On the measure of flower color: Examples from the Brassicaceae

Background/Question/Methods

Flower color has fascinated biologists for centuries, yet most studies have focused on polymorphisms with discrete color variants. Most species in nature with flower color polymorphisms exhibit a continuous distribution of pigmentation. In such cases, accurate quantification of flower color is essential to understanding the ecological and evolutionary roles of these pigments. Although many measures of flower color have been proposed, no single approach is used consistently and the relationships among different flower color quantification methods remain unknown. Focusing on the most common floral pigments, anthocyanins, we examined a variety of methodological approaches to quantifying flower color in three diverse mustard lineages – Parrya nudicaulis, Raphanus sativus, and the Streptanthus glandulosus complex. For each of these three taxa, we sampled individuals representing a range of pigmentation and measured UV-Vis reflectance spectra, anthocyanin concentrations, and HPLC-DAD analysis of the flavonoid intermediates. For Parrya nudicaulis, we extended the comparison to include gene expression of a candidate locus (chalcone synthase; CHS). The relationships among each flower color quantification method were determined with regression and path analyses.

Results/Conclusions

UV-Vis spectral chroma measurements showed the strongest correlation with the concentration of anthocyanins. Within the numerous measures of chroma, the ratio of reflectance at 550 nm relative to background reflectance (450nm & 700nm) exhibited the most significant correlation with anthocyanin concentration. We found significant correlations between the concentration of most flavonoid intermediates and anthocyanins – the strongest relationship was with the predominant flavonol, quercetin. For P. nudicaulis, gene expression of CHS was correlated with anthocyanins and chroma, but not flavonol concentration. Path analysis confirmed the expected relationships between CHS expression, anthocyanin concentration and chroma, but only detected a weak causal effect of CHS expression on flavonol concentration. Based on the continuous flower color variation quantified in these three species of Brassicaceae, a chroma measurement that is informed by the underlying biochemistry of the anthocyanins is the best predictor of color.