Hydrogen-dependent dissimilatory nitrate reduction to ammonium enables growth of Campylobacterota isolates

Abstract

Dissimilatory nitrate reduction to ammonium (DNRA) is a key process used by diverse microorganisms in the global nitrogen cycle. For long, DNRA has been considered primarily as an organotrophic reaction, despite evidence that oxidation of inorganic electron donors also supports DNRA. Evidence of DNRA coupling with molecular hydrogen (H2) oxidation has been reported for several microbial isolates; however, the underlying physiology of the microbial process remains understudied. In this study, we report the isolation of two Campylobacterota strains, Aliarcobacter butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2, which grow using H2 as the sole electron donor for DNRA, and physiological insights gained from a close examination of hydrogenotrophic DNRA in these isolates. In both batch and continuous cultures, DNRA sensu stricto (i.e. NO3- reduction that includes stoichiometric NO2--to-NH4+ reduction) was strictly dependent on the presence of H2 and exhibited stoichiometric coupling with H2 oxidation, indicating that electrons required for NO2- reduction were unequivocally derived from H2. Successful chemostat incubation further demonstrated that hydrogenotrophic DNRA is viable under NO3--limiting, H2-excess conditions. Genomic and transcriptomic analyses identified group 1b [NiFe]-hydrogenase and cytochrome c552 nitrite reductase as the key enzymes catalyzing hydrogenotrophic DNRA. In addition, metagenomic surveys revealed that bacteria capable of hydrogenotrophic DNRA are taxonomically diverse and abundant in various ecosystems, particularly in the vicinity of deep-sea hydrothermal vents. These findings, integrating physiological, genomic, and transcriptomic analyses, clarify that H2 can solely serve as a growth-supporting electron donor for DNRA and suggest potential significance of this microbial process in nitrogen- and hydrogen-related environmental biogeochemical cycles.

Keywords: Campylobacterota; dissimilatory nitrate reduction to ammonium (DNRA); hydrogenases; nitrogen cycling; respiratory hydrogen oxidation.

Figure 1

Phylogenetic affiliation, cellular morphology, and growth characteristics of A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2 during hydrogenotrophic DNRA. (A) Maximum-likelihood phylogenetic tree of 92 Campylobacterota genomes, including the two isolates, constructed using IQ-TREE v2.3.6. The tree is based on the alignment of the concatenated amino acid sequences from 120 single-copy bacterial marker genes identified with GTDB-Tk v2.4.0 (release220). Three Nitrosomonas spp. genomes served as the outgroup. Branch support values were derived from 1000 ultrafast bootstrap replicates and the SH-aLRT single-branch test; bifurcations with both support values > 80% are marked with filled black circles. A grid plot to the right of the tree visualizes the inventories of functional genes encoding dissimilatory NO₃⁻/NO₂⁻ reductases and hydrogenases. (B) SEM (top) and TEM (bottom) images of A. butzleri hDNRA1 (left) and Sulfurospirillum sp. hDNRA2 (right). Scale bars in the micrographs represent 1 μm and 0.2 μm for SEM and TEM images, respectively. (C) Batch-culture growth curves of A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2, grown with H₂, acetate, and NO₃⁻. Each data point represents the mean of three biological replicates (n = 3), with error bars indicating standard deviations.

Figure 2

Hydrogenotrophic DNRA and H₂ oxidation kinetics in A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2. (A, B, E, F) Reductive transformation of NO₃⁻ (2 mM; 200 μmol per bottle) was monitored in batch cultures under two conditions: in the absence (A, E) and presence (B, F) of 5% (v/v) H₂ in the headspace. Concentrations of NO₃⁻, NO₂⁻, NH₄⁺, and acetate were monitored, with dotted vertical lines indicating the NO₂⁻ peak (t=10 h) and the depletion of NO₃⁻/NO₂⁻ (t=16 h). (C, G) H₂ consumption during the experiments (B, F) was tracked over time to assess its role as an electron donor for DNRA. Each data point represents the mean of three biological replicates (n = 3), with error bars indicating standard deviations. (D, H) Michaelis–Menten kinetics of H₂ oxidation were evaluated for whole cells pregrown on hydrogenotrophic DNRA. Initial H₂ oxidation rates were calculated from batch incubations (n = 3) initiated with the indicated molar concentrations of dissolved H₂ on the x-axis.

Figure 3

Coupling of H₂ oxidation with DNRA in H₂-limited batch cultures and chemostat reactor of A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2. (A, B) Hydrogen-limited batch cultures were initiated with a stoichiometrically limiting amount of H₂ (200 μmol bottle⁻¹; H₂-to-NO₃⁻ molar ratio of 1) and replenished with H₂ (200 μmol bottle⁻¹) after its depletion (indicated by black arrows). The concentrations of NO₃⁻, NO₂⁻, NH₄⁺, H₂, and acetate were monitored until the complete depletion of NO₃⁻/NO₂⁻ and are presented as total amounts in the bottles. Each data point represents the mean of three biological replicates (n = 3), with error bars indicating standard deviations. (C, D) Hydrogenotrophic DNRA was demonstrated in lab-scale chemostat reactors. For Sulfurospirillum sp. hDNRA2, experimental conditions were optimized during incubation to achieve a steady state, with vertical dotted lines marking the time points of changes (details provided below the figure). At the end of the incubation, H₂ was removed from the gas stream to evaluate its impact on DNRA activity.

Figure 4

Genomic and transcriptomic insights into hydrogenotrophic DNRA in A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2. (A, B) Gene clusters putatively associated with hydrogenotrophic DNRA identified from the complete genomes of A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2. (C, D) Differential gene expression analyses of functional genes involved in hydrogen oxidation and nitrogen metabolism (top). Genome-wide transcriptional changes under H₂-free (acetate and formate used as alternate electron donors for A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2, respectively) versus H₂-supplemented conditions are shown in volcano plots (bottom). Datasets from three biological replicates (n = 3) are presented for each condition. Maximum-likelihood phylogenetic trees generated with the amino acid sequences of the catalytic subunits of the group 1b [NiFe]- and group 2d [NiFe]-hydrogenases (E), and cytochrome c₅₅₂ nitrite reductase (F) are presented to illustrate the phylogenetic positioning of these key functional genes in the two hydrogenotrophic DNRA isolates.

Figure 5

Analyses of Campylobacterota MAGs from the GTDB, containing at least one copy of each of the napA, nrfA, and hynB genes. (A) Maximum-likelihood phylogenetic tree of 74 Campylobacterota high-quality MAGs (CheckM completeness ≥ 90%, contamination ≤5%), constructed using IQ-TREE v2.3.6. Colored shades indicate family-level taxonomic affiliations, and colored dots next to the taxon descriptions represent the ecosystem classifications of their source metagenomes. (B) Geographical origins of the analyzed Campylobacterota MAGs, visualized on a world map. (C) Violin plots depicting the relative abundances (percentage of quality-trimmed raw reads from the source metagenome mapped onto each MAG) of Campylobacterota MAGs in their source metagenomes, categorized into three ecosystem types. Box plots within the violin plots display medians, interquartile ranges, and boundaries for non-outliers. (D, E) Maximum-likelihood phylogenetic trees constructed with in-silico-translated amino acid sequences of hydrogenotrophic DNRA marker genes nrfA (D) and hynB (E) identified from the Campylobacterota MAGs. Only non-truncated sequences are included. Heatmaps next to the trees indicate the amino acid identity between these sequences and those from A. butzleri hDNRA1 and Sulfurospirillum sp. hDNRA2.