Evidence for phylogenetically and catabolically diverse active diazotrophs in deep-sea sediment

Abstract

Diazotrophic microorganisms regulate marine productivity by alleviating nitrogen limitation. However, we know little about the identity and activity of diazotrophs in deep-sea sediments, a habitat covering nearly two-thirds of the planet. Here, we identify candidate diazotrophs from Pacific Ocean sediments collected at 2893 m water depth using ¹⁵N-DNA stable isotope probing and a novel pipeline for nifH sequence analysis. Together, these approaches detect an unexpectedly diverse assemblage of active diazotrophs, including members of the Acidobacteria, Firmicutes, Nitrospirae, Gammaproteobacteria, and Deltaproteobacteria. Deltaproteobacteria, predominately members of the Desulfobacterales and Desulfuromonadales, are the most abundant diazotrophs detected, and display the most microdiversity of associated nifH sequences. Some of the detected lineages, including those within the Acidobacteria, have not previously been shown to fix nitrogen. The diazotrophs appear catabolically diverse, with the potential for using oxygen, nitrogen, iron, sulfur, and carbon as terminal electron acceptors. Therefore, benthic diazotrophy may persist throughout a range of geochemical conditions and provide a stable source of fixed nitrogen over geologic timescales. Our results suggest that nitrogen-fixing communities in deep-sea sediments are phylogenetically and catabolically diverse, and open a new line of inquiry into the ecology and biogeochemical impacts of deep-sea microorganisms.

Fig. 1. Sampling location and previously measured N₂ fixation in the sediment core used in this study.

a Map of sampling site at Monterey Canyon, California, USA. Sampled location marked with a white circle. Contour lines show 500 m depth intervals. b ¹⁵N₂ assimilation in sediment microcosms measured using isotope ratio mass spectrometry (reproduced from Dekas et al. [12]). Microcosm headspace gas is indicated. Each circle represents a biological replicate, with bars indicating the average ¹⁵N₂ assimilation. Large, yellow circles indicate the replicate bottle used for ¹⁵N-SIP and nifH analyses.

Fig. 2. Taxonomic diversity of ¹⁵N-incorporators and nifH-containing taxa identified via ¹⁵N-SIP and DNA sequencing, respectively.

a Taxa listed by class for Proteobacteria and phylum for all other groups. b Deltaproteobacteria ¹⁵N-incorporators and nifH ASVs shown separately to accommodate for different vertical axis scales. The fraction of sequences identified as Desulfobacterales (dots), Desulfuromonadales (stripes), and other Deltaproteobacteria (no pattern) is additionally indicated. n.d. = not detected.

Fig. 3. Relative abundance of ¹⁵N-incorporators in the incubation in which they were identified, as well as the corresponding raw (unincubated) sediment.

Relative abundances shown for 0–3 cmbsf samples incubated with argon (a) or methane (b) and 9–12 cmbsf samples incubated with argon (c) or methane (d). Lineages that accounted for >0.1% of 16S rRNA gene reads in at least one sample are colored by class (for Proteobacteria only) or phylum. Low abundance lineages (i.e., <0.1% of reads in each sample) are grouped as ‘Other’. Internal bar lines show relative abundances of individual ASVs.

Fig. 4. Diversity of recovered nifH sequences.

Bold branches show nifH ASVs (n = 434) placed at their maximum likelihood positions on the reference nifH tree (n = 6040). Shaded regions indicate taxonomic identities of select clades for reference. Gray wedge shows collapsed branches of nifH homologs; the full tree including these sequences can be found in Supplementary Fig. S7. Reference sequences on terminal branches with lengths >1.0 were pruned for clarity (n = 34). Phyla within clade labeled “Multiple phyla”: Acidobacteria, Actinobacteria, Bacteroidetes, Chlorobi, Chloroflexi, Cyanobacteria, Elusimicrobia, Lentisphaerae, Nitrospirae, Proteobacteria, Spirochaetes, and Verrucomicrobia. Star and diamond insets display example nifH ASV placements at finer phylogenetic scales. ‘% nuc. identity’ in table shows range of percent nucleotide identities between nifH ASVs and closest references for each taxonomic group. *clades which include one or two nifH sequences from other phyla (clades without asterisks contain sequences from one phylum).

Fig. 5. Metabolic profiles of the ¹⁵N-incorporators’ closest characterized relatives [, –95].

Empty cells indicate metabolisms not tested or not found to support growth. Bold taxon names indicate ¹⁵N-incorporators within phyla or classes (Proteobacteria only) for which nifH sequences were also detected. Incubation column: Indicates the headspace (‘Ar’ or ‘CH₄′) and sediment horizon (0–3 cmbsf, shaded top left corner; 9–12 cmbsf, shaded bottom right corner; both, full shaded square) for which taxa were identified as ¹⁵N-incorporators. Identity column: ‘# ASVs’ indicates number of ASVs identified as ¹⁵N-incorporators that shared the same closest relative; ‘% id’ indicates percent sequence identity to closest relative across the amplified V4/V5 16S rRNA gene region (~370 bp). N₂ fixation column: ‘Fix N?’ indicates if closest relative previously shown to fix nitrogen (dark square); ‘nifH?’ indicates if relative contains a copy of nifH in its genome (dark square) or the relative’s lowest taxonomic rank that contains an organism with a copy of nifH (P phylum, C class, O Order, F Family, G Genus); ‘Recovered?’ indicates if one of the recovered nifH ASV’s closest nifH reference sequence was the copy from that relative (dark square) or the lowest common taxonomic rank shared with that relative. Tree rooted to 16S rRNA gene from Methanocaldococcus jannaschii (GenBank accession no. NR_113292). †: ¹⁵N-incorporator shared equal sequence identity to M. methanicus and M. superfactus. The 16S rRNA gene sequence from M. methanicus was used for tree construction and the metabolic profiles for both organisms are shown. ‡: nifH sequence not available at the time of analysis.