Carbon Dioxide Concentration Mechanisms in Natural Populations of Marine Diatoms: Insights From Tara Oceans

Marine diatoms, the most successful photoautotrophs in the ocean, efficiently sequester a significant part of atmospheric CO(2) to the ocean interior through their participation in the biological carbon pump. However, it is poorly understood how marine diatoms fix such a considerable amount of CO(2), which is vital information toward modeling their response to future CO(2) levels. The Tara Oceans expeditions generated molecular data coupled with in situ biogeochemical measurements across the main ocean regions, and thus provides a framework to compare diatom genetic and transcriptional flexibility under natural CO(2) variability. The current study investigates the interlink between the environmental variability of CO(2) and other physicochemical parameters with the gene and transcript copy numbers of five key enzymes of diatom CO(2) concentration mechanisms (CCMs): Rubisco activase and carbonic anhydrase (CA) as part of the physical pathway, together with phosphoenolpyruvate carboxylase, phosphoenolpyruvate carboxykinase, and malic enzyme as part of the potential C4 biochemical pathway. Toward this aim, we mined >200 metagenomes and >220 metatranscriptomes generated from samples of the surface layer of 66 globally distributed sampling sites and corresponding to the four main size fractions in which diatoms can be found: 0.8–5 μm, 5–20 μm, 20–180 μm, and 180–2,000 μm. Our analyses revealed that the transcripts for the enzymes of the putative C4 biochemical CCM did not in general display co-occurring profiles. The transcripts for CAs were the most abundant, with an order of magnitude higher values than the other enzymes, thus implying the importance of physical CCMs in diatom natural communities. Among the different classes of this enzyme, the most prevalent was the recently characterized iota class. Consequently, very little information is available from natural diatom assemblages about the distribution of this class. Biogeographic distributions for all the enzymes show different abundance hotspots according to the size fraction, pointing to the influence of cell size and aggregation in CCMs. Environmental correlations showed a complex pattern of responses to CO(2) levels, total phytoplankton biomass, temperature, and nutrient concentrations. In conclusion, we propose that biophysical CCMs are prevalent in natural diatom communities.