected on day 455 was used for metagenomic sequencing. All kits used were obtained from Life Technologies (Life Technologies, Carlsbad, CA, USA). DNA was sheared for 5 min using the IonXpress™ Plus Fragment Library Kit following the manufacturer’s instructions. Further library preparation was performed using the Ion Plus Fragment Library Kit following the manufacturer’s instructions, with size selection using an E-gel 2% agarose gel. The library was used for two sequencing runs, for which emulsion PCR was performed using the OneTouch 200 bp kit. Sequencing was then conducted on an Ion Torrent PGM using the Ion PGM 200 bp sequencing kit and an Ion 318 chip, resulting in 734.4 Mbp (ERR4672195) and 583.5 Mbp (ERR4761584) of sequencing data in single reads (raw data are available under ENA accession number PRJEB40748).

Metagenomic data analysis

Kmer analysis of the metagenomic datasets was performed using Jellyfish (Marçais and Kingsford 2011). Metagenome assembly was performed using the CLC genomics workbench (version 8, CLCbio, Arhus, Denmark) with word size 35 and bubble size 5000. GC content of the resulting 7078 contigs was calculated using “fasta_to_gc_cov_length_tab.pl” script (https://github.com/dspeth/bioinfo_scripts/blob/master/binning/manual_R_binning/fasta_to_gc_cov_length_tab.pl). Contigs were binned based on contig sequencing depth and GC content using R. After manual improvement, an AOA draft genome divided over 25 contigs was obtained and a NOB draft genome on 207 contigs. BLAST search of the metagenome assembly before binning using Nitrospina 16S rRNA gene sequences resulted in the retrieval of a full-length 16S rRNA Nitrospina gene sequence from the data.

Genome completeness was estimated using CheckM2 (https://doi.org/10.1038/s41592-023-01940-w). The AOA draft genome (NmT12) was explored using the MicroScope platform (Vallenet et al. 2020). For comparative genomics of Nitrosarchaeum marinum T12, the Genome Taxonomy Database (GTDB, v226 released Apr 16, 2025) was queried for reference genomes (Parks et al. 2022). Additional genomes were retrieved using NCBI datasets command-line tool (https://doi.org/10.1038/s41597-024-03571-y) using the options: ‘taxon 1 007 082 –assembly-source all –assembly-version latest –exclude-atypical’. The genome of Nitrosarchaeum haohaiensis sp. nov. CL1^T^ (https://doi.org/10.1111/1758-2229.70100), was provided by the authors upon request. All available genomes of the genera Nitrosarchaeum (68 genomes) and Nitrosarchaeum_A (4 genomes) were selected and contig sequences were downloaded from the NCBI GenBank or NCBI RefSeq databases (French et al. 2012, O’Leary et al. 2016, Bollmann et al. 2024, Ghimire-Kafle et al. 2024). In addition, 10 high-quality genomes of the genus Nitrosopumilus were selected. For phylogenomic tree construction, 11 additional, more distantly related genomes of the order Nitrososphaerales were downloaded. A description of all genomes used in the analyses is provided in Table S1.

