Environmental shaping of codon usage and functional adaptation across microbial communities

Abstract

Microbial communities represent the largest portion of the Earth's biomass. Metagenomics projects use high-throughput sequencing to survey these communities and shed light on genetic capabilities that enable microbes to inhabit every corner of the biosphere. Metagenome studies are generally based on (i) classifying and ranking functions of identified genes; and (ii) estimating the phyletic distribution of constituent microbial species. To understand microbial communities at the systems level, it is necessary to extend these studies beyond the species' boundaries and capture higher levels of metabolic complexity. We evaluated 11 metagenome samples and demonstrated that microbes inhabiting the same ecological niche share common preferences for synonymous codons, regardless of their phylogeny. By exploring concepts of translational optimization through codon usage adaptation, we demonstrated that community-wide bias in codon usage can be used as a prediction tool for lifestyle-specific genes across the entire microbial community, effectively considering microbial communities as meta-genomes. These findings set up a 'functional metagenomics' platform for the identification of genes relevant for adaptations of entire microbial communities to environments. Our results provide valuable arguments in defining the concept of microbial species through the context of their interactions within the community.

Figure 1.

Codon usage is metagenome-specific. Soil versus Santa Cruz Whale fall Santa Cruz Bone codon usage (CU) frequencies. (A) The distance (MILC, outlined in ‘Materials and Methods’ section) of each gene’s CU frequency to overall CU frequencies of two microbial communities. Genes [red in whale carcass (N = 33 422) and blue in Waseca soil (N = 88 696) metagenome] are predominantly closer to their respective metagenome of origin therefore forming two distinct groups (the distribution of log2 ratio of the two distances for each gene are shown in the inset). If the amino acid composition of metagenomes is kept constant and the codons randomly chosen, CU bias of each metagenome would be eliminated resulting in uniform distribution of CU distances and overlap of two colours, as shown in (B).

Figure 2.

Codon usage variability between same species in different metagenomes is larger than within a metagenome. ORFs from each identified species (using MEGAN) were compared against their originating metagenome (dark grey, total comparisons N = 2058) and against same-species ORFs in a different metagenome (light grey, total comparisons N = 1029 comparisons). ICC measures were calculated, representing how ‘close’ the CU profiles match, with ICC = 1 denoting the perfect match. The light grey distribution shows less variability and is shifted towards higher ICC values, denoting the closer overall match of species’ CU to their metagenome of origin.

Figure 3.

Environmental variability of codon usage. Variability of codon usage per COG Category in six strains of R. palustris and in 12 strains of P. acnes. The codon usage variability (calculated as median CU distance from the ribosomal set within an orthologous group to its centroid CU) for the strains of P. acnes (N = 15 436), living in consistent environmental conditions, is shifted to the left, i.e. shows smaller variation and higher bias, than for the R. palustris strains (N = 24 071) living in diverse environmental conditions.

Figure 4.

Enrichment of functions within highly expressed genes in metagenomes. Enrichment or depletion of functional annotations in the 3% genes with highest predicted expression (highest MELP measure) relative to the abundance of each COG supercategory in the whole metagenome for the OZ EBPR sludge (N = 29 754), Waseca farm soil (N = 88 696), acid mine biofilm (N = 79 257), Sargasso Sea (N = 688 539), US EBPR sludge (N = 20 175), Whale fall Santa Cruz microbial mat (N = 40 916), Whale fall Antarctic bone (N = 30 503), Whale fall Santa Cruz bone (N = 33 422), obese mouse gut (N = 4058), lean mouse gut (N = 4955) and human gut (N = 47 765), Santa Cruz whale fall bone (N = 33 422) and acid mine (N = 79 257). Metagenomes show different functional enrichment patterns that are consistent with environmental requirements (e.g. metabolite transport functions [E] in the Sargasso Sea or energy conversion [C] in the whale carcass metagenome). Non-significant enrichments are shown in white. Letters at the top represent COG supercategories: [J] Translation, ribosomal structure and biogenesis; [A] RNA processing and modification; [K] Transcription; [L] Replication, recombination and repair; [B] Chromatin structure and dynamics; [D] Cell cycle control, cell division, chromosome partitioning; [Y] Nuclear structure; [V] Defence mechanisms; [T] Signal transduction mechanisms; [M] Cell wall/membrane/envelope biogenesis; [N] Cell motility; [Z] Cytoskeleton; [W] Extracellular structures; [U] Intracellular trafficking, secretion and vesicular transport; [O] Posttranslational modification, protein turnover, chaperones; [C] Energy production and conversion; [G] Carbohydrate transport and metabolism; [E]Amino acid transport and metabolism; [F] Nucleotide transport and metabolism; [H] Coenzyme transport and metabolism; [I] Lipid transport and metabolism; [P] Inorganic ion transport and metabolism; [Q] Secondary metabolites biosynthesis, transport and catabolism; [R] General function prediction only; [S] Function unknown; [X] Uncharacterized.

Figure 5.

Correlation of metaproteomic and metagenomic data. Spearman’s correlation coefficient between metaproteomic and metagenomic data is shown for (A) the Sargasso Sea (N = 257)—each proteins’ number of spectra in the metaproteome versus the corresponding gene’s MELP value and (B) the human gut—the median NSAF value per protein COG versus the median MELP value per gene COG (N = 116) and (C) the Human gut median NSAF value per protein COG versus the median MELP value per Sargasso Sea gene COG. Spearman’s rho correlation is positive and greater for comparisons of metaproteomes with their own metagenomes then with foreign metagenomes.