A multi-tissue approach to assess the effects of lipid extraction on the isotopic composition of deep-sea fauna

Abstract

Stable isotope analysis (SIA) of carbon and nitrogen is now a common tool to investigate trophic relationships and food-web structure in aquatic ecosystems. However, species-specific and tissue-specific lipid content sometimes hinders the correct interpretation of δC and δN values because lipids are C-depleted with respect to proteins, and thus tissue lipid extraction is generally invoked. We assessed the effects of lipid extraction on δC and δN compositions and C:N ratios of muscle, liver (or hepatopancreas) and gonads of three common Mediterranean deep-sea species with different locomotory activity, buoyancy mechanisms and feeding modes: the benthic-feeder shrimp Aristeus antennatus, the nektobenthic-feeder shark Galeus melastomus and the pelagic-feeder fish Micromesistius poutassou to evaluate both the effect of lipid extraction on SIA data and the validity of δC lipid correction models. Results showed that the effect of lipid extraction is not unique but some common patterns can be identified. Lipid extraction resulted in increased δN values in the liver and the muscle and in decreased δN values in the gonads, in increased δC values in all the tissues of the three species, except in the hepatopancreas of A. antennatus, and in decreased C:N ratios. The magnitude of the changes was species- and tissue-specific. We assessed the validity of δC correction equations for lipid content in muscle and liver tissues available from literature in the species from this study. Such equations provided corrected δC values equivalent to those obtained through lipid extraction when applied to a species with similar characteristics (i.e., taxon, behavior, etc.) to those for which the equations were designed. Our results for muscle tissue showed that not one of the equations tested was valid for the deep-sea shark G. melastomus, and we propose a species-specific model with a fairly feasible model efficiency. Besides, tissue-specific equations for liver in G. melastomus provided non-significant differences between δC and δC values, but all model efficiencies were fairly low. Thus, lipid extraction trials on elasmobranch muscle and liver tissue to determine effects on δC and δN values on a species-by-species basis are recommended. Our research together with a comprehensive literature review on this topic, highlights that there is no accepted or mandated standard of treatment for lipids when using stable isotope analyses