Phylogenomic tree construction

For phylogenomic reconstruction, a curated set of 136 non-ribosomal markers (NM) was used (https://doi.org/10.1093/molbev/msv015); (https://doi.org/10.1038/s41564-024-01647-4). Open reading frames (ORFs) in all genomes were predicted using Pyrodigal v3.6.3 (https://doi.org/10.21105/joss.04296); (https://doi.org/10.1186/1471-2105-11-119). Predicted ORFs were queried against the set of 136 conserved markers using DIAMOND blastp v2.1.12 (https://doi.org/10.1038/nmeth.3176). Best hits were aligned with MAFFT v7.526 (https://doi.org/10.1093/molbev/mst010) using the options ‘–maxiterate 1000 –localpair’. Alignments were trimmed with trimAl v1.5.rev0 (https://doi.org/10.1093/bioinformatics/btp348) using the automated1 algorithm and subsequently concatenated using seqkit v2.10.1 (https://doi.org/10.1002/imt2.191). The maximum-likelihood phylogeny was inferred from 94 sequences and 41 803 distinct patterns using IQ-TREE 3 v3.0.1 (https://doi.org/10.32942/X2P62N) under the LG+C30+F+G4 substitution model. Branch support was assessed using 1000 ultrafast bootstrap replicates (UFBoot; –ufboot 1000) (https://doi.org/10.1093/molbev/msx281) with additional bootstrap NNI optimization (–bnni), and 1000 SH-like aLRT replicates (–alrt 1000). The resulting phylogeny was visualized in R v4.3.0 (R Core Team. R: A Language and Environment for Statistical Computing. 2022) using ggtree v3.8.2 (https://doi.org/10.1002/imt2.56) (Yu et al., 2020), ggtreeExtra v1.10.0 (https://doi.org/10.1093/molbev/msab166), and ggstar v1.0.4 (https://cran.r-project.org/web/packages/ggstar/index.html).

Pangenomic analysis

Clusters of orthologous genes shared between Nitrosarchaeum and Nitrosopumilus genomes were determined using the Anvi’o v8 (marie) pangenomics workflow (Eren et al 2015, Eren et al. 2021). A summary of the analysis parameters and the output table is given in the supplementary file ‘Nitrosarchaeum_Anvio_Summary.zip’. For the functional enrichment analysis, the ‘anvi-compute-functional-enrichment-in-pan’ algorithm was used (Shaiber et al. 2020). Individual gene clusters were investigated independently using the gene cluster (GC) identity as function and the genome origin (‘Habitat’; Fig. 3, Table S1) was used as category variable. The resulting output table was further processed using the dplyr v1.1.4 (Wickham et al. 2022) and the tidyr v1.3.1 (Wickham and Girlich 2022) packages implemented in R v4.3.0 (R Core Team 2022). The results were filtered for habitats containing more than one genome and gene cluster enrichment scores with an adjusted q-value <0.05 (Shaiber et al. 2020). NMDS was performed using the ‘metaMDS’ function of the vegan v2.6–4 package (Oksanen et al. 2022), with Bray–Curtis distance and 1 000 maximum random starts to find a stable solution. Plots were created using the ggplot2 v3.5.1 package (Wickham 2016).

Results and discussion

Enrichment and isolation of a novel AOA belonging to genus Nitrosarchaeum

The initial enrichment of the North Sea water in glass bottles, followed by incubation in a bioreactor, resulted in a nitrifying co-culture of an archaeal ammonia oxidizer affiliated to genus Nitrosarchaeum and a bacterial nitrite oxidizer affiliated to genus Nitrospina (Fig. S1). Previous enrichment studies using from same North Sea location resulted in a co-culture of a bacterial ammonia oxidizer of the genus Nitrosomonas and the nitrite-oxidizer ‘Ca. Nitrospira salsa’ (Haaijer et al. 2013). The main differences between the two approaches were the use of lower substrate concentrations in the current study, particularly at the start of the enrichment (180 µM ammonium versus 500 µM ammonium) and the filtering of the seawater to select for the smaller AOA cells.

Interestingly, the dominant archaeal ammonia oxidizer changed during the initial pre-enrichment phase, from a Nitrosopumilus-like species to a Nitrosarchaeum-like species (see Supplemental Results). This phase involved two major parameter changes: the incubation method was switched from shaking to static incubation and the filtered North Sea water medium was replaced with a synthetic high-salinity medium. Either of these parameters, or a combination of both, apparently led to the selection of the Nitrosarchaeum strain. Whether changing these parameters would reproducibly lead to selective enrichment of Nitrosarchaeum should be further explored in the future. The enrichment showed a maximum cell-specific ammonia oxidation activity of 0.71 fmol day^−1^ (Fig. S1 and Supplemental Material). Further details on the enrichment and the reactor performance can be found in the Supplemental Material.

