Prevalent genome streamlining and latitudinal divergence of planktonic bacteria in the surface ocean

Abstract

Planktonic bacteria dominate surface ocean biomass and influence global biogeochemical processes, but remain poorly characterized owing to difficulties in cultivation. Using large-scale single cell genomics, we obtained insight into the genome content and biogeography of many bacterial lineages inhabiting the surface ocean. We found that, compared with existing cultures, natural bacterioplankton have smaller genomes, fewer gene duplications, and are depleted in guanine and cytosine, noncoding nucleotides, and genes encoding transcription, signal transduction, and noncytoplasmic proteins. These findings provide strong evidence that genome streamlining and oligotrophy are prevalent features among diverse, free-living bacterioplankton, whereas existing laboratory cultures consist primarily of copiotrophs. The apparent ubiquity of metabolic specialization and mixotrophy, as predicted from single cell genomes, also may contribute to the difficulty in bacterioplankton cultivation. Using metagenome fragment recruitment against single cell genomes, we show that the global distribution of surface ocean bacterioplankton correlates with temperature and latitude and is not limited by dispersal at the time scales required for nucleotide substitution to exceed the current operational definition of bacterial species. Single cell genomes with highly similar small subunit rRNA gene sequences exhibited significant genomic and biogeographic variability, highlighting challenges in the interpretation of individual gene surveys and metagenome assemblies in environmental microbiology. Our study demonstrates the utility of single cell genomics for gaining an improved understanding of the composition and dynamics of natural microbial assemblages.

Keywords: comparative genomics; marine microbiology; microbial ecology; microbial microevolution; operational taxonomic unit.

Fig. 1.

Genomic differences between SAGs and cultured bacterioplankton. PCA of general genome characteristics (A) and encoded amino acid frequency (B) of SAGs (solid colored symbols) and cultures of marine bacterioplankton (open circles) are shown. Cultures belonging to the same taxonomic group as SAGs have the same color. The two Actinobacteria SAGs were excluded from the genome characteristics analysis because they are Gram-positive bacteria, which have a different cell wall architecture, and were not included in the development of the trophic strategy model of Lauro et al. (17). (Insets) Variable vectors corresponding to each PCA plot. The following input variables were used for the genome characteristic analysis: abundance of genes encoding proteins localized in the cytoplasm; cytoplasmic membrane, periplasm, outer membrane, extracellular, and multiple locations; COG categories I, K, Q, T, and V; %NC, % noncoding DNA.

Fig. 2.

Genome size and paralogous gene frequency of SAGs and bacterioplankton cultures. The percentages of genes belonging to paralog families in SAGs (solid colored circles) and cultures (open circles) were estimated using BLASTCLUST. Cultures belonging to the same taxonomic group as SAGs have the same color. (Inset) Results of least squares linear regression between genome size and paralog frequency.

Fig. 3.

Global distribution of SAG-related microorganisms, as determined by metagenomic fragment recruitment. SAGs are listed along the y-axis, where color bars indicate source locations. Color bars along the x-axis indicate the surface ocean climate zone (SI Appendix, Table S5 provides locations). Metagenomes are in the same order as presented in SI Appendix, Fig. S10 along the top x-axis. The scale bar indicates the percentage of aligned metagenome sequences with alignments ≥200 bp long and ≥95% identity, normalized by the length of each SAG assembly. Percentages of aligned sequences from each metagenome to all SAGs, and from all metagenomes to individual SAGs, are presented as gray bars on the y-axis and x-axis, respectively. Med. Sea, Mediterranean Sea; NP, North Pacific; SA, South Atlantic; Roseo, Roseobacter; ARCTIC, ARCTIC96-BD19 cluster; Bacteroid, Bacteroidetes; MGA, Marine Group A; Verruco, Verrucomicrobia; Actino, Actinobacteria. A threshold of ≥95% nucleotide sequence identity of alignments ≥200 bp was applied for the BLASTN-based recruitment.

Fig. 4.

Capacity of available genomes to represent surface ocean bacterioplankton assemblages, as related to genetic divergence and geographic differences. (A) Fraction of marine metagenome reads recruited by SAGs, genomes of bacterioplankton cultures, and the combined set of genomes using a range of genomic DNA identity thresholds. (B) Ratio of recruitment in the SAGs’ native versus nonnative environment as a function of genomic DNA identity. Averages of values calculated for each metagenome (A) or genome (B) are provided. The scale of the SSU rRNA gene divergence was estimated using a Bacteria domain-wide correlation between SSU rRNA gene identity and the average nucleotide identity of available genomes (31). A threshold of ≥200-bp alignment was applied for the BLASTN-based recruitment.