Metabolomic profiles associated with exposure to per- and polyfluoroalkyl substances (PFASs) in aquatic environments

Matthew D. Taylor *abc, Jennifer Bräunig b, Jochen F. Mueller b, Marcus Crompton c, R. Hugh Dunstan c and Sandra Nilsson b
aPort Stephens Fisheries Institute, New South Wales Department of Primary Industries, Locked Bag 1, Nelson Bay, NSW 2315, Australia. E-mail:
bThe University of Queensland, Queensland Alliance for Environmental Health Sciences, 20 Cornwall Street, Woolloongabba, QLD 4102, Australia
cSchool of Environmental and Life Sciences, University of Newcastle, New South Wales 2308, Australia

Received 27th August 2019 , Accepted 17th September 2019

First published on 19th September 2019

Per- and polyfluoroalkyl substances (PFASs) are frequently detected in aquatic environments. Longer chained perfluoroalkyl acids (PFAAs), in particular, have been found to bioaccumulate in a broad range of aquatic biota. PFAAs have a physiochemical similarity to naturally occurring fatty acids and could potentially disrupt metabolic processes, however, there has been limited study in this area, especially in aquatic species. In this study, the associations between PFAAs and metabolite profiles were investigated in crustaceans. Eastern School Prawn (Metapenaeus macleayi) were obtained from three different locations (n = 15 per location) with similar environmental conditions but different levels of PFAA contamination. The concentrations of PFAAs, fatty acids and amino acids were analysed and differences in PFAA and metabolite profiles were evaluated. Different PFAA profiles were mirrored by significant differences in the composition of both fatty acid and amino acid profiles, indicating a potential association between PFAA concentration and the composition of metabolites in prawns. These results highlight a need for further research investigating the impacts of PFAA exposure, with the current study providing a foundation for further investigation of the relationship between PFAA bioaccumulation and organism metabolism.

Environmental significance

Per- and polyfluoroalkyl substances (PFASs) are emerging contaminants which are increasingly detected in aquatic environments. Potential effects of PFAS exposure are not yet fully understood for aquatic biota, and expanding this body of knowledge is essential to defining the full range of impacts of PFAS contamination. Eastern School Prawn was used as a study organism to investigate associations between PFAS exposure and changes in the metabolome. The results show an association between PFAA (perfluoroalkyl acid) concentration, and differences in metabolite composition in prawns. This could be indicative of biochemical effects of PFAS contamination, and highlights the need for further research investigating the broad range of potential impacts of exposure in aquatic organisms.


Per- and polyfluoroalkyl substances (PFASs) are a group of manmade fluorinated organic compounds, which are widely applied in industrial processes and products such as stain repellents, coatings and aqueous film-forming foams.1 Both manufacturing of these products, and the extensive usage of products containing PFASs, has led to substantial emissions of these contaminants into the environment. Due to high persistence and bioaccumulation potential of PFASs, but limited knowledge of their impacts, this group of compounds is of emerging concern. Consequently, the occurrence and toxicity of PFASs are currently the subject of intensive research.2–4

The introduction of PFASs into aquatic ecosystems results in exposure of aquatic animals to these compounds.2 However, the potential eco-toxicological effects of PFAS exposure are not yet fully understood. PFASs, especially longer chained PFAAs (perfluoroalkyl acids) have been reported to bioaccumulate in a broad range of aquatic biota, including fish and crustaceans.5,6 PFAAs have charged functional groups (e.g. carboxylate or sulfonate) attached to a fluorinated carbon chain, with a structure which is similar to fatty acids. They may therefore have similar behaviour to some natural fatty acids, and possibly compete with endogenous fatty acids to bind to proteins such as transporters and receptors, thus disrupting metabolic processes.7,8

In aquatic species, bioaccumulation of PFASs is highly variable among individuals, species, and environments.5,9,10 In estuarine systems, previous work has also indicated that bioaccumulation of PFAAs is often higher in crustacean species.9,11 While the proximal factors contributing to this are unknown, it is likely that aspects of crustacean physiology impact toxicokinetics (principally bioaccumulation, tissue distribution and elimination), which in turn could potentially affect sensitivity to contaminants.12 Consequently, crustaceans present useful model taxa to study the relationship between contaminants such as PFAAs, and physiological responses.

Metabolomics is the study of naturally occurring organic metabolites within the tissues of organisms. Ecometabolomics is the application of metabolomics to characterise the interactions between organisms and their environment,13 both as a direct measure of the response of an organism to environmental change, as well as responses to contaminant exposure within these environments.14–16 The approach measures concentrations of different molecules that may vary as a result of alterations to biochemical and cellular processes, and thus provides data which reflects the biological function or regulation within an organism.17 While this can identify metabolic pathways potentially affected by contaminants,18 it is also useful in generation of hypotheses surrounding metabolic changes for further investigation under controlled conditions.13 Fatty acids and amino acids are commonly measured in the study of ecometabolomics.18 Fatty acids and lipids have multiple functions in cellular metabolism and are integral components of cell membranes. Changes in cellular composition of fatty acids will therefore provide an indication of the dynamic physiology of these structures in response to stress. Amino acids have less commonly been examined in ecometabolomic studies of aquatic animals (see summary in Sardens et al.18), but there is some evidence for substantial changes in the amino acid pool for aquatic species exposed to contaminants.19 Fatty acids and amino acids can provide evidence of changes in a range of metabolic processes in estuarine species, which may occur in response to potential stressors such as PFASs.19,20

In recent years, metabolomics has seen application in assessment of the toxicological and ecotoxicological effects of PFAS contamination, including humans,21,22 earthworms,23 fish16,24,25 and planktonic crustaceans.26–29 Across these organisms, studies have shown PFASs to impact several metabolites, including a suite of fatty acid22 and amino acid compounds.23,29 Work with crustaceans has been limited to laboratory experiments with Daphnia magna, a freshwater planktonic species, but showed concentration dependent relationships between perfluorooctane sulfonate (PFOS) and amino acids which suggested protein degradation. Noting limited work on estuarine crustaceans, and the lack of any study examining associations in free-ranging aquatic species, we sought to examine associations between environmentally relevant concentrations of PFASs (mainly PFAAs) and potential alterations in fatty acid and amino acid compositions as a first step in examining metabolic implications of contamination with PFAAs in these taxa. We used Eastern School Prawn (Metapenaeus macleayi, hereafter referred to as School Prawn)30 as a study organism, given previous evidence of PFAA contamination in this species,5 marked temporal and spatial variability in PFAA concentrations,31,32 and recent work on the metabolome of this species in the wild.33 We measured corresponding PFAA, fatty acid and amino acid concentrations in School Prawn muscle tissue, and evaluated associations between dominant PFAAs and metabolites across multiple locations with similar environmental conditions but differing levels of PFAA contamination.


