Fatty Acids
Analysis of 41 dietary fatty acids was carried out using procedures
described by Chavarie et al. (2016b) (Table 1). Muscle samples were
freeze-dried and subsequently homogenized with a mortar and pestle.
Lipids were extracted overnight from 1 g of the homogenate in a 2:1
chloroform-methanol solution containing 0.01% BHT (v/v/w) at −20ºC
(Folch et al., 1957). After extraction, samples were filtered through
Whatman Grade 1 Qualitative filter paper and the filter paper/sample was
rinsed twice with 2 ml of the 2:1 chloroform:methanol. Sample extract
was collected in a test tube and 7 ml of 0.88 N NaCl solution was added
to encourage fatty acids to move into the organic (chloroform)
layer. The aqueous layer was discarded after which the chloroform was
dried with sodium sulfate prior to total lipid measurement. The
extracted lipid was used to prepare fatty acid methyl esters (FAME) by
transesterification with Hilditch reagent (0.5 N
H2SO4 in methanol) (Morrison et al.,
1964). Samples were heated for 1 h at 100 °C. Gas chromatographic
(GC) analysis was performed on an Agilent Technologies 7890N GC equipped
with a 30 m J&W DB-23 column (0.25 mm I.D; 0.15 μm film thickness). The
GC was coupled to a Flame Ionization Detector operating at 350oC. Hydrogen was used as carrier gas flowing at 1.25
ml/min for 14 minutes, and increased to 2.5 ml/min for 5 min. The
split/splitless injector was heated to 260 oC and run
in splitless mode. The oven program was as follows: 60oC for 0.66 min, increasing by 22.82oC/min to 165 oC with a 1.97 min
hold; increasing by 4.56 oC/min to 174oC and by 7.61 oC/min to 200oC with a six min hold. Peak areas were quantified
using Agilent Technologies ChemStation software. Fatty acids standards
were obtained from Supelco (37 component FAME mix) and Nuchek (54
component mix GLC-463).
All fatty acid values were converted to a mass percentage of the total
array, and were named according the IUPAC nomenclature as X:Y n-z, where
X is the number of carbon atoms in the fatty acids, Y is the number of
methylene-interrupted double bonds in the chain, and n-z denotes the
position of the last double bond relative to the methyl terminus
(Ronconi et al., 2010). Fatty acids suggested by Iverson et al. (2004)
as important dietary fatty acids, which transfer from prey to predator,
were used in our analyses. Fatty acids profiles (% of fatty acids) were
transformed using arcsin square-root function. Fatty acids groups were
identified using a multivariate analysis R Package (Team, 2017),
FactoMineR, using a hierarchical clustering analysis based on principal
components (Husson et al., 2012). To reduce the number of variables
used, A SIMPER (similarity percentage routine) was used to assess which
fatty acids were primarily responsible for observed differences among
groups (King et al., 1999). A principal components analysis (PCA) was
performed on the fatty acid profiles with PC-ORD version 6 (McCune et
al., 2011) among piscivorous groups to provide inferences about patterns
of resource use as defined by Chavarie et al. (2016b). Two-way
Permutational Multivariate Analysis of Variance (PERMANOVA), a
non-parametric analog of Multivariate analysis of variance (MANOVA), was
used to test for differences in fatty acid composition among the groups
identified by FactoMineR and among arms (i.e., to investigate any
spatial variations within the piscivorous ecotype). Two-way PERMANOVA
were performed in PAST 3 (Hammer et al., 2001) using 9999 permutations.
Pairwise post-hoc comparison (Bonferroni corrected) followed to
test differences among groups defined by FactoMineR and among arms.
Pairwise post-hoc comparison (Bonferroni corrected) also followed
to test differences among arms (i.e., spatial variation). Finally, the
fatty acid groups determined by FactoMineR were tested for differences
in depth of capture using one-way analysis of similarities (ANOSIM) with
9999 permutations using PAST 3.