To isolate the AOA strain T12, the enrichment culture was transferred multiple times in liquid medium supplemented with antibiotics, including kanamycin, streptomycin, spectinomycin, and ciprofloxacin. The sequential treatment with antibiotics ultimately yielded a pure culture of strain T12, which was PCR-negative for primers targeting bacteria, a single morphotype under the microscope, and free of heterotrophic bacterial growth on solid medium, for which we propose the name N. marinum on the basis that comparative genomics indicated sufficient evolutionary distance to the previously described Nitrosarchaeum species (see section ‘Genome and phylogeny of N. marinum T12’). After removing the co-occurring bacteria, prolonged lag phases and reduced growth rates of N. marinum T12 were observed. It was hypothesized that the growth decline may be due to oxidative stress, as previously reported for other AOA strains (Kim et al. 2016). The pure N. marinum T12 cultures were therefore supplemented with catalase, which restored more rapid growth.

Physiology of Nitrosarchaeum marinum T12

The pure N. marinum T12 culture grew chemolithoautotrophically with ammonia as the sole energy source. The highest growth rate was 0.24 d^−1^, which equates to an approximate doubling time of 2.9 days. Whilst the observed growth rate was lower than reported for the marine AOA Nitrosopumilus maritimus (growth rate 0.65 d^−1^) and freshwater enrichment cultures BO1 and AC2 belonging to genus Nitrosarchaeum (∼0.35 and 0.48 d^−1^, respectively), the rate was like that of Nitrosarchaeum limnium SFB1 (0.2 d^−1^) (Mosier et al. 2012, French et al. 2012, Ghimire-Kafle et al. 2024). This is perhaps unsurprising as growth rates of different species within an AOA genus often vary quite substantially, not only in marine genera like Nitrosopumilus (0.31–0.88 d^−1^) (Bayer et al. 2019) but also in other AOA genera like Nitrosotalea (0.29–0.60 d^−1^) and Nitrosocaldus (0.13–0.72 d^−1^) (Lehtovirta-Morley et al. 2014, Abby et al. 2018, Daebeler et al. 2018). Growth was furthermore tested at a range of different salinities (Fig. 1A). Nitrosarchaeum marinum T12 was able to grow over a surprisingly broad salinity range of 1-60 g l^−1^ (or 1–60‰) (Fig. 1A). This contrasts with previously characterized Nitrosopumilus isolates, which cannot grow at salinities below 10–15‰ [Bayer et al. 2019 (salinity adjusted by amending NaCl), Qin et al. 2017]. It is also distinctly different from the only previous Nitrosarchaeum isolate, N. koreense MY1, which grows in a narrow salinity range of only 1‰–4‰, possibly reflecting its terrestrial origin (Jung et al. 2011). While the full salinity range of Nitrosarchaeum limnium enrichment has not been reported, its growth was reduced at 75% seawater salinity (∼26‰), which resulted in only partial conversion of the supplied ammonium [Mosier et al. 2012 (salinity adjusted by diluting seawater medium)]. Recently, a new aquatic AOA, Nitrosarchaeum haohaiensis CL1 was cultivated from estuarine waters in East China Sea (Li et al. 2025), which, although not from a fully saline environment, also exhibits a wider salinity range than the previously cultivated Nitrosarchaeum strains. For N. marinum T12, the highest growth rates were observed at a concentration of 5‰–20‰. Even at 35‰, a salinity similar to seawater and to typical growth media of marine AOA from the genus Nitrosopumilus, the growth rate of N. marinum T12 was 0.19 d^−1^, which is ∼80% of the highest growth rate recorded. These results challenge the idea that the genus Nitrosarchaeum is specifically adapted to fresh and brackish water. Although amoA sequences related to Nitrosarchaeum have been detected in fully marine habitats (Alves et al. 2018), it had not been clear until now whether Nitrosarchaeum strains could thrive in such environments. Many marine AOA are adapted to low ammonium concentrations and while the oligotrophic open ocean has very low ammonium concentrations, coastal regions can experience higher ammonium loads. To investigate the effect of ammonium on growth, N. marinum T12 was grown over a range of ammonium concentrations, from 10 μM to 10 mM (Fig. 1B). The highest growth rates occurred at ammonium concentrations of 1 mM and below. Although the culture was able to grow even in the presence of 10 mM ammonium, the growth rate was reduced to 0.10 d^−1^ (Fig. 1B). The ammonium range supporting the growth of our isolate is in line with observations from N. koreense MY1 and several Nitrosopumilus representatives (Jung et al. 2011, Bayer et al. 2019), and slightly broader than that of strain AC2 (French et al. 2012). Nitrosarchaeum marinum T12 grew over a pH range of 6.0–9.0, with fastest growth at pH 6.5–8.5 (Fig. 1C). This pH range is similar or slightly broader than previously reported for N. koreense MY1 and AC2 (Jung et al. 2011, French et al. 2012) and various Nitrosopumilus isolates (Qin et al. 2017, Bayer et al. 2019). Supplementation with pyruvate or catalase was required for growth and no, or only very little, nitrite accumulation, was detected in the control treatments without their addition (Fig. 1D). This has been previously reported in other AOA strains and supports the idea that some AOA rely on other members of the microbiome to detoxify reactive oxygen species (Kim et al. 2016).

