RESUMEN
How many subcellular targets of the beneficial silicon effect do exist in salt-stressed rice? Here, we investigate the effects of silicon on the different components of salt stress, i.e., osmotic stress, sodium, and chloride toxicity. These components are separated by multivariate analysis of 18 variables measured in rice seedlings (Oryza sativa L.). Multivariate analysis can dissect vectors and extract targets as principal components, given the regressions between all variables are known. Consequently, the exploration of 153 correlations and 306 regression models between all variables is essential, and regression parameters for variables of shoot (silicon, sodium, chloride, carotenoids, chlorophylls a and b, and relative growth rate) and variables of shoot and root (hydrogen peroxide, ascorbate peroxidase (APX), catalase (CAT), fresh weight, dry weight, root-to-shoot ratio) are determined. The regression models [log (y) = y0 + a × log (x)] are confirmed by variance analysis of global goodness of fits (p < 0.0001). Thereby, logarithmic transformation yields linearization for multivariate analysis by Pearson's correlation. Four principal components are extracted: two targets of osmotic stress, one target of sodium toxicity, and one target of chloride toxicity. Thereby, silicon improves salt tolerance by increasing APX and CAT activities and decreasing hydrogen peroxide, salt ion accumulation, photosynthetic pigment losses, and growth inhibition. Salt stress increases silicon uptake pointing to a physiological regulation of plant salt stress in the presence of silicon. This mechanism and its four components are promising targets for further agricultural application.