Uncovering the sub-lethal impacts of plastic ingestion by shearwaters using fatty acid analysis

Our study is the first to describe the fatty acid composition of the liver, breast muscle, and adipose tissue from flesh-footed (Ardenna carneipes) and short-tailed shearwaters (A. tenuirostris) and uses this technique to explore possible sub-lethal effects of plastic ingestion on these species.


Introduction
In <70 years, plastic waste has become persistent in almost all terrestrial and aquatic habitats to such an extent that plastic debris is present in the stratigraphy of most sedi-mentary deposits and is one of the primary indicators of the human-induced, geological epoch known as the Anthropocene (Waters et al., 2016, Zalasiewicz et al., 2016. Once in marine environments, plastic fragments into small particles known as micro-(1-5 mm) and nano-plastics (<1μm; Barnes et al., 2009, Provencher et al., 2017. Plastic can then be readily consumed by a range of marine life from krill (Dawson et al., 2018) to large marine mammals (Fossi et al., 2018, Nelms et al., 2018. Seabirds, particularly Procellariiformes (i.e. albatross, petrels and shearwaters), have been recorded with some of the highest plastic ingestion rates, which have been attributed to factors such as foraging strategy, plastic colour and odour (Lavers and Bond, 2016, Savoca et al., 2016, Tavares et al., 2017, Verlis et al., 2013. Impacts from the ingestion of plastic may include damage to tissues, morbidity and starvation (Auman et al., 1997, Lavers et al., 2014. Lipid-derived fatty acids (FAs) serve a functional purpose within an animal's body where they may be stored or metabolized in a variety of tissues (Parrish et al., 2015). Different FAs serve different metabolic functions in animals and most FAs are combined into a range of lipid classes. These include triglycerides (TAG), which are key to the storage of energy, phospholipids, which comprise the structural components of cell membranes, or wax esters, which are stored in various tissue structures (Ramos and Gonzalez-Solis, 2012). The majority of adipose tissue consists of TAG, which is a crucial energy source for young birds during periods of parental neglect and in preparation for fledging (Ricklefs and Schew, 1994). In migratory birds, adipose tissue containing abundant FAs, such as 16:0 palmitic acid, 18:1n9 (an omega-9 FA) and 18:2n6 (an omega-6 FA), is stored in excess and metabolized rapidly in the extreme energy bursts experienced during migration (McWilliams et al., 2004).
Migratory marine megafauna, which travel large distances to exploit variable food sources, rely heavily on lipid reserves to survive these journeys (Ramos and Gonzalez-Solis, 2012). In some species, adipose deposition has been scored and compared to morphometric measurements to determine body condition (Auman et al., 1997, Cousin et al., 2015, Schultner et al., 2013. Application of these methods to seabirds has received criticism as some measures of condition are thought to be subjective ( van Franeker, 2004). Chick growth can be highly variable due to the wide range of prey sources exploited and infrequent feeding of seabird chicks (Angel et al., 2015, Connan et al., 2005, Taylor and Konarzewski, 1989, Williams et al., 2009. As a result, validation of current measures of body condition using direct, precise analytical techniques (e.g. dietary lipids) would inform our understanding of bird health (Schamber et al., 2009). The application of FA analysis of tissues has been used to describe spatial and temporal shifts in diet, revealing trophic interactions (Haynes et al., 2015, Karnovsky et al., 2012. For example, in fledgling seabirds, adipose tissue reflects the dietary FAs consumed by an individual over a period of ∼1-2 months (Williams et al., 2009), providing a rapid method to assess short-term diet. The description of FA compositions as a response to toxicological factors, and issues such as the sub-lethal impact of plastic ingestion, has not yet been explored.
Some seabird groups ingest high quantities of plastic (e.g. 56% of Procellariformes; Kühn et al., 2015), which can be a major cause of morbidity and altered physiology (Auman et al., 1997, Lavers et al., 2014; however, very few studies have described this relationship successfully (Carey, 2011, Cousin et al., 2015, Ryan, 1988. The application of FAs may have the potential to explore these important questions in wild, free-living seabirds as current seabird linear morphometric measurements, used as indicators of condition, have not been validated against FA or protein data (Schamber et al., 2009), and these measurements often vary among fledglings within a colony as adults return to feed chicks at different intervals. Here we investigate the relationship between FA composition, linear morphometric measurements and plastic ingestion in two species of pelagic seabird: flesh-footed shearwaters (Ardenna carneipes) and short-tailed shearwaters (Ardenna tenuirostris) to better understand the potential sublethal impacts of ingested plastic.