Description of study sites

The Hunter River estuary is a mature wave-dominated barrier estuary on the mid-north coast of New South Wales, Australia (Fig. 1), and supports an important fishery particularly for School Prawn.34 The catchment is primarily agricultural and forested, however the lower estuary is highly industrialised and urbanised around the City of Newcastle. The Hunter River estuary is impacted by a distinct aqueous film-forming-foam (AFFF)-derived point source of PFAS contamination flowing into Fullerton Cove from the nearby Royal Australian Airforce base at Williamtown.5 The estuary may also be impacted by other industrial and waste-water treatment sources around the catchment. Previous work has identified bioaccumulation of PFASs in all exploited species sampled from this estuary, but with concentrations generally decreasing with increasing distance to the main point source.31 Wallis Lake estuary is also a wave dominated estuary on the mid-north coast of New South Wales with a primarily agricultural/forested catchment, and exploited species captured from this estuary have previously been shown to have either non-detectable or low-levels of PFAA bioaccumulation.6,31 Three locations of similar water physico-chemistry (see Table 1, as well as recent ref. 33–35) were examined across these two estuaries (Fig. 1): (1) Fullerton Cove in the Hunter River estuary, highly impacted by an AFFF-derived PFAS point source; (2) Tomago in the Hunter River estuary, moderately impacted by an AFFF-derived PFAS point source and potentially other industrial sources; and (3) Wallis Lake estuary, low-level impact likely from more diffuse PFAS sources. The three locations were approximately equidistant from the mouths of the respective estuaries. In the Hunter River, salinities at the two locations generally oscillate between 30-35 under non-flood conditions (see Tyler et al.35), and the salinity measurements collected here (Table 1) matched previous salinity measurements at these locations in both estuaries.33,34
image file: c9em00394k-f1.tif
Fig. 1 Map showing locations (blue circles) sampled in each estuary, with estuary locations on the eastern Australian seaboard indicated by arrows. The AFFF-derived PFAS point-source in the Hunter River estuary is indicated with a red circle.
Table 1 Physicochemical conditions measured at the time of sampling School Prawn in each location
Hunter River estuary Wallis Lake estuary
Fullerton Cove Tomago
Salinity 34.4 30.7 35.4
Temperature (°C) 23.4 24.5 23.3
Dissolved oxygen (mg L−1) 6.4 6.0 5.6

Sample collection

School Prawn samples were collected from the Hunter River estuary using a chartered commercial prawn trawling vessel fishing with a Florida Flyer otter trawl. Following retrieval of the net, prawns (n = 15 per location) were immediately sorted from the catch and euthanized on ice, followed by rapid progression of the following processing steps. Each prawn was measured for carapace length (CL, cm), had their head removed, was de-shelled and de-veined, and had their muscle tissue divided down the centreline into two separate vials. Following division of the muscle tissues, sample vials were immediately immersed in ice to maintain sample integrity until they could be placed in storage at −80 °C (the trawl vessel proceeded immediately back to port, followed by a 15 minute drive to the laboratory). School Prawn samples from the Wallis Lake estuary were similarly handled with the exception that prawns were captured using a prawn seine net. Prawns arrived at the laboratory within ∼30–120 minutes of initial processing (the longest time was for Wallis Lake). Physico-chemical parameters were measured at the time of sampling using a Horiba U-52 Multiparameter Water Quality Meter (Pasadena, TX, USA).

Extraction and analysis of PFAAs

PFASs were extracted from 1 g w/w of homogenized muscle tissue, according to a method described by Baduel et al.36 Briefly, all samples were spiked with 0.2 ppm mass-labelled PFAS standard mix, digested by 0.4 mL 200 mM sodium hydroxide in methanol and extracted with 4 mL acetonitrile. The supernatants were cleaned-up by liquid–liquid extraction using n-hexane, followed by pushing the samples through carbon cartridges (Bond Elut, 100 mg, Agilent Technologies). Prior to instrumental analysis, samples were spiked with 0.2 ppm instrument standard. PFAS analysis was performed using high performance liquid chromatography (HPLC, Nexera HPLC, Shimadzu Corp, Kyoto Japan) coupled to a tandem mass spectrometer (SCIEX Triple Quad 6500+, Concord, Ontario, Canada) as described in Taylor et al.6 Target analytes included perfluorobutanoic acid (PFBA), perfluoropentanoic acid (PFPeA), perfluorohexanoic acid (PFHxA), perfluoroheptanoic acid (PFHpA), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnDA), perfluorododecanoic acid (PFDoDA), perfluorotridecanoic acid (PFTriDA), perfluorotetradecanoic acid (PFTreDA), perfluorobutanesulfonic acid (PFBS), perfluorohexanesulfonic acid (PFHxS), perfluorooctanesulfonic acid (PFOS), perfluorodecanesulfonic acid (PFDS), and the fluorotelomer sulfonates 8:2 FTS, 6:2 FTS, and 4:2 FTS. Limits of detections (LODs) were set to three times the standard deviation of the concentration of the lowest standard after 10 injections of the standard, with a signal to noise superior to 3. Limits of Reporting (LORs) were set 10 times this standard deviation and ranged from 0.02–0.38 μg kg−1 for different PFASs (data presented in ESI). Details of the chemicals and reagents that were used, as well as further details on extraction, analysis and quality assurance are presented in the ESI.

Fatty acid analysis of prawn muscle

Fatty acids were extracted from 20 mg (wet weight) lyophilised prawn muscle using a methanol/toluene solution (4[thin space (1/6-em)]:[thin space (1/6-em)]1 v/v) to form the fatty acid methyl ester (FAME) products in the upper toluene phase as described previously by Taylor et al.33 Free cholesterol which was liberated during the process does not form methyl esters, and thus 20 μL of BSTFA was then added to form the trimethylsilyl derivative of cholesterol for analysis by GC-MS.37 Quality assurance data are presented in the ESI.