Growth characteristics of N. marinum T12. A. Growth rates at different salinities. Salinity (expressed as g l−1) includes the weight of hydrates in magnesium and calcium salts to enable comparison with previously published studies on AOA. B. Growth rates at different ammonium concentrations. C. Growth rates at different pH (black circles: media buffered with 2.5 mM MES, white circles: media buffered with 10 mM HEPES). D. Effect of pyruvate and catalase on growth (white squares: negative control, black squares: 25 U mL−1 catalase, black circles: 10 µM sodium pyruvate, grey circles: 100 µM sodium pyruvate, white circles: 1 mM sodium pyruvate). Error bars represent the standard error of biological triplicates and may be smaller than the data point.

The cells are small rods (<1 µm in length) and similar in size and appearance to representatives from the genus Nitrosopumilus and ‘Ca. Nitrosarchaeum limnium SFB1’ and Nitrosarchaeum koreense MY1 (Konneke et al. 2005, Jung et al. 2011, Bayer et al. 2019). Both electron-light and electron-dense areas were previously reported in the thin sections of cells of N. koreense MY1, along with a structure consisting of an inner membrane and a periplasmic space (Jung et al. 2011). Transmission electron microscopy of the cells of N. marinum T12 suggested subcellular compartmentalization, but a structure comparable to that of N. koreense MY1 was not visible (Fig. S2).

Genome and phylogeny of Nitrosarchaeum marinum T12

A metagenome of the bioreactor was generated at the enrichment stage. The reactor was dominated by a co-culture of an AOA species belonging to Nitrosarchaeum and a previously uncultured species of Nitrospina, and their genomes constituted 81.9% and 7.2% of the total single reads, respectively. The classification of the Nitrospina strain is further described in the Supplemental Material (Fig. S3).

