The hidden face of food phenolic composition.

Plant polyphenols are extremely diverse, due to the occurrence of several basic structures, numerous substitutions and, for some groups, of polymers (tannins). Plant polyphenol composition depends on the plant species and organ, with some molecules specific of particular plant families while others are ubiquitous. The polyphenol content is classically assessed by global analysis methods, which lack specificity and accuracy. These methods have been replaced with high performance liquid chromatography (HPLC), that enables accurate determination of individual molecules, provided they can be unambiguously identified and calibration curves can be established. However, HPLC analysis is restricted to simple compounds and difficult to apply in the case of complex extracts. Further difficulties encountered in the case of polymers include irreversible adsorption on the stationary phases. Proanthocyanidin analysis by HPLC after acid-catalysed depolymerisation in the presence of a nucleophile permits to overcome these problems and shows that proanthocyanidins predominate in the polyphenol composition of most plants. Large varietal differences in tannin quantitative and qualitative composition were observed for all plant species studied. Moreover, analysis is usually performed after extraction, which may lead to significant underestimation of the polyphenol content, since a large proportion is not extracted by usual solvents. This may be due to covalent binding to other plant constituents and to non-covalent adsorption on plant solids. Such matrix effect also influences the taste perception of polyphenols and their fate in the digestive tract, from in-mouth interactions with salivary proteins to their metabolism by colon microflora, with potential influence on bioavailability.

Litter mixture effects on tropical tree seedling growth--a greenhouse experiment.

Decomposing litter provides critical nutrients for plants, particularly in nutrient-poor ecosystems such as tropical forests. We hypothesised that decomposing litter improves the performance of a variety of tropical tree seedlings, and that this litter effect varies depending on the species of litter present in litter mixtures. We addressed these hypotheses with a large pot experiment manipulating a range of different litter mixtures of contrasting quality and using seedlings of four tree species from the Amazonian forest of French Guiana. In contrast to our initial hypothesis, decomposing litter had either neutral or negative impacts on seedling growth, despite strongly different growth rates, biomass allocation patterns and leaf and root traits among tree species. Tree species varied in their responses to litter additions, which were further modified by species identity of the added litter. Our data show litter species-specific effects on growth, biomass allocation and leaf and root traits of tropical tree seedlings. These results suggest that a net nutrient release from decomposing litter does not necessarily improve tree seedling growth, even under nutrient-limiting conditions. In conclusion, litter layer composition may affect seedling establishment and recruitment success beyond litter-derived plant nutrient availability, which may contribute to tree species composition and dynamics in the studied tropical forest.