Variability in Phytoplankton Morphology and Macromolecular Composition With Nutrient Starvation and The Implications for Oceanic Elemental Stoichiometry

Trait based modeling efforts are an important tool for predicting the distribution of phytoplankton communities in the ocean and their interaction with elemental stoichiometry. The elemental stoichiometry of phytoplankton is based on their macromolecular composition. Many phytoplankton species accumulate C-rich storage products (carbohydrates and lipids) and reduce N and P-rich functional components (proteins and nucleic acids) upon N- or P-starvation. Reconciling global patterns in C:N:P stoichiometry and phytoplankton community structure and succession requires a better understanding of how phytoplankton macromolecular composition varies across taxa, size class, and growth conditions. We examined changes in cell size and composition from exponential growth to nitrogen starvation in four common phytoplankton species representing two size classes each of chlorophytes and diatoms. Variation in cell size, cell mass, and length of stationary growth phase appeared to be size dependent. The larger species of chlorophyte and diatom had a significant increase in cell mass and cell size with N-starvation and showed no significant change in cell density after starvation for 5-7 days. The smaller size species of both phyla showed no significant change in cell size or mass upon N-starvation and a consistent decline in cell density 1-2 days after peak densities were reached. All species had a similar significant increase in C quota, but changes in N quota and C:N were more variable and species-specific. We also present changes in macromolecular composition and C, N, and P-allocation due to N-starvation and their implications for elemental stoichiometry under natural conditions. These results are compared to field observations of C:N:P stoichiometry and phytoplankton community structure to examine the physiological plasticity that may underlie global oceanic C:N:P variability and demonstrate the importance of this plasticity in trait based models.