The quantification of pollutant metabolites in fish bile is an efficient approach to xenobiotic pollution monitoring in freshwaters since these measurements directly address exposure. Fluorescence excitation-emission matrix spectroscopy (EEMS) has demonstrated to be a highly specific and cost-effective technique for polycyclic aromatic hydrocarbon (PAH) and PAH-metabolite identification and quantification. EEMS ability to quantify these compounds strongly depends on the intensity and variability of the bile baseline fluorescence (BBF). We found large differences in BBF among Aequidens metae (AME) individuals and of these with Piaractus orinoquensis (PIO). Moreover, BBF was large enough that solvent dilutions of over 1:400 were needed to avoid inner filter effects. We used parallel factor analysis (PARAFAC) to model the intra- and inter-species BBF variability. PARAFAC successfully decomposed the EEMS set into three fluorophores present in all samples, although in concentrations spreading over ~ 3 orders of magnitude. One of the factors was identified as tryptophan. Tryptophan and Factor 2 were covariant and much more abundant in AME than in PIO, while Factor 3 was ~ 6 times more abundant in PIO than in AME. Also, tryptophan was ~ 10x more abundant in AME specimens immediately caught in rivers than in their laboratory-adapted peers. The PARAFAC decomposition effectiveness was confirmed by the positive proportionality of scores to dilution ratios. A large inner filter indicates that Factor 2 is as strong a light absorber as tryptophan. Our results stress the need to include bile matrix variable components for the detection and quantification of pollutant metabolites using PARAFAC.
Read full abstract