Stoichio-Metagenomics of Ocean Waters: A Molecular Evolution Approach to Trace the Dynamics of Nitrogen Conservation in Natural Communities

Abstract

Nitrogen is crucially limiting in ocean surface waters, and its availability varies substantially with coastal regions typically richer in nutrients than open oceans. In a biological stoichiometry framework, a parsimonious strategy of nitrogen allocation predicts nitrogen content of proteins to be lower in communities adapted to open ocean than to coastal regions. To test this hypothesis we have directly interrogated marine microbial communities, using a series of metagenomics datasets with a broad geographical distribution from the Global Ocean Sampling Expedition. Analyzing over 20 million proteins, we document a ubiquitous signal of nitrogen conservation in open ocean communities, both in membrane and non-membrane proteins. Efficient nitrogen allocation is expected to specifically target proteins that are expressed at high rate in response to nitrogen starvation. Furthermore, in order to preserve protein functional efficiency, economic nitrogen allocation is predicted to target primarily the least functionally constrained regions of proteins. Contrasting the NtcA-induced pathway, typically up-regulated in response to nitrogen starvation, with the arginine anabolic pathway, which is instead up-regulated in response to nitrogen abundance, we show how both these predictions are fulfilled. Using evolutionary rates as an informative proxy of functional constraints, we show that variation in nitrogen allocation between open ocean and coastal communities is primarily localized in the least functionally constrained regions of the genes triggered by NtcA. As expected, such a pattern is not detectable in the genes involved in the arginine anabolic pathway. These results directly link environmental nitrogen availability to different adaptive strategies of genome evolution, and emphasize the relevance of the material costs of evolutionary change in natural ecosystems.

Keywords: marine microbial communities; material costs; molecular evolution; nitrogen limitation; stoichiogenomics.

FIGURE 1

Nitrogen content of transmembrane and non-transmembrane proteins from coastal and open ocean samples: a box plot visualization. The overall variation between open ocean and coastal samples is estimated as the relative difference in mean nitrogen content. Statistical significance is based on unpaired, one-sided approximative Wilcoxon–Mann–Whitney test.

FIGURE 2

Nitrogen content and transmembrane protein topology. Mean nitrogen content of cytoplasmic, transmembrane, and periplasmic/extracellular domains from coastal and open ocean samples. Solid bars represent transmembrane proteins with more than one transmembrane domain, while striped bars represent transmembrane proteins with only one transmembrane domain. The overall variation between open ocean and coastal samples is estimated as the relative difference in mean nitrogen content. Statistical significance is based on unpaired, one-sided approximative Wilcoxon–Mann–Whitney test. Standard errors of the mean are <0.001 for all samples (thus not visible in the present plot).

FIGURE 3

Nitrogen content allocation in NtcA and ArgF. Nitrogen content in fast and slow evolving residues (for details, see section “Materials and Methods”) in (A) NtcA and in (B) ArgF. N indicates the number of sequences, and n the number of sites in each dataset. Statistical significance is based on unpaired, one-sided Wilcoxon rank sum test.

FIGURE 4

Nitrogen content of proteins up- and down-regulated in response to nitrogen depletion. Scatter plot of the variation in mean nitrogen content in open ocean relative to coastal samples in fast and slow evolving residues for the different set of proteins (circles for the NtcA-induced cascade, and squares for the arginine biosynthetic pathway). Functional identification is based on TIGRFAM model (see section “Materials and Methods”). Statistical significance is reported for variable residues, and is based on unpaired, two-sided Wilcoxon rank sum test.