Global soil metagenomics reveals distribution and predominance of Deltaproteobacteria in nitrogen-fixing microbiome

Abstract

Background: Biological nitrogen fixation is a fundamental process sustaining all life on earth. While distribution and diversity of N₂-fixing soil microbes have been investigated by numerous PCR amplicon sequencing of nitrogenase genes, their comprehensive understanding has been hindered by lack of de facto standard protocols for amplicon surveys and possible PCR biases. Here, by fully leveraging the planetary collections of soil shotgun metagenomes along with recently expanded culture collections, we evaluated the global distribution and diversity of terrestrial diazotrophic microbiome.

Results: After the extensive analysis of 1,451 soil metagenomic samples, we revealed that the Anaeromyxobacteraceae and Geobacteraceae within Deltaproteobacteria are ubiquitous groups of diazotrophic microbiome in the soils with different geographic origins and land usage types, with particular predominance in anaerobic soils (paddy soils and sediments).

Conclusion: Our results indicate that Deltaproteobacteria is a core bacterial taxon in the potential soil nitrogen fixation population, especially in anaerobic environments, which encourages a careful consideration on deltaproteobacterial diazotrophs in understanding terrestrial nitrogen cycling. Video Abstract.

Keywords: metagenomics; microbial community; nitrogen fixation; soil microbiome.

Fig. 1

Currently known diversity of diazotrophs. a A genome-based phylogenetic tree consisting of potential diazotrophic bacteria [i.e., the genomes of which harbor all three core genes of nitrogenase (nifH, nifD, and nifK)], including the genomes of new isolates of Anaeromyxobacteraceae and Geobacteraceae (Table 1). The colors of branches and the band surrounding the tree denote the phyla and proteobacterial classes. Genomes of families Anaeromyxobacteraceae and Geobacteraceae, the foci of the present study, are highlighted with circled letters (A and G) and colored backgrounds (blue and pink, respectively). b Growth curves of the type strains of two type species within the family Geobacteraceae, namely Geomonas oryzae S43^T and Oryzomonas japonica Red96^T. The two isolates were grown on MFM medium with N₂ as the sole nitrogen source. Average and standard deviation of each time point (n = 3) are indicated. Some error bars are shorter than the symbol size

Fig. 2

Overview of metagenomic datasets used in this study and distribution of diazotrophs therein. a The sampling locations for each metagenomic datasets used in this study. Six types of environments are differentiated by the shapes and colors of symbols. b Filtering procedure of metagenomic datasets. The filtering criteria, as well as the aggregation of geographically similar samples, are explained in the panel. c Phylum-level prokaryotic community structure of the 321 metagenomic datasets estimated by 16S rRNA gene sequences. d The dominance of nitrogen-fixing population in each environment. For each of the 321 metagenomic datasets, the ratio of reads per kilobase of reference sequence per million sample reads (RPKM) of nitrogenase genes to the RPKM of ribosomal protein genes is displayed. The letters on the right side of the box indicates the statistical significance in RPKM ratio between different environmental categories (P < 0.05, Brunner–Munzel test with Bonferroni’s correction). First, second, and third quantiles are indicated by solid lines. The whiskers, if any, denote 1.5*[interquartile range] from first or third quartile

Fig. 3

Phylogenetic compositions of nitrogenase genes in the metagenomic datasets with at least 50 reads of nifD and nifK (n = 88 in total). a Upper panel: phylum- and proteobacterial class-level composition. Lower panel: breakdown of deltaproteobacterial composition at the family level. The category “Possibly Geobacteraceae or Anaeromyxobacteraceae” comprises deltaproteobacterial reads that were unannotated at the family level but received higher-level annotations consistent with family Geobacteraceae or family Anaeromyxobacteraceae. b Family-level distribution of nifD and nifK reads. The correspondence with the phylum- and proteobacterial class-level taxonomy is noted in parentheses: Delta, Deltaproteobacteria; Alpha, Alphaproteobacteria; Gamma, Gammaproteobacteria; Beta, Betaproteobacteria; Firm, Firmicutes; Actino, Actinobacteria; Cyano, Cyanobacteria. The area size (not the radius) of each plot is proportional to the relative abundance of each family within each dataset

Fig. 4

Contribution of nitrogenase gene sequences from newly isolated strains in the bioinformatic analyses of metagenomes. a A schematic of the analysis. NifD/K sequences annotated as Anaeromyxobacteraceae or Geobacteraceae in metagenomes were mapped onto already known sequences of NifD/K (right-upper) and those in our new isolates (right-bottom). Only the top hit for each query sequence (i.e., one from metagenomes) was considered. b Relative abundance of metagenome-derived NifD/K sequences that were most similar to already known sequences (yellow) and those from our new isolates (green), as well as those equally similar to the nitrogenase genes of already known genomes and our new isolates (dim green), are summarized. Only datasets with 10 or more sequences of NifD/K for each family are displayed