Microbial and Viral Genome and Proteome Nitrogen Demand Varies across Multiple Spatial Scales within a Marine Oxygen Minimum Zone

Abstract

Nutrient availability can significantly influence microbial genomic and proteomic streamlining, for example, by selecting for lower nitrogen to carbon ratios. Oligotrophic open ocean microbes have streamlined genomic nitrogen requirements relative to those of their counterparts in nutrient-rich coastal waters. However, steep gradients in nutrient availability occur at meter-level, and even micron-level, spatial scales. It is unclear whether such gradients also structure genomic and proteomic stoichiometry. Focusing on the eastern tropical North Pacific oxygen minimum zone (OMZ), we use comparative metagenomics to examine how nitrogen availability shapes microbial and viral genome properties along the vertical gradient across the OMZ and between two size fractions, distinguishing free-living microbes versus particle-associated microbes. We find a substantial increase in the nitrogen content of encoded proteins in particle-associated over free-living bacteria and archaea across nitrogen availability regimes over depth. Within each size fraction, we find that bacterial and viral genomic nitrogen tends to increase with increasing nitrate concentrations with depth. In contrast to cellular genes, the nitrogen content of virus proteins does not differ between size fractions. We identified arginine as a key amino acid in the modulation of the C:N ratios of core genes for bacteria, archaea, and viruses. Functional analysis reveals that particle-associated bacterial metagenomes are enriched for genes that are involved in arginine metabolism and organic nitrogen compound catabolism. Our results are consistent with nitrogen streamlining in both cellular and viral genomes on spatial scales of meters to microns. These effects are similar in magnitude to those previously reported across scales of thousands of kilometers. IMPORTANCE The genomes of marine microbes can be shaped by nutrient cycles, with ocean-scale gradients in nitrogen availability being known to influence microbial amino acid usage. It is unclear, however, how genomic properties are shaped by nutrient changes over much smaller spatial scales, for example, along the vertical transition into oxygen minimum zones (OMZs) or from the exterior to the interior of detrital particles. Here, we measure protein nitrogen usage by marine bacteria, archaea, and viruses by using metagenomes from the nitracline of the eastern tropical North Pacific OMZ, including both particle-associated and nonassociated biomass. Our results show higher genomic and proteomic nitrogen content in particle-associated microbes and at depths with higher nitrogen availability for cellular and viral genomes. This discovery suggests that stoichiometry influences microbial and viral evolution across multiple scales, including the micrometer to millimeter scale associated with particle-associated versus free-living lifestyles.

Keywords: comparative metagenomics; marine particles; microbial genome evolution; stoichiometry; viral genome evolution.

FIG 1

Environmental contextual data and sampling locations for this study. (A) Map showing sampling station locations for the 2013 and 2014 field collection efforts. (B) Representative oxygen saturation profile from Station 6 in the upper 300 m. (C) Dissolved nitrate and nitrite profile for Station 6, measured in 2013. (D) Average GC content of the metagenomic reads from all of the metagenomic samples that were taken from the upper 300 m, representing a >1.6 μm particle-associated size fraction and a planktonicd size fraction of >0.2 μm.

FIG 2

Stoichiogenomic profiles identify the size fraction and depth structure across domains. Redundancy analysis (RDA) ordination was constrained by the sequencing depth, and the first two axes uncorrelated with sequencing depth are presented. The panels are separated by the putative taxonomic assignment of the genes that were used to construct the ordinations: (A) bacterial genes, (B) archaeal genes, (C) viral genes. Point color indicates the sample depth, where yellow is above the average depth of the oxic-anoxic interface (80 m) and blue is below the interface. Point shape indicates particle-associated (circle) versus free-living (triangle) samples.

FIG 3

Microbial and viral genes exhibit stoichiogenomic structure across size fraction and depth. Each point represents a metagenome, where the value is the coverage-weighted average GC content (A), the nitrogen content of the amino acid side chains (B), and the total side chain N:C ratio (C) of all archaeal, bacterial, or viral annotated genes in that metagenome. Boxplots indicate the median, 25th, and 75th percentiles of the metagenomes within a size fraction. (D) The distribution of estimated average bacterial genome length between size fractions, as determined by the relative abundances of the single copy core genes in each metagenome.

FIG 4

Domains have unique amino acid usage patterns to structure protein nitrogen content. The regularized amino acid regression parameters on the average N:C ratios of widely conserved functional proteins across samples are presented. Positive values indicate increases in the protein N:C ratio, and negative values indicate decreases. Amino acids are colored by the numbers of N atoms in their side chains. Note the differences in scale between domains. (A) Results from bacterial rpoZ sequences. (B) Results from archaeal ftsZ sequences. (C) Results from viral Gp23 sequences.

FIG 5

Viral Gp23 amino acid composition varies with depth. The x axis represents the coverage-weighted average frequency of each amino acid in a Gp23 sequence for each sample. Amino acids are labeled as polar (hydrophilic) (A to C), neutral (D), or hydrophobic (E and F). The smoothed lines, which are used to demonstrate overall trends, were fitted via general additive model smoothing, with shaded regions representing the 95% confidence intervals of the smoothing.

FIG 6

Bacterial functional gene content differences between size fractions. KEGG orthologues are grouped by metabolic pathways and are identified by KEGG gene nomenclature (y axis). The log ratios of the relative abundance of each gene to sdhC, a core functional gene that is present across all of the samples, are plotted on the x axis (see Materials and Methods). Point color indicates the size fraction.