Materials and methods
We sampled 18 fledgling shearwaters (∼80-90 days old) of each species for this study. Freshly dead (road-kill, beachwashed) flesh-footed shearwater fledglings were collected on Lord Howe Island, New South Wales, Australia (31.554 • S, 159.084 • E) from 26 April to 12 May 2017 (Fig. 1). Shorttailed shearwater fledglings were harvested by local hunters under recreational (individually held) collection permits from Great Dog Island in the Furneaux Group, Tasmania, Australia (40.247 • S, 148.249 • E) in late April 2017 (Fig. 1). Harvested birds were selected randomly and included a range of body sizes.
Bird body mass (±10 g) was determined using a spring balance, wing chord (unflattened and straightened; ±1 mm) using a stopped ruler, culmen and head + bill length using Vernier callipers (±0.1 mm). Visible ingested plastic items (>1 mm) from the proventriculus and gizzard were dried and weighed to the nearest 0.0001 g using an electronic balance. Partial breast muscle (pectoralis major), one lobe of the liver and 2-5 g subcutaneous adipose tissue from the breast were removed from each bird. Tissues were wrapped in aluminium foil, placed in individually labelled bags and stored at −20 • C. Tissues were weighed and then freeze-dried for 72 h for dry matter and percent moisture determination and homogenized using a mortar and pestle. The lipids were extracted using a modified Bligh and Dyer (1959) protocol. Briefly, 0.1 g of the tissue was added to a chloroform/methanol/distilled water (1/2/0.8, v/v/v) solution in a glass flask, shaken multiple times and left overnight for extraction. Phase separation was induced by the addition of chloroform/water (1/1, v/v) and shaking on the following day. The chloroform phase that contained the lipids was concentrated under reduced pressure at 42 • C and the total lipid content was determined gravimetrically. Then, 1.5 ml of chloroform was added to the lipid extract and 750 μl (Bligh and Dyer, 1959)   acid methyl ester (FAME) dissolved in CHCl 3 (500 mg/l). The mixture was concentrated under a stream of nitrogen gas. Next, 3 ml of saponification reagent (5% (w/v) KOH in 80:20 (v/v) MeOH:H 2 O) was added and the whole solution heated at 80 • C for 3 h. Once cooled to room temperature, the mixture was diluted by adding 1 ml of distilled water. To create a phase separation, 1.5 ml of hexane:chloroform (4:1) was added. The top aqueous phase containing the nonsaponifiable neutral lipids was removed. The lower layer was mixed by hand with 1 ml 2 M hydrochloric acid and 1.5 ml of hexane chloroform ((Hex:CHCl 3 ) 4:1 (v/v)) and the resulting top aqueous layer placed in another vial. Then, 1 ml of methylation reagent was added to this new solution and left to rest for 1 h at 80 • C before adding 1 ml H 2 O, and then 1.5 ml of Hex:CHCl 3 solution was added and the sample concentrated under a nitrogen gas stream. Finally, 1000 μl of internal standard working solution (50 μg/ml C19 FAME in CHCl 3 ) was added and the extracted FAME sample was stored at −20 • C. Standard working solution was added to each sample at the final step of FAME extraction to calculate the potential sample loss. Blank samples were used to calibrate for outside contaminants during the gas chromatography-mass spectrometry (GC-MS) analysis. FAME samples were analysed using a Varian CP-3800 gas chromatograph equipped with a CP-8400 autosampler, coupled to a Brüker 300-MS triple quadrupole mass spectrometer. Stationary phase was an Agilent DB-5MS column, 30 m × 0.25 mm, with 0.25 μm phase thickness. Helium was used as the carrier gas. Electron ionization mass spectra of FAME was recorded in full scan mode.