The near-complete (>99%) draft genome of the AOA strain comprises 1 738,407 bp across 25 contigs, with an average coverage of 500-fold and a GC content of 32.65% (Fig. S4). This is similar to the GC content of the published Nitrosarchaeum and Nitrosopumilus genomes (Table S1) and lower than in many soil AOA (Qin et al. 2020). The genome was most closely affiliated with the genus Nitrosarchaeum. The highest average nucleotide identity (ANI) revealed closest similarities to Ca. Nitrosarchaeum limnium BG20 (GCF_000241145.1; 92.37%) and Ca. Nitrosarchaeum limnium SFB1 (GCF_000204585.1; 92.03%; (Table S2)) (Mosier et al., 2012a; Blainey et al., 2011). This is below the recommended ANI species cut-off of 95% (Jain et al. 2018, Lee et al. 2019), indicating that the North Sea AOA represents a novel species within the genus Nitrosarchaeum. In phylogenomic analysis, strain T12 forms a clade with the two aforementioned closely related ‘Ca. N. limnium’ strains (Fig. 2) enriched from the San Francisco Bay estuary, an environment exposed to fluctuating salinity gradients (Blainey et al. 2011, Mosier et al. 2012b). Interestingly, these three marine/estuarine genomes are affiliated with a clade containing genomes obtained from highly diverse environments (Fig. 2). In contrast, the remaining genomes of Nitrosarchaeum were all retrieved from freshwater ecosystems and seem to cluster predominantly according to their habitat. Whereas representatives of genus Nitrosopumilus are generally abundant in many marine environments, with relatively few exceptions (e.g. Walker et al. 2010, Alves et al. 2018, Qin et al. 2020, Herber et al. 2020), genus Nitrosarchaeum have been considered as a freshwater or low-salinity lineage (Mosier et al. 2012b, Ren and Wang 2022). Previous distribution analysis of amoA sequences suggested high occurrence of Nitrosarchaeum in coastal and estuarine habitats (Alves et al. 2018), but sequences have also been retrieved from freshwater, soil, and even marine origins. Thus, the term cosmopolite might best describe the distribution of this AOA group in the environment (Alves et al. 2018), even though sub-clades within the Nitrosarchaeum genus might be more specialized and adapted to distinct freshwater environments (Fig. 2).

Maximum-likelihood phylogeny of Nitrosarchaeum and Nitrosopumilus genomes. The tree was inferred from a concatenated alignment of 136 conserved non-ribosomal markers and rooted using distantly related Nitrososphaerales genomes; outgroup branch lengths are omitted (dashed branches). Branch support is indicated by dots on internal nodes: white, <75%; gray, ≥75%; black, ≥90%. The scale bar indicates amino acid substitutions per site. Clade colors denote genera. Genome environmental origin (habitat) is shown as colored squares preceding tip labels. Black bars indicate genome completeness. Colored bars indicate gene clusters: core (present in 70–83 genomes), red; cloud (present in 2–69 genomes), purple; unique (present in a single genome), yellow.

Comparative genomics of Nitrosarchaeum and Nitrosopumilus

To gain further insights into the environmental adaptation strategies of Nitrosarchaeum, we analyzed the pangenome of all Nitrosarchaeum and closely related Nitrosopumilus genomes. 11 gene clusters were shared by all 83 genomes included in the analysis, representing the core genome in the strict sense (Fig. 2). This resulted in an extended core set of 1058 gene clusters conserved in the Nitrosarchaeum-Nitrosopumilus pangenome (ANI of 75–99%). These core genes accounted for ∼50% of the genes in the complete genomes and up to 78% in the incomplete genomes (Fig. 2).

The number of gene clusters unique to a genome was highly variable (Fig. 2) and partially depended on genome completeness. We identified a total of 13,564 gene clusters, with 1–684 unique genes per genome (an average of 104 unique genes per genome, or 73 genes per Mbp). Previous studies reported an average of ∼111 novel, unique genes per Mbp with each new marine AOA strain sequenced (Qin et al. 2020). In our study, 73 genes per Mbp were found, which is within a similar range and confirms the extensive genomic diversity of AOA (Qin et al. 2020).

The number of genes not annotated using the COG database was, as expected, much higher in the cloud and unique genome than in the core genome (Fig. 3). In the core genome, the largest number of gene clusters was assigned to the category ‘Translation’ (∼14% of all gene clusters), whereas genes related to ‘Signal transduction’, ‘Cell wall’, ‘Mobilome’, and ‘Defense mechanisms’ were more abundant in the cloud and unique genome (Fig. 3A).