Amino acid analysis of prawn muscle

Five milligrams of wet prawn muscle tissue was lyophilised and extracted with 500 μL acetonitrile/methanol/water (40/40/20) together with norvaline as the internal standard, for 20 minutes at −20 °C.38 The extract was centrifuged at 16[thin space (1/6-em)]000 × g and the supernatant was transferred to a clean borosilicate glass tube. The pellet was re-extracted twice with 200 μL of acetonitrile/methanol/water and the supernatants were combined. The combined extracted supernatants were dried in centrifugal vacuum concentrator and redissolved in 200 μL of 0.1 M HCl. The samples were then processed using the EZ:Faast™ derivatisation kit for amino acid analyses, and analysed by gas chromatograph coupled with a flame ionisation detector.39

Data handling and statistical analyses

All analyses were conducted using R v. For PFAAs, values <LOR were specified as a value equivalent to 0.5 × LOR for data analysis. Total PFAA, total fatty acid, and total amino acid in School Prawn collected at each location were compared using a single-factor analysis-of-variance (ANOVA). To evaluate differences in PFAA and metabolite profiles among locations, and molecules important for driving these differences, each dataset was centered and scaled using the preprocess function in the caret package (which includes a Box–Cox transformation41) and analysed using a discriminant function analysis (lda function in the MASS package42). Differences in PFAA and metabolite profiles among locations were also tested for significance using a single-factor permutational analysis-of-variance (using function adonis), with significant differences further evaluated by pairwise analyses among factor levels. Direct relationships between the main variables driving distinct groupings in the discriminant function analysis for each linear discriminant were evaluated using simple linear regression.


Spatial variation in PFAAs

Forty-five School Prawn of carapace length 2.00 ± 0.04 cm (mean ± SE) were analysed for PFAAs, fatty acids and amino acids. PFAAs were detected in all School Prawn samples collected from Hunter River sites at Fullerton Cove and Tomago, and in 66% of School Prawn sampled from Wallis Lake (all be it at low concentrations). PFOS-linear, PFOS-branched and PFHxS were the major contaminants measured in the Fullerton Cove and Tomago sites as shown in Table 2. There were significant differences in total PFAA concentrations among the three locations (F1,42 = 201.1, P ≪ 0.001), with pairwise comparisons (P < 0.001) confirming a significant decrease in total PFAA concentration from Fullerton Cove, to Tomago, to Wallis Lake (Fig. 2). Discriminant function analysis showed good separation among samples from different locations based on their PFAA profile (Fig. 3). Fullerton Cove and Tomago samples were separated from Wallis Lake samples along LD1, but Fullerton Cove and Tomago samples also grouped differently to each other, separated primarily along LD2. Permutational ANOVA confirmed a significant difference among locations based on their PFAA profile (pseudo-F1,42 = 18.8, P = 0.005), and pairwise comparisons indicated all locations significantly differed from all other locations (Padj < 0.003). Fullerton Cove and Tomago samples were characterised by higher PFOS-linear and PFOA concentrations relative to Wallis Lake. Hunter River locations were further separated along LD2 by elevated PFHxS concentrations in Fullerton Cove and elevated PFDA and PFDoDA concentrations in Tomago samples (Fig. 3, Table 2).
Table 2 Mean PFAA concentrations (μg kg−1, SE in brackets, n = 15) in muscle tissue of School Prawn collected at three locations of differing levels of PFAA contamination. The molecules with the five largest parameter coefficients for the two LD axes shown in Fig. 3 are indicated as superscripts (i.e. LD1, LD2) in the notation column
Notation Hunter River estuary Wallis Lake estuary
Fullerton Cove Tomago
PFOS-linearLD1, LD2 8.13 (0.14) 5.21 (0.31) 0.06 (<0.01)
PFOS-branchedLD2 2.73 (0.07) 2.47 (0.28) 0.03 (<0.01)
PFHxSLD1, LD2 2.82 (0.15) 0.73 (0.06) 0.03 (<0.01)
PFOALD1 0.31 (0.01) 0.14 (0.01) 0.03 (<0.01)
PFDALD1, LD2 0.10 (<0.01) 0.13 (0.01) 0.01 (<0.01)
PFBA 0.02 (<0.01) 0.01 (<0.01) 0.01 (<0.01)
PFPeA 0.01 (<0.01) 0.03 (0.01) 0.01 (<0.01)
PFHpA 0.02 (<0.01) 0.01 (<0.01) 0.01 (<0.01)
PFNALD1 0.14 (<0.01) 0.10 (<0.01) 0.03 (<0.01)
PFUnDA 0.02 (<0.01) 0.04 (<0.01) 0.02 (<0.01)
PFDoDALD2 0.06 (<0.01) 0.07 (<0.01) 0.06 (<0.01)

image file: c9em00394k-f2.tif
Fig. 2 Mean ± SE total fatty acid (white bars, primary y-axis), total amino acid (black bars, black secondary y-axis) and total PFAA concentrations (grey bars, grey secondary y-axis) in muscle tissue of School Prawn collected at three locations of differing levels of PFAA contamination.

image file: c9em00394k-f3.tif
Fig. 3 Results of linear discriminant function analysis of PFAA concentrations measured in muscle tissue from School Prawn at three locations of differing levels of PFAA contamination (see figure legend). Group centroids are shown as filled circles, and the variation explained by each linear discriminant is given in brackets on the axes labels. Vectors indicate relative loading of linear coefficients for important PFAAs, and symbol size provides a relative indication of total PFAA concentration.

