Growth rate peaks at intermediate cell size in marine photosynthetic picoeukaryotes

We have performed an in situ test of Raven’s prediction that there is a reversal of the relationship between cell size and maximum achievable growth rate in unicellular algae at the low end of size classes. In a natural population of marine phytoplankton, including the smallest picoeukaryote known to date, and under both nutrient sufficiency and deficiency, we find a maximum in growth rate (4.8 and 3.3 divisions d21, respectively) in the 2–3-mm size class represented by coccoid Chlorella-like cells, with lower growth rates in both higher and lower size classes. This 2–3-mm size class is also the most robust under nutrient deficiency, reducing its growth rate by 14% only relative to nutrient-sufficient conditions, versus 50–60% for the lowermost and uppermost size classes, respectively. The relationship between cell size and maximum achievable growth rate in unicellular algae has been shown to be generally inverse (Chan 1978; Banse 1982; Tang 1995). This is due to both a thinner diffusion boundary layer and a greater surface to volume ratio which confer to small cells a greater capacity to acquire nutrients and efficiency in their use for growth and maintenance (Raven 1998). Such properties constitute a competitive advantage for picophytoplankton (,2 mm in size) in oligotrophic waters because they are able to use scarce resources more efficiently than larger cells (Chisholm 1992). The size of the prokaryotic and eukaryotic picophytoplankton are close to the minimum possible size estimated from the occurrence of nonscalable essential cell components, i.e., their genome, the plasmalemma, and other membranes (Raven 1986, 1994, 1998). Thus, the cell sizes of two major components of autotrophic picoplankton in oceans, the coccoid cyanobacteria of the genera Prochlorococcus (0.6 mm) and Synechococcus (0.9 mm), are significantly smaller than the arbitrary upper limit of picoplanktonic cell size. Among the photosynthetic picoeukaryote community, Ostreococcus tauri is the smallest free-living eukaryote know to date (0.95 mm, Courties et al. 1994). It has a minimal cellular organization with a single chloroplast and mitochondrion, and the smallest genome described among freeliving eukaryotic cells (Derelle et al. 2002). Its small cellular and genome sizes unveiled the ‘‘bare limits’’ of life as a free-living photosynthetic eukaryote (Derelle et al. 2006). The extensive size reduction of picoplanktonic cells and the necessary occupation of a larger fraction of the biomass by nonscalable components put such constraints on maximum achievable growth rate that it has been suggested (Raven 1986, 1994) that, below a certain size, growth decreases with decreasing size of unicellular algae both when resources are saturating and when they are limiting. The different level of the structural organization between prokaryotic and eukaryotic algae has important physiological consequences (Raven 1986; Weisse 1993). Compared with cyanobacteria, eukaryotes must devote a larger fraction of their internal metabolism to maintenance processes because of the presence of internal membranes. Consequently, at a given size, eukaryotes have a higher specific metabolism than prokaryotes. On the other hand, the genome and the thickness of membranes impose higher limits on the smallest size of eukaryotic than of prokaryotic cells (Raven 1986, 1994, 1998). Among the eukaryotic algae, the reversal of allometric relationship is supported by observations from laboratory cultures of chlorophytes highlighting a reduction of the growth rate for spherical cells whose diameter was less than 4 mm (Raven 1994). Despite its attractiveness, the prediction that nonscalable components can reduce the maximum specific growth rate of very small cells has never been tested with field data. In the natural environment, the first estimates of growth rates of autotrophic picoeukaryotes by size class (Dupuy et al. 2000) were in agreement with Raven’s theory, with an inversion of the allometric relationship for picoeukaryotes smaller than 2 mm and a maximum growth rate (3.3 or 4.7 divisions d21) for autotrophic eukaryotes belonging to the 2–4-mm equivalent sphere diameter size class. However, this was a single experiment and without rigorous statistical treatment. Here we test this reversal of allometric relationship on a more extensive data set in natural populations of marine phytoplankton less than 10 mm in cell diameter, including the smallest autotrophic eukaryote, O. tauri (Courties et al. 1994). To examine growth rates in various size classes of natural eukaryotic phytoplankton, we sampled in the marine Mediterranean Thau lagoon where O. tauri was initially described. Cell-based gross growth rates were determined by the dilution technique (Landry and Hassett 1982) under both nutrient-sufficient and nutrient-deficient conditions. Flow cytometry was used to measure cell densities in different size classes, which were determined by both standard beads and cultures of unicellular algae of known cell diameters determined by optical microscopy. Methods—Experimental procedure: Samples from the Thau lagoon (Southern France, 43u229N, 3u359E) were obtained on a monthly basis between May and September 1999, and from June to August 2001. Data from other