Core, cloud, and unique genomes of Nitrosarchaeum and Nitrosopumilus species (83 genomes). The core genome was defined as gene clusters present in ≥70 genomes. The cloud genome comprises gene clusters present in 2–69 genomes, and the unique genome comprises gene clusters present in a single genome. a, Number of gene clusters in the core, cloud, and unique genomes. b, Number of gene clusters with multiple annotations and without annotation. c, Abundance of gene clusters assigned to the core, cloud, and unique genomes, stratified by COG category. d, Distribution of gene clusters encoding multicopper oxidases and plastocyanins across genomes. Detailed COG function and KOfam annotations for gene clusters are provided in Supplementary Table S5. Colors: red, core; purple, cloud. MC, undefined multicopper oxidase; NirK, NirK-like multicopper oxidase.

To investigate variation in gene-cluster composition across AOA genomes, we performed non-metric multidimensional scaling (NMDS) on a genome-by-gene-cluster matrix derived from the Anvi’o pangenome output. Bray–Curtis dissimilarities were calculated between genomes and ordinated in two dimensions (k = 2; stress = 0.158; Fig. S5). Nitrosopumilus and Nitrosarchaeum separated primarily along NMDS1, consistent with their phylogenetic separation and habitats of origin (Figs 2 and 4, S4, and Supplemental Material).

Pangenomic comparison of Nitrosarchaeum and Nitrosopumilus genomes. a, Enrichment of gene clusters in genomes from similar habitats (marine, brackish/transition, and freshwater; adjusted q-value < 0.05; Table S6). b, Number of gene clusters with multiple COG annotations, a single COG annotation, or no COG annotation. c, Abundance of gene clusters with a single COG category annotation, summarized by COG category.

Genomic insights into chemotaxis and motility

The ability to respond to the environment by swimming can be advantageous in dynamic ecosystems where environmental parameters, such as substrate or salt concentrations, change rapidly. Therefore, the genomes were compared to identify gene clusters responsible for chemotaxis and motility. The analysis revealed 54 gene clusters associated with chemotaxis and 48 gene clusters encoding proteins involved in archaeal secretion-pili systems (Fig. 5, Table S5 and Supplemental Material). Gene clusters encoding the core chemotaxis proteins CheA and CheW were found only in six of the analyzed Nitrosarchaeum and Nitrosopumilus genomes (Table S7 and Supplemental Material). Besides N. marinum T12, the gene clusters are present in three Nitrosarchaeum genomes (GCA_021414365.1, GCF_000204585.1, GCF_000241145.1), and two Nitrosopumilus strains, N. adriaticus NF5 (GCF_000956175.1) and N. ureiphilus PS0 (GCF_013407185.1). All these genomes also possess gene clusters encoding the CheC system and for archaellum biosynthesis. Therefore, motility appears to be a relatively rare feature in aquatic AOA, considering that the majority of the available aquatic AOA genomes lack the capacity for archaellum biosynthesis. In contrast, nearly all the potentially motile aquatic AOA originate from coastal waters or sediments (Qin et al. 2017, Bayer et al. 2019). Although N. marinum T12 was not motile under the growth conditions of this study, and archaellum and pili biosynthesis may only be induced in response to specific environmental triggers, micro-organisms in general can exhibit osmotaxis (Rosko et al. 2017). Virtually nothing is known about environmental cues required for AOA motility, although N. limnium SFB1 was reported to be motile in laboratory culture (Blainey et al. 2011).

Gene clusters linked to chemotaxis and the archaeal secretion–pili system. Gene clusters encoding chemotaxis system (Che system) proteins are shown in green, gene clusters oncoding Secretion−pili system are shown in purple. ‘Potential Che’ (putative chemotaxis-associated proteins) includes gene clusters likely belonging to the Che system, including methyl-accepting chemotaxis proteins (MCPs) and CheY-like and CheB-like response regulators. The group ‘Other’ includes gene clusters encoding additional histidine kinases (HKs) and response regulators (RR) that were linked to chemotaxis based on COG or KOfam annotations. T4SS, type IV secretion system. Colored squares to the right of accession numbers denote habitat.