Spatial variation in the metabolome and associations with PFAA concentrations

Total fatty acid concentrations appeared similar among locations (Fig. 2), and this was confirmed by the lack of any significant differences (F1,42 = 1.4, P = 0.271). The average levels of composite contributions from saturated fatty acids (SFA), mono-unsaturated fatty acids (MUFA) and poly-unsaturated fatty acids (PUFA), are presented in Table 3. These show subtle shifts in composition of the odd- and even-chain SFA across the three locations while the total SFA remain relatively constant. Shifts were also apparent in the concentrations of MUFA and PUFA from the three locations. Locations formed distinct groupings along the LD (linear discriminant) axes based on their fatty acid profiles (Fig. 4), maintaining a similar configuration to groupings for PFAAs (Fig. 3); Hunter River and Wallis Lake were separated along LD1, and Hunter River locations were further separated along LD2. Distinct fatty acid molecules were important in driving differentiation of groups along each LD axes, and these are highlighted in Table 3 and the loadings (linear coefficients) on the LD axes are indicated in Fig. 4 (for the molecules with the 4 highest linear coefficients). Samples within locations grouped more tightly for fatty acid profiles, relative to PFAA profiles. Permutational ANOVA confirmed differences in fatty acid profiles among locations (pseudo-F1,42 = 5.7, P = 0.005), and pairwise analyses indicated that all locations were different from one another (Padj < 0.05). Further investigation of linear relationships between PFAA concentrations and fatty acids important in driving the groupings indicated that C19:0 (nonadecanoic acid) concentration tended to increase with linear PFOS, PFHxS, PFOA and PFNA, whereas concentrations of C20:4n6 (arachidonic acid), aiC17:0 (anteiso-margaric acid) and C16:0 (palmitic acid) decreased with increasing concentration of major PFAA molecules (Table 4).
Table 3 Mean fatty acid concentrations (μg mg−1, SE in brackets, n = 15) in muscle tissue of School Prawn collected at three locations of differing levels of PFAA contamination. The molecules with the five largest parameter coefficients for the two LD axes shown in Fig. 4 are indicated as superscripts (i.e. LD1, LD2) in the notation column. Also appended at the bottom of the table is the sum of concentrations for saturated fatty acids (SFA, odd and even-chained), mono-unsaturated fatty acids (MUFA), and poly-unsaturated fatty acids (PUFA), for each group
Notation Formula Common name Hunter River estuary Wallis Lake estuary
Fullerton Cove Tomago
C14:0 C14H28O2 Myristic acid 0.125 (0.009) 0.156 (0.017) 0.110 (0.013)
C16:1 C16H30O2 Palmitoleic acid 0.559 (0.035) 0.581 (0.047) 0.502 (0.069)
C16:0LD2 C16H32O2 Palmitic acid 3.241 (0.003) 3.533 (0.004) 3.285 (0.005)
iC17:0LD1 C17H34O2 Iso-margaric acid 0.117 (0.009) 0.134 (0.015) 0.144 (0.011)
aiC17:0LD2 C17H34O2 Anteiso-margaric acid 0.076 (0.005) 0.090 (0.010) 0.120 (0.007)
C17:1 C17H34O2 Heptadecenoic acid 0.955 (0.088) 0.473 (0.120) 0.197 (0.021)
C17:0 C17H34O2 Margaric acid 0.821 (0.052) 0.605 (0.067) 0.482 (0.039)
C18:2n6 C18H32O2 Linoleic acid 0.138 (0.015) 0.222 (0.045) 0.198 (0.038)
C18:1n7 C18H34O2 Octadecenoic acid 0.690 (0.030) 0.705 (0.047) 0.573 (0.039)
C18:1n9 C18H34O2 Oleic acid 0.293 (0.012) 0.341 (0.028) 0.364 (0.024)
C18:0 C18H36O2 Stearic acid 1.566 (0.052) 1.695 (0.094) 1.870 (0.084)
iC19:0 C22H46O2 Iso-nonadecanoic acid 0.024 (0.003) 0.038 (0.009) 0.028 (0.003)
aiC19:0LD1 C19H38O2 Anteiso-nonadecanoic acid 0.043 (0.003) 0.033 (0.003) 0.048 (0.004)
C19:0LD1 C19H38O2 Nonadecanoic acid 0.227 (0.015) 0.165 (0.025) 0.099 (0.009)
C20:4n6LD2 C20H32O2 Arachidonic acid 0.534 (0.025) 0.563 (0.043) 1.099 (0.075)
C20:5n3 C20H30O2 Eicosapentaenoic acid 1.476 (0.045) 1.677 (0.134) 0.989 (0.075)
C20:3n3 C20H34O2 Eicosatrienoic acid 0.035 (0.002) 0.038 (0.003) 0.048 (0.003)
C20:1 C20H38O2 Eicosenoic acid 0.091 (0.002) 0.103 (0.010) 0.090 (0.005)
C20:0 C20H40O2 Arachidic acid 0.101 (0.009) 0.124 (0.014) 0.207 (0.016)
iC20:0 C20H40O2 Iso-arachidic acid 0.047 (0.003) 0.048 (0.005) 0.121 (0.013)
C21:0 C21H42O2 Heneicosanoic acid 0.085 (0.007) 0.067 (0.011) 0.048 (0.009)
C22:6n3LD1, LD2 C22H32O2 Docosahexaenoic acid (DHA) 0.051 (0.004) 0.051 (0.007) 0.063 (0.008)
C22:5n3LD1 C22H34O2 Docosapentaenoic acid (DPA) 0.926 (0.049) 0.925 (0.082) 0.602 (0.048)
C22:0 C22H44O2 Behenic acid 0.085 (0.005) 0.101 (0.011) 0.109 (0.016)
C23:0 C23H46O2 Tricosanoic acid 0.051 (0.012) 0.053 (0.011) 0.029 (0.003)
C24:0LD2 C14H28O2 Myristic acid 0.053 (0.015) 0.028 (0.005) 0.045 (0.006)
Saturated fatty acids (SFA) 6.662 6.870 6.745
Odd-chain fatty SFA 1.444 1.185 0.998
Even chain SFA 5.218 5.685 5.747
Mono-unsaturated fatty acids (MUFA) 2.588 2.203 1.726
Poly-unsaturated fatty acids (PUFA) 3.160 3.476 2.999

image file: c9em00394k-f4.tif
Fig. 4 Results of linear discriminant function analysis of fatty acid concentrations measured in muscle tissues from School Prawn at three locations of differing levels of PFAA contamination (see figure legend). Group centroids are shown as filled circles, and the variation explained by each linear discriminant is given in brackets on the axes labels. Vectors indicate relative loading of linear coefficients for important fatty acids, and symbol size provides a relative indication of total PFAA concentration.
Table 4 Relationship between concentration of PFAAs and fatty acids contributing to separation among groups (see Table 3), evaluated using simple linear regression. The direction of the slope (+ or −), and significance level (* indicates P < 0.05, ** indicates P < 0.01) are indicated, and grey-shaded cells indicates that R2 > 0.2
image file: c9em00394k-u1.tif

Total amino acid concentrations showed significant (F1,42 = 3.3, P = 0.048, Fig. 2) differences among locations, and pairwise analysis indicated this was driven by Tomago having significantly higher (∼30 nmol mg−1 higher) amino acid concentrations than Wallis Lake (Padj = 0.048). Discriminant function analysis grouped samples within locations along LD axes similarly to the results for PFAAs and fatty acids, but with more variability within locations than fatty acids (Fig. 5). Despite this variability, permutational ANOVA indicated significant differences in amino acid profiles among locations (pseudo-F1,42 = 3.0, P = 0.014), and pairwise analyses indicated that all locations were different from one another (Padj < 0.05). Simple linear regression between PFAA concentrations and amino acids important in driving the groupings along each LD axis (Table 5) indicated that isoleucine negatively correlated with dominant PFAA molecules (PFOS-linear, PFHxS and PFOA, Table 6). However, of these correlations, isoleucine was only weakly correlated with PFOS-linear (R2 < 0.2), the most abundant PFAA. Tyrosine was also weakly correlated with PFDA.

