ALBERT

Description: Photoacclimation models are a prerequisite for accurate estimates of primary production in aquatic environments under typically variable light conditions. They generally start from empirical functions of the internal chlorophyll a (chl a) or nutrient quota (e.g. the Droop model). We propose that physiological variations in phytoplankton reflect phenotypic adaptation which maximizes the growth rate. Growth maximization has to account for indirect effects of the enhancement of carbon (C) acquisition by acclimation, primarily through concomitant changes in the intracellular nitrogen (N) budget. Our model expresses, for the first time, the indirect effect of alterations in N uptake on C assimilation by a parameter-free trade-off between the 2 uptake functions. The model explicitly prescribes optimal protein partitioning between N and C uptake and sub-partitioning into carboxylation (1,5-bisphosphate carboxylase/oxygenase, Rubisco) and light harvesting. Applications to various published experimental data for different phytoplankton species support the validity of the optimality hypothesis and point to different flexibility in the re-organization of chloroplasts between taxa as well as to different time-scales on which photoacclimation operates. Simulations of a batch culture with the haptophyte Isochrysis galbana show that a decoupling in pigment N:C from cellular N:C may explain observed lag phases in chl a:C regulation. For diatoms, seemingly stronger constraints in intra-cellular stoichiometry determine the photoacclimative response to variable light regimes, as simulated and reported for Skeletonema costatum. N and chl a quotas correlate well in nutrient-limited chemostats of Thalassiosira fluviatilis, but in part decouple under light limitation. In N limited growth, non-linearity in N:C as expressed by the Droop function results from a combination of a linear quota dependency, down-regulation of relative carboxylation capacity, and increasing N costs of chl a synthesis at elevated growth rates. Our optimality assumption that includes indirect feed-backs through the concept of protein partitioning generates an accurate model for adaptation in physiological traits.