Genomic potential for osmoprotection

As the isolated strain T12 is capable of growth across a large range of salinity, we were interested in potential salt adaptation mechanisms, including biosynthesis of compatible solutes and transport systems for the uptake of solutes or potassium (Gunde-Cimerman et al. 2018, Bremer and Kraemer 2019). Mechanosensitive channels (MSC) enable a very quick reaction to changing osmotic conditions by opening in response to mechanical extension of the cytoplasmic membrane, and MscL channels have a large gate opening compared to McsS channels (Kung et al. 2010 and references therein). Most of the Nitrosarchaeum and Nitrosopumilus genomes possessed MSC-encoding gene clusters (Fig. 6, Table S6). Interestingly, only the Nitrosarchaeum species encoded a large MSC protein (MscL), whereas all species possess small MSC proteins (MscS). This is consistent with previous observations (Blainey et al. 2011) and with the idea that MscL is required for adaptation to low-salinity habitats. In addition, species within Nitrosarchaeum and Nitrosopumilus possess distinct sets of MscS-encoding gene clusters. An exception is the cosmopolite Nitrosarchaeum subgroup, including strain T12 (Fig. 6). This group possesses a great number of MSC proteins (MscS and MscL) and partly shares the gene clusters of both genera, Nitrosarchaeum and Nitrosopumilus (Fig. 6, Table S6). A combination of many different channels enables the cell to have a transient and energy-efficient response to different osmotic conditions. The marine Nitrosopumilus genomes analyzed in our study lack MscL channels, but all possess multiple MscS channels, ranging from two to four. This might be an adaptation to the higher, but more consistent osmolarity in ocean waters. The role of MscS in adaptation to high salinity has also been experimentally verified in N. maritimus SCM1 and exposure of N. maritimus to high salinity causes a transcriptional upregulation of the mscS gene (Widderich et al. 2016). Nitrosarchaeum marinum T12 possesses a large MscL channel, as all other Nitrosarchaeum strains. However, the genome also encodes five different MscS channels, like the marine Nitrosopumilus species.

Gene clusters linked to osmoprotection. Gene clusters encoding proteins potentially involved in the biosynthesis of compatible solutes are colored brown (GlBe, glycine betaine; Spmd, spermidine). Gene clusters encoding transporters/channels for ions or compatible solutes are colored blue (MSC, mechanosensitive channels; SSS, sodium: solute symporter family). Colored squares to the right of accession numbers denote habitat.

In addition to MSCs, potassium transporters play an essential role in cell homeostasis and osmoadaptation (Beagle and Lockless 2021). A putative voltage-gated potassium channel (GC_00 002 545; Table S6) was also enriched in the marine AOA genomes. We identified a single gene cluster encoding a TrkG-like transport protein which is present in nearly all genomes (TrkG1; Fig. 6) and most Nitrosarchaeum species possess an additional gene cluster. Few genomes, including the cosmopolite Nitrosarchaeum group, contain multiple copies of one of these gene clusters. Nevertheless, the corresponding TrkA protein required for regulation was only found in eight genomes, including strain T12 (Fig. 6, Table S6). The Nitrosopumilus genomes and those of the cosmopolite Nitrosarchaeum group also possess NhaP-type cation/proton antiporters, which mediate the transport of sodium or potassium ions across the membrane (Masrati et al. 2018) and may provide the cells with additional flexibility to cope with osmotic stress. All genomes but one encoded Kef-type proteins involved in potassium efflux (Lyngberg et al. 2011). All Nitrosopumilus and freshwater Nitrosarchaeum species possess, on average, two to six gene clusters. An exception is again the cosmopolite Nitrosarchaeum subgroup, in which genomes possess two to eight different KefB-encoding gene clusters.