image file: c9em00394k-f5.tif
Fig. 5 Results of linear discriminant function analysis of amino acid concentrations measured in muscle tissues from School Prawn at three locations of differing levels of PFAA contamination (see figure legend). Group centroids are shown as filled circles, and the variation explained by each linear discriminant is given in brackets on the axes labels. Vectors indicate relative loading of linear coefficients for important amino acids, and symbol size provides a relative indication of total PFAA concentration.
Table 5 Mean amino acid concentrations (nmol mg−1, SE in brackets, n = 15) in muscle tissue of School Prawn collected at three locations of differing levels of PFAA contamination. The molecules with the five largest parameter coefficients for the two LD axes shown in Fig. 5 are indicated as superscripts (i.e. LD1, LD2) in the notation column
Notation Formula Common name Hunter River estuary Wallis Lake estuary
Fullerton Cove Tomago
AAA C6H11NO4 Alpha-aminoadipic acid 0.296 (0.071) 0.593 (0.158) 0.740 (0.149)
ALA C3H7NO2 Alanine 15.83 (1.948) 17.053 (1.387) 16.015 (2.035)
ASNLD1 C4H8N2O3 Asparagine 6.538 (0.956) 8.108 (0.751) 7.533 (0.573)
CTH C7H14N2O4S Cystathionine 0.048 (0.016) 0.092 (0.027) 0.060 (0.014)
GLN C5H10N2O3 Glutamine 6.747 (2.711) 9.078 (2.514) 5.585 (1.789)
GLU C5H9NO4 Glutamic acid 3.103 (0.517) 2.524 (0.338) 3.280 (0.514)
GLY C2H5NO2 Glycine 85.642 (3.743) 107.754 (4.723) 94.050 (4.109)
GPR C7H12N2O3 Glycine-proline dipeptide 0.211 (0.052) 0.350 (0.079) 0.169 (0.032)
HISLD1, LD2 C6H9N3O2 Histidine 1.369 (0.199) 1.406 (0.147) 1.159 (0.142)
HYP C5H9NO3 Hydroxyproline 2.332 (0.681) 1.619 (0.432) 1.589 (0.328)
ILELD1, LD2 C6H13NO2 Isoleucine 1.018 (0.12) 1.939 (0.171) 1.740 (0.140)
LEU C6H13NO2 Leucine 1.128 (0.154) 2.068 (0.173) 1.827 (0.166)
LYS C6H14N2O2 Lysine 0.935 (0.167) 1.410 (0.196) 1.059 (0.104)
MET C5H11NO2S Methionine 0.645 (0.064) 0.805 (0.090) 0.836 (0.059)
ORN C5H12N2O2 Ornithine 0.104 (0.046) 0.096 (0.030) 0.166 (0.042)
PHE C9H11NO2 Phenylalanine 0.387 (0.033) 0.493 (0.045) 0.502 (0.054)
PRO C5H9NO2 Proline 47.901 (3.831) 39.189 (4.418) 30.786 (2.917)
SERLD2 C3H7NO3 Serine 10.222 (1.676) 8.546 (0.991) 9.505 (1.107)
THRLD1, LD2 C4H9NO3 Threonine 5.311 (0.806) 6.479 (0.570) 4.785 (0.466)
TRP C11H12N2O2 Tryptophan 0.018 (0.009) 0.013 (0.007) 0.013 (0.006)
TYRLD, LD21 C9H11NO3 Tyrosine 0.518 (0.082) 0.617 (0.066) 0.301 (0.050)
VAL C5H11NO2 Valine 5.714 (0.707) 10.276 (1.202) 6.902 (0.568)

Table 6 Relationship between concentration of PFAAs and amino acids contributing to separation among groups (see Table 5), evaluated using simple linear regression. The direction of the slope (+ or −), and significance level (* indicates P < 0.05, ** indicates P < 0.01) are indicated, and grey-shaded cells indicates that R2 > 0.2
image file: c9em00394k-u2.tif


Both concentrations and profiles of PFASs in School Prawn muscle tissue differed significantly among the three locations sampled in this study. The highest and lowest concentrations were found in School Prawn from Fullerton Cove and Wallis Lake respectively, and the concentrations were relatively consistent with previously reported concentrations in these areas.5,6,31 Both Fullerton Cove and Tomago had a higher concentration of PFOS-linear and PFOA than School Prawn from Wallis Lake, however, they were differentiated by different concentrations of PFHxS, PFDA and PFUnDA (although the latter was close to LOR).

Differences in PFAA profiles were mirrored by differences in both fatty acid and amino acid profiles, although there was more variability in amino acid profiles relative to fatty acids. While the magnitude of change in absolute concentrations were relatively small for some fatty acids and amino acids, changes were substantial for some molecules when relative concentrations among locations were considered. Unfortunately, there is little information in the literature with which to evaluate the impacts of changes in individual fatty acids in penaeid prawns, although examples from other taxa indicate that PFAAs can affect fatty acid homeostasis43 and lipid metabolism.21–25,29 Luebker et al.7 showed that PFAAs, and PFOS in particular, can inhibit binding to fatty acid carrier proteins in rodents. Fatty acid carrier proteins in shrimp show similarities with vertebrate carrier proteins, and play an important role in both lipid metabolism and immunity in invertebrates.44 While there was no significant difference in total fatty acid concentration among locations differentially impacted by PFAS contamination, certain free fatty acids individually correlated with PFOS concentrations, but in a contrasting manner. Nonadecanoic and docosapentaenoic acid (DPA) concentration positively correlated with PFOS (and other PFAAs) concentration, whereas palmitic, anteiso-margaric, and arachidonic acid negatively correlated. The differences in the fatty acid profiles among the three sites also indicated that higher levels of odd-chain fatty acids may be associated with PFAA contamination. This was reflected in a strong correlation between the most abundant odd-chain fatty acid C19:0 and the PFAA contaminants (Table 4). Also, the ratio of odd-chain[thin space (1/6-em)]:[thin space (1/6-em)]even-chain saturated fatty acid concentrations is higher with increasing levels of PFAA contamination (0.28 at Fullerton Cove, 0.21 at Tomago and 0.17 at Wallis Lake). The odd-chain fatty acids are most likely to be derived from phytoplankton and bacteria present in the sites, and this could be related to some impact of PFAA on food resources (although this is highly speculative). Thus, alterations in the fatty acid composition within the prawn muscle tissues could be arising from both the influence of PFAA on nutritional resources and/or direct effects of the bioaccumulated PFAA in lipid metabolism. These results provide a basis for further investigation, under laboratory controlled conditions, on further links between individual PFAAs and lipid metabolism in prawn muscle tissue. These relationships could also be further investigated in the hepatopancreas, however the small size of this organ means that samples will need to be pooled for analysis.

