Nitrogenase gene amplicons from global marine surface waters are dominated by genes of non-cyanobacteria

Abstract

Cyanobacteria are thought to be the main N(2)-fixing organisms (diazotrophs) in marine pelagic waters, but recent molecular analyses indicate that non-cyanobacterial diazotrophs are also present and active. Existing data are, however, restricted geographically and by limited sequencing depths. Our analysis of 79,090 nitrogenase (nifH) PCR amplicons encoding 7,468 unique proteins from surface samples (ten DNA samples and two RNA samples) collected at ten marine locations world-wide provides the first in-depth survey of a functional bacterial gene and yield insights into the composition and diversity of the nifH gene pool in marine waters. Great divergence in nifH composition was observed between sites. Cyanobacteria-like genes were most frequent among amplicons from the warmest waters, but overall the data set was dominated by nifH sequences most closely related to non-cyanobacteria. Clusters related to Alpha-, Beta-, Gamma-, and Delta-Proteobacteria were most common and showed distinct geographic distributions. Sequences related to anaerobic bacteria (nifH Cluster III) were generally rare, but preponderant in cold waters, especially in the Arctic. Although the two transcript samples were dominated by unicellular cyanobacteria, 42% of the identified non-cyanobacterial nifH clusters from the corresponding DNA samples were also detected in cDNA. The study indicates that non-cyanobacteria account for a substantial part of the nifH gene pool in marine surface waters and that these genes are at least occasionally expressed. The contribution of non-cyanobacterial diazotrophs to the global N(2) fixation budget cannot be inferred from sequence data alone, but the prevalence of non-cyanobacterial nifH genes and transcripts suggest that these bacteria are ecologically significant.

Figure 1. World map of sampling locations showing the distribution of nifH Clusters.

Pie charts display the distribution of nifH Clusters within each sample. Clusters containing <10 sequences (shown in grey) were not phylogenetically designated. Note that Cluster I is split into Proteobacteria and Cyanobacteria, but that Cluster III also contains some Proteobacteria. For the Sargasso Sea samples, which were prefiltered (10 µm) to avoid filamentous cyanobacteria, the pie charts for DNA and cDNA samples are shown in the bottom left corner.

Figure 2. Phylogeny and relative composition of the sequenced nifH assemblages.

Neighbor-joining phylogenetic tree of 92% similarity clustered nifH sequences (79 090) and nearest relatives from ten sampling locations. The relative number of sequences in each cluster is shown in a heatmap with the median value of the dataset indicated in dark blue and the highest value in pink. White indicates that there are no sequences from the cluster present in the sample. Bootstrap values (500 replicates) >50% are indicated with grey circles proportional to the size of the bootstrap value. Accession numbers of reference sequences, sequence codes for each cluster and the absolute number of sequences for each sample are shown in Figure S4.

Figure 3. Principal Coordinate Analysis plot depicting nifH assemblage similarity between samples.

The plot illustrates the compositional and phylogenetic similarity between samples based on weighted UniFrac. Temperature and chlorophyll (chl a) are indicated by color and size of the dots for each sample in the plot.