Antioxidative defence and photoprotection in pine needles under field conditions. A multivariate approach to evaluate patterns of physiological responses at natural sites

Measurements of antioxidants and chloroplast pigments have been widely used as markers of stress and vitality of conifers in the field. However, due to the high variability of these data and the multiple environmental influences tress are exposed to, a quantification of physiological stress responses has only scarcely been possible. Physiological stress responses cannot be monitored by single stress markers, but are governed by many different interactions. The objective of this study was to evaluate patterns of biochemical stress markers in an objective and repeatable manner. For this purpose, a data set of 12 stress‐physiological variables (chloroplast pigments, epoxidation state of the xanthophyll cycle, α‐tocopherol, ascorbate and dehydroascorbate, GSH and GSSG) measured on field‐grown Pinus canariensis needles (n = 90) was subjected to explorative statistical techniques. Four principal components (PC), which explained 80% of the variance of the original data, were extracted by principal component analysis. According to stress‐physiological principles, complex responses were assigned to these PCs. Principal component 1 was positively affected by concentrations of a‐tocopherol and total ascorbate, and negatively by the proportion of epoxides in the xanthophyll cycle and by α‐carotene contents. Principal component 2 was composed of chlorophyll, lutein, neoxanthin, and β‐carotene contents, PC 3 contained information about GSH concentrations and the proportions of GSSG and dehydoascorbate; and PC 4 mainly comprised the pool size of the xanthophyll cycle. These components could be ascribed physiological principles such as antioxidative response in chloroplasts (PC 1), pigment content (PC 2), or antioxidant regeneration (PC 3). Via cluster analysis a classification of samples was made based on the patterns of their PC scores. The resulting clusters represented typical physiological response patterns: Cluster 1 was related to initial stages of oxidative damage, cluster 2 to antioxidative responses, whereas cluster 3 represented healthy trees. The spatial distribution of members of these clusters among field plots revealed that different response patterns could be observed at the same plot, a fact that might be ascribed to small scale differences and/or individually differing resistances, and something that is frequently overlooked in the field.