Like fatty acids, amino acid profiles were significantly different among locations. This could have arisen from alterations in nutritional resources influenced by PFAA and/or direct influence on amino acid homeostasis. The correlation analyses in Table 6 suggests that the branch chain amino acids (BCAA) leucine and isoleucine may have been directly influenced by PFAA bioaccumulation. These molecules differed to those that had shown responses to PFAS exposure in earthworms,23 and Zebrafish,16 but the crustacean Daphnia magna, showed decreased concentrations of isoleucine (a glucogenic amino acid) with increasing exposure to PFOS.29 The authors suggested that this could be evidence of increased energy metabolism through disruption of important metabolic pathways, which impacts other essential processes such as gluconeogenesis. While consideration of previous work in the literature points to some potential biological implications of the observed relationships, the majority of experiments have concentrated on PFOS exposure (and not other PFAAs). As for fatty acids, the snapshot of potential PFAA impacts on amino acids in free-ranging prawns presented in our study both highlights the need, and lays the foundation, for further investigation under tightly controlled laboratory conditions. Such work will aid further interpretation of relationships between PFAAs and amino acid homeostasis.

Noting that this project was designed as a first step to explore potential responses of important metabolite classes in relation to PFAS exposure under natural conditions, we have held back from a deeper interpretation of the individual metabolite differences observed, which requires more targeted experimentation. However, other examples highlight the use of “omic” studies to examine the impact of persistent organic pollutants (POPs, including endocrine disrupting contaminants) on aquatic taxa. Sanchez et al.45 reported a response of the liver proteome in Largemouth Bass (Micropterus salmoides) exposed to polychlorinated biphenyl (PCB) and other contaminants, including proteins associated with ion homeostasis and energy production. Shi et al.46 found exposure of Zebrafish (Danio rerio) to PFOS altered expression of multiple proteins, including functional protein classes involved in lipid transport, energy metabolism, and cell structure. The work of Samuelsson et al.47 highlights the potential of metabolomics as a discovery-driven approach to identify metabolic consequences of chemical exposure.19 These examples point to the broader potential of this approach for screening putative metabolic impacts of biological toxins (such as PFAAs) in natural aquatic environments.

Technical comments and limitations

It is important to identify potential caveats on the patterns reported here, to ensure results are not over-interpreted. The metabolome can be affected by multiple environmental variables, including both abiotic and biotic variables.48 While we took care to sample locations for which environmental variables were similar (Table 1) but contamination levels were different, we cannot completely exclude the potential for other extrinsic factors to have influenced amino acid and fatty acid concentrations. For example, some of the amino acids which showed changes among locations are also associated with osmoregulation, and there is a chance that relatively minor differences in salinities could have contributed to the patterns observed. Total fatty acid concentrations were similar among School Prawn sampled under similar conditions, regardless of whether they were from different estuaries or different locations within the same estuary,33 suggesting that this might not be a major issue for this species.

There were also slight differences in the length of time samples remained on ice following capture and processing, for different estuaries; there is a chance that this also could be a confounding factor contributing to minor differences in the metabolome among locations. Bottino et al.49 showed there was no significant change in fatty acid concentrations (fatty acids of chain length C14–C22 were tested) in the taxonomically related Brown Shrimp (Farfantepenaeus aztecus) over 18 days storage on ice, so such an impact is unlikely for fatty acids. The effect of storage on ice on the stability of amino acids is unclear, so this remains a potential confounding factor to consider in the interpretation of the results. In future studies, investigators should consider snap freezing animals upon capture, as well as undertaking more comprehensive analysis of water chemistry at collection locations to aid in interpretation of data. Also, analysing a greater sample size will provide increased statistical power to evaluate hypotheses surrounding linkages between the organism metabolome and PFAA contamination in free-ranging animals.


While it is not possible to explore detailed relationships between PFAA bioaccumulation and organism metabolism with the existing data set, the results presented here do indicate an association between PFAAs concentration and individual metabolites in shrimp. These relationships are suggestive of broadscale changes in homeostasis across differentially exposed animals, which may be of consequence for exposed prawns and other crustaceans. It is important to note that our study animals could be considered to have comparatively low concentrations, relative to crustaceans impacted elsewhere (e.g. van de Vijver et al.50), and the observed changes may thus indicate a conservative estimate of potential impact. The consequence of these changes is difficult to establish given the lack of detailed studies on the metabolome in penaeid prawn species, however our results lay the foundation for more in-depth investigation of these relationships under controlled conditions, and potential development of dose–response curves for metabolites of interest (or metabolites of known consequence to organism function). Expanding this body of knowledge is essential to fully appreciate the impact of PFAS contamination on marine animals, and the fisheries that rely on them.

Conflicts of interest

There are no conflicts of interest to declare.


This study was funded by the NSW Department of Primary Industries, and sample analysis undertaken using the resources of Queensland Alliance for Environmental Health Sciences, The University of Queensland. The Queensland Alliance for Environmental Health Sciences, The University of Queensland, gratefully acknowledges the financial support of the Queensland Department of Health. We wish to thank D. Cruz, D. Elliot, C. McLuckie, and J. Hewitt for assistance with sample collection. Sample collection was conducted under a Section 37 Scientific Collection Permit (permit P01/0059).