We furthermore screened the genomes for genes involved in the synthesis or transport of compatible solutes (Fig. 6, Table S6). Of the many small organic compounds potentially involved in osmoprotection (Mueller et al. 2005), we identified ectoine biosynthesis in 10 genomes, a potential spermidine synthase in 14 genomes and a choline dehydrogenase-like enzyme (BetA), which is hypothetically involved in glycine betaine synthesis, in only 7 genome. Only few compatible solute transporters, with a potential sodium/solute symporter (SSS family; PutP) were present in mostly freshwater Nitrosarchaeum genomes. Additionally, two gene clusters encoding potential urea/proton symporters [Urea1(1,2)] were identified in several genomes. However, as the genomes encoding these symporters also encode for the urease enzyme (Table S7), urea might rather be used as a substrate and not as a compatible solute. Another gene cluster (Urea2) was annotated as substrate-binding protein of a putative urea ABC transporter. This gene cluster is present in six genomes, independent of the presence of the urease-encoding genes (Fig. 6, Table S6, Table S7). It was also identified in the genome of N. marinum T12 and may be associated with a function in osmoadaptation. Lastly, almost all genomes contain aquaporin-encoding gene clusters that might be involved in an early adaptation to osmotic stress (Roberts, 2005).

Conclusion

Archaeal ammonia oxidizers are among the most ubiquitous micro-organisms in the biosphere. Despite extensive research over the past decades, only few enrichment cultures and even fewer isolates are described to date. In this study, we presented the isolation, physiology and comparative genomics of a novel Nitrosarchaeum species enriched from North Sea waters. Nitrosarchaeum marinum T12 belongs to a group of AOA found in a diverse set of environments, including soil, freshwater, brackish/transition, and engineered aquatic. It is the first Nitrosarchaeum enriched from a truly marine habitat and can grow over an exceptionally broad range of pH and salinity. Comparative genomics of Nitrosarchaeum and selected Nitrosopumilus strains revealed a highly diverse genetic repertoire, with many genes unique to individual genomes, even within this small phylogenetic distance. It may help these AOA to quickly adapt and be successful in various environments. Nitrosarchaeum marinum T12 possesses genes related to chemotaxis and motility and encodes a high abundance of different transport systems potentially involved in osmoadaptation. Here, it shares a basic set of genes common to all Nitrosarchaeum strains, but also possesses additional gene clusters only enriched in a subgroup of apparently cosmopolitan Nitrosarchaeum species. Furthermore, it contains gene clusters otherwise only present in marine Nitrosopumilus species. This high diversity of transport systems may be a key to flexibility and adaptation to growth over such a wide range of salinity. Strains of the genus Nitrosarchaeum thrive in many environments severely impacted by salinization as a result of climate change and anthropogenic activities. Strain T12 may present an example of adaptation to increasing salinity, suggesting that these AOA might be, to a certain extent, resilient to such changes. Nevertheless, the rate and extent of such an adaptation still need to be investigated. Furthermore, experimental research is needed to verify the involvement of the proteins encoded by the proposed gene clusters in osmoadaptation. Having successfully cultivated N. marinum T12 as a pure culture in the laboratory makes such research possible.

Description of Nitrosarchaeum marinum sp. nov.

Nitrosarchaeum marinum (ma.ri´num. L. neut. adj. marinum of the sea, marine).

Cells are small and slender, rod-shaped, with a length of < 1 µm. They grow by oxidizing ammonia to nitrite chemolithoautotrophically. Ammonia is their sole energy source. T12^T^ represents the type strain, isolated from marine North Sea waters close to the island of Texel, the Netherlands. No archaellum could be seen under the microscope, however, strain T12^T^ has the genomic potential to express an archaellum. It grows over a salinity range of 1 to 60 g l^−1^, a pH range of 6.0–9.0 and at 21°C–22°C. Strain T12^T^ requires the addition of pyruvate or catalase (but not both) to grow and can tolerate up to 10 mM ammonium. The genome is available under ENA accession number PRJEB40748 and GenBank accession number GCA_904848535.1.

Conflicts of interest

None declared.

Supplementary Material

fiag013_Supplemental_Files