  1. L. Ahrens, Polyfluoroalkyl compounds in the aquatic environment: a review of their occurrence and fate, J. Environ. Monit., 2011, 13, 20–31 RSC.
  2. L. Ahrens and M. Bundschuh, Fate and effects of poly- and perfluoroalkyl substances in the aquatic environment: A review, Environ. Toxicol. Chem., 2014, 33, 1921–1929 CrossRef CAS PubMed.
  3. K. E. Murray, S. M. Thomas and A. A. Bodour, Prioritizing research for trace pollutants and emerging contaminants in the freshwater environment, Environ. Pollut., 2010, 158, 3462–3471 CrossRef CAS PubMed.
  4. A. G. Paul, K. C. Jones and A. J. Sweetman, A first global production, emission, and environmental inventory for perfluorooctane sulfonate, Environ. Sci. Technol., 2009, 43, 386–392 CrossRef CAS PubMed.
  5. M. D. Taylor and D. D. Johnson, Preliminary investigation of perfluoroalkyl substances in exploited fishes of two contaminated estuaries, Mar. Pollut. Bull., 2016, 111, 509–513 CrossRef CAS PubMed.
  6. M. D. Taylor, S. Nilsson, J. Bräunig, K. C. Bowles, V. Cole, N. A. Moltschaniwskyj and J. F. Mueller, Do conventional cooking methods alter concentrations of per- and polyfluoroalkyl substances (PFASs) in seafood?, Food Chem. Toxicol., 2019, 127, 280–287 CrossRef CAS.
  7. D. J. Luebker, K. J. Hansen, N. M. Bass, J. L. Butenhoff and A. M. Seacat, Interactions of fluorochemicals with rat liver fatty acid-binding protein, Toxicology, 2002, 176, 175 CrossRef CAS PubMed.
  8. T. Fletcher, T. S. Galloway, D. Melzer, P. Holcroft, R. Cipelli, L. C. Pilling, D. Mondal, M. Luster and L. W. Harries, Associations between PFOA, PFOS and changes in the expression of genes involved in cholesterol metabolism in humans, Environ. Int., 2013, 57–58, 2–10 CrossRef CAS PubMed.
  9. M. D. Taylor, Survey design for quantifying perfluoroalkyl acid concentrations in fish, prawns and crabs to assess human health risks, Sci. Total Environ., 2019, 652, 59–65 CrossRef PubMed.
  10. M. Houde, A. O. De Silva, D. C. Muir and R. J. Letcher, Monitoring of perfluorinated compounds in aquatic biota: An updated review, Environ. Sci. Technol., 2011, 45, 7962–7973 CrossRef CAS PubMed.
  11. M. Habibullah-Al-Mamun, M. K. Ahmed, M. Raknuzzaman, M. S. Islam, M. M. Ali, M. Tokumura and S. Masunaga, Occurrence and assessment of perfluoroalkyl acids (PFAAs) in commonly consumed seafood from the coastal area of Bangladesh, Mar. Pollut. Bull., 2017, 124, 775–785 CrossRef CAS PubMed.
  12. A.-M. Nyman, K. Schirmer and R. Ashauer, Importance of toxicokinetics for interspecies variation in sensitivity to chemicals, Environ. Sci. Technol., 2014, 48, 5946–5954 CrossRef CAS PubMed.
  13. J. G. Bundy, M. P. Davey and M. R. Viant, Environmental metabolomics: a critical review and future perspectives, Metabolomics, 2008, 5, 3 CrossRef.
  14. G. Signa, R. Di Leonardo, A. Vaccaro, C. D. Tramati, A. Mazzola and S. Vizzini, Lipid and fatty acid biomarkers as proxies for environmental contamination in caged mussels Mytilus galloprovincialis, Ecol. Indicat., 2015, 57, 384–394 CrossRef CAS.
  15. Y.-K. Kwon, Y.-S. Jung, J.-C. Park, J. Seo, M.-S. Choi and G.-S. Hwang, Characterizing the effect of heavy metal contamination on marine mussels using metabolomics, Mar. Pollut. Bull., 2012, 64, 1874–1879 CrossRef CAS PubMed.
  16. S. S. Y. Huang, J. P. Benskin, B. Chandramouli, H. Butler, C. C. Helbing and J. R. Cosgrove, Xenobiotics produce distinct metabolomic responses in Zebrafish Larvae (Danio rerio), Environ. Sci. Technol., 2016, 50, 6526–6535 CrossRef CAS PubMed.
  17. C. Y. Lin, M. R. Viant and R. S. Tjeerdema, Metabolomics: Methodologies and applications in the environmental sciences, J. Pest. Sci., 2006, 31, 245–251 CrossRef CAS.
  18. J. Sardans, J. Peñuelas and A. Rivas-Ubach, Ecological metabolomics: overview of current developments and future challenges, Chemoecology, 2011, 21, 191–225 CrossRef CAS.
  19. L. M. Samuelsson and D. G. J. Larsson, Contributions from metabolomics to fish research, Mol. BioSyst., 2008, 4, 974–979 RSC.
  20. R. N. Finn and H. J. Fyhn, Requirement for amino acids in ontogeny of fish, Aquacult. Res., 2010, 41, 684–716 CrossRef CAS.
  21. T. L. Alderete, R. Jin, D. I. Walker, D. Valvi, Z. Chen, D. P. Jones, C. Peng, F. D. Gilliland, K. Berhane, D. V. Conti, M. I. Goran and L. Chatzi, Perfluoroalkyl substances, metabolomic profiling, and alterations in glucose homeostasis among overweight and obese Hispanic children: A proof-of-concept analysis, Environ. Int., 2019, 126, 445–453 CrossRef CAS PubMed.
  22. S. Salihovic, T. Fall, A. Ganna, C. D. Broeckling, J. E. Prenni, T. Hyötyläinen, A. Kärrman, P. M. Lind, E. Ingelsson and L. Lind, Identification of metabolic profiles associated with human exposure to perfluoroalkyl substances, J. Exposure Sci. Environ. Epidemiol., 2019, 29, 196–205 CrossRef CAS PubMed.
  23. B. Lankadurai, V. Furdui, E. Reiner, A. Simpson and M. Simpson, 1H NMR-based metabolomic analysis of sub-lethal perfluorooctane sulfonate exposure to the earthworm, Eisenia fetida, in soil, Metabolites, 2013, 3, 718–740 CrossRef PubMed.
  24. E. Ortiz-Villanueva, J. Jaumot, R. Martínez, L. Navarro-Martín, B. Piña and R. Tauler, Assessment of endocrine disruptors effects on zebrafish (Danio rerio) embryos by untargeted LC-HRMS metabolomic analysis, Sci. Total Environ., 2018, 635, 156–166 CrossRef CAS PubMed.
  25. A. Arukwe and A. S. Mortensen, Lipid peroxidation and oxidative stress responses of salmon fed a diet containing perfluorooctane sulfonic- or perfluorooctane carboxylic acids, Comp. Biochem. Physiol., Part C: Pharmacol., Toxicol. Endocrinol., 2011, 154, 288–295 CAS.
  26. N. D. Wagner, A. J. Simpson and M. J. Simpson, Metabolomic responses to sublethal contaminant exposure in neonate and adult Daphnia magna, Environ. Toxicol. Chem., 2017, 36, 938–946 CrossRef CAS PubMed.
  27. V. Kovacevic, A. J. Simpson and M. J. Simpson, The concentration of dissolved organic matter impacts the metabolic response in Daphnia magna exposed to 17α-ethynylestradiol and perfluorooctane sulfonate, Ecotoxicol. Environ. Saf., 2019, 170, 468–478 CrossRef CAS PubMed.
  28. N. D. Wagner, P. A. Helm, A. J. Simpson and M. J. Simpson, Metabolomic responses to pre-chlorinated and final effluent wastewater with the addition of a sub-lethal persistent contaminant in Daphnia magna, Environ. Sci. Pollut. Res., 2019, 26, 9014–9026 CrossRef CAS PubMed.
  29. M. Kariuki, E. Nagato, B. Lankadurai, A. Simpson and M. Simpson, Analysis of sub-lethal toxicity of perfluorooctane sulfonate (PFOS) to Daphnia magna using 1H nuclear magnetic resonance-based metabolomics, Metabolites, 2017, 7, 15 CrossRef PubMed.
  30. M. D. Taylor, K. C. Bowles, D. D. Johnson and N. A. Moltschaniwskyj, Depuration of perfluoroalkyl substances from the edible tissues of wild-caught invertebrate species, Sci. Total Environ., 2017, 581, 258–267 CrossRef PubMed.
  31. M. D. Taylor, J. Beyer-Robson, D. D. Johnson, N. A. Knott and K. C. Bowles, Bioaccumulation of perfluoroalkyl substances in exploited fish and crustaceans: Spatial trends across two estuarine systems, Mar. Pollut. Bull., 2018, 131, 303–313 CrossRef CAS PubMed.
  32. M. D. Taylor, Factors affecting spatial and temporal patterns in perfluoroalkyl acid (PFAA) concentrations in migratory aquatic species: A case study of an exploited crustacean, Environ. Sci.: Processes Impacts, 2019 10.1039/c9em00202b.
  33. M. D. Taylor, N. A. Moltschaniwskyj, M. J. Crompton and R. H. Dunstan, Environmentally-driven changes in fatty acid profiles of a commercially important penaeid prawn, Estuaries Coasts, 2019, 42, 528–536 CrossRef CAS.
  34. M. D. Taylor, B. Fry, A. Becker and N. A. Moltschaniwskyj, The role of connectivity and physicochemical conditions in effective habitat of two exploited penaeid species, Ecol. Indicat., 2017, 80, 1–11 CrossRef CAS.
  35. K. J. Tyler, A. Becker, N. M. Moltschaniwskyj and M. D. Taylor, Rapid salinity changes impact the survival and physiology of a penaeid prawn: Implications of flood events on recruitment to the fishery, Fish. Manag. Ecol., 2017, 24, 478–487 CrossRef.
  36. C. Baduel, F. Y. Lai, K. Townsend and J. F. Mueller, Size and age-concentration relationships for perfluoroalkyl substances in stingray livers from eastern Australia, Sci. Total Environ., 2014, 496, 523–530 CrossRef CAS PubMed.
  37. M. J. Crompton and R. H. Dunstan, Evaluation of in situ fatty acid extraction protocols for the analysis of staphylococcal cell membrane associated fatty acids by gas chromatography, J. Chromatogr. B: Biomed. Sci. Appl., 2018, 1084, 80–88 CrossRef CAS PubMed.
  38. J. D. Rabinowitz and E. Kimball, Acidic acetonitrile for cellular metabolome extraction from Escherichia coli, Anal. Chem., 2007, 79, 6167–6173 CrossRef CAS PubMed.
  39. C. Evans, H. R. Dunstan, T. Rothkirch, T. K. Roberts, K. L. Reichelt, R. Cosford, G. Deed, L. B. Ellis and D. L. Sparkes, Altered amino acid excretion in children with autism, Nutr. Neurosci., 2008, 11, 9–17 CrossRef CAS PubMed.
  40. R Core Team, R: A language and environment for statistical computing v 3.5.2, 2016 Search PubMed.
  41. M. Kuhn, Building predictive models in R using the caret package, J. Stat. Softw., 2008, 28, 1–26 Search PubMed.
  42. W. N. Venables and B. D. Ripley, Modern Applied Statistics with S, Springer, New York, 4th edn, 2002 Search PubMed.
  43. I. Curran, S. L. Hierlihy, V. Liston, P. Pantazopoulos, A. Nunnikhoven, S. Tittlemier, M. Barker, K. Trick and G. Bondy, Altered fatty acid homeostasis and related toxicologic sequelae in rats exposed to dietary potassium perfluorooctanesulfonate (PFOS), J. Toxicol. Environ. Health, 2008, 71, 1526–1541 CrossRef CAS.
  44. Q. Ren, Z.-Q. Du, X.-F. Zhao and J.-X. Wang, An acyl-CoA-binding protein (FcACBP) and a fatty acid binding protein (FcFABP) respond to microbial infection in Chinese white shrimp, Fenneropenaeus chinensis, Fish Shellfish Immunol., 2009, 27, 739–747 CrossRef CAS.
  45. B. C. Sanchez, K. J. Ralston-Hooper, K. A. Kowalski, H. Dorota Inerowicz, J. Adamec and M. S. Sepúlveda, Liver proteome response of largemouth bass (Micropterus salmoides) exposed to several environmental contaminants: Potential insights into biomarker development, Aquat. Toxicol., 2009, 95, 52–59 CrossRef CAS.
  46. X. Shi, L. W. Yeung, P. K. Lam, R. S. Wu and B. Zhou, Protein profiles in zebrafish (Danio rerio) embryos exposed to perfluorooctane sulfonate, Toxicol. Sci., 2009, 110, 334–340 CrossRef CAS.
  47. L. M. Samuelsson, L. Förlin, G. Karlsson, M. Adolfsson-Erici and D. G. J. Larsson, Using NMR metabolomics to identify responses of an environmental estrogen in blood plasma of fish, Aquat. Toxicol., 2006, 78, 341–349 CrossRef CAS.
  48. M. R. Viant, Metabolomics of aquatic organisms: the new "omics" on the block, Mar. Ecol.: Prog. Ser., 2007, 332, 301–306 CrossRef CAS.
  49. N. R. Bottino, M. L. Lilly and G. Finne, Fatty acid stability of Gulf of Mexico Brown Shrimp (Penaeus aztecus) held on ice and in frozen storage, J. Food Sci., 1979, 44, 1778–1779 CrossRef CAS.
  50. K. I. Van de Vijver, P. T. Hoff, W. Van Dongen, E. L. Esmans, R. Blust and W. M. De Coen, Exposure patterns of perfluorooctane sulfonate in aquatic invertebrates from the Western Scheldt estuary and the southern North Sea, Environ. Toxicol. Chem., 2003, 22, 2037–2041 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c9em00394k

This journal is © The Royal Society of Chemistry 2019