Statistical analysis
All statistical analysis was completed using R 3.4.3 (R Core Team, 2018) in RStudio (v.1.1.453, Boston, Massachusetts, USA). The relationship between mass of ingested plastic items and bird morphometrics was explored using a linear regression and a Cook's Distance of > 3 identified statistical outliers (Rousseeuw and Leroy, 2005). Results where P < 0.05 were regarded as statistically significant. Following a similar statistical approach to Williams et al. (2009), all FAs that were expressed in the tissues of the individuals at < 0.01% were normalized and the values logtransformed: x trans = log x i /c r where x i denotes the percentage composition of a FA, x trans is the transformed FA and c r is a random reference FA found in all samples (in our case, 18:0). Because a log transformation cannot be performed on 0 values, all FA values were altered by adding 0.01; logged values were then analysed using a principal component analysis (PCA). To examine the relationships between FA composition among tissues and species, and in relation to ingested plastic, we used a multivariate analysis of variance (MANOVA) with principle component (PC) scores as the response. A Wilk's (λ) test statistic was used and relationships were considered significant when P < 0.05.
A total of 37 FA were found in each of the three tissues of both species tested (Supplementary Table 1). Monounsaturated FA (MUFA) were the predominant FA class, accounting for 35.9% of all FA on average and ranging from 28.9-49.9% in the tissues of short-tailed and from 20.7-37.6% in the tissues of flesh-footed, respectively. The second most abundant FA class were the saturated FAs (SFAs, 31.3-38.1%, 21.8% and 31.2%). The content of the remaining poly-unsaturated FA (PUFA) ranged from 15.9-33.0% and 36.8-57.5% in the two species, respectively.
A PCA was run on all FAs detected in the samples (Fig. 2). The factor loadings for PCs 1-4 accounted for 62% of the variance in the data. The two species grouped distinctly with little overlap between the FA compositions (Fig. 2). PC 1 and 3 were strongly driven by low percentages of FAs 16:1, 18:1 and 18:2n6 while PC 2 and 4 were strongly positively correlated with low proportions of FAs 15:0, 17:0, 17:1, 20:5n3 and 22:6n3 FAs (Table 1).

Differences in FA compositions
Palmitic acid (16:0) is the most abundant FA found in animals, and is commonly the product of de novo synthesis of 14carbon FAs within the liver of seabirds (Käkelä et al., 2009). This particular FA was the most abundant in the tissues of both shearwater species examined (Supplementary Table 1), as well as in the breast muscle of the closely related sooty shearwater (Ardenna grisea) (Al-Amer et al., 2016) and stomach oil of sub-Antarctic breeding albatrosses (Connan et al., 2014). Palmitic acid may be stored in the adipose tissue or used rapidly as an energy substrate (Williams and Buck, 2010). The abundance of this FA is believed to be important for the release of energy during migration, as observed in songbirds (McWilliams et al., 2004) and may explain why it is found in such high levels in fledgling shearwaters.
Essential fatty acids (EFAs) must be obtained through diet and cannot be synthesized de novo. The EFAs 20:5n3 (eicosapentaenoic acid, EPA) and 22:6n3 (docosahexaenoic acid, DHA) were found in high concentrations in the liver and muscle tissues of both species (Supplementary Table 1). EPA was found in relative abundance in short-tailed shearwaters (6.5% ± 2.7%) compared to flesh-footed shearwaters (2.2% ± 1.2%) and other members of the Procellariidae (Al-Amer et al., 2016, Wang et al., 2007, Woodward et al., 1995. Overall, EFAs are crucial for avian health, including the development of polar lipid structures (i.e. muscle) in young birds (as has been shown in chickens; Maldjian et al., 1995).
This study demonstrates the scope of intra-genus physiological differences among the two study species of shearwaters. Differences in FA composition among tissue types can likely be attributed to the function of the tissues (e.g. the liver in bioprocessing and de novo synthesis of FAs), the turnover rate of FAs within certain tissues and the typical lipid classes that are associated with each tissue type (Ramos andGonzalez-Solis, 2012, Williams andBuck, 2010). Results from the PCA (Fig. 2) indicate flesh-footed and short-tailed shearwaters form two distinct groups based on their FA composition. The two species feed in different environments on prey from low trophic, sub-Antarctic environments (short-tailed shearwaters) or high trophic, sub-tropical environments (flesh-footed shearwaters); therefore, the FA composition of prey species is likely driving this difference, since the majority of FAs is derived from prey (Connan et al., 2005). Flesh-footed shearwaters feed primarily on mesopelagic fishes and squid (Gould et al., 1997), prey that are typically high in 16:0, 18:1n9 and 22:6n3 FAs (Lewis, 1967). Conversely, short-tailed shearwaters forage on smaller, lower trophic level prey, with Antarctic krill (Euphausia superba) and small cephalopods being the most dominant species (Carey, 2011, Connan et al., 2005. Antarctic krill are abundant in 14:0, 16:0, 16:1 and 20:5n3 FAs (Fricke et al., 1984), and this may be reflected in the FA outputs of these birds (Supplementary Table 1). We found higher amounts of 18:1n9 in flesh-footed shearwater tissues (29.4% ± 6.8%) and more 20:5n3 in short-tailed shearwaters (6.5% ± 2.7%), supporting the diet-driven differences above. While Procellariiformes exhibit some variation in their FA composition (Al-Amer et al., 2016, Wang et al., 2007, Woodward et al., 1995  stearic acid are most common in marine prey species (Speake et al., 1999). From the limited information available, the FA composition of flesh-footed and short-tailed shearwaters, though individually different, is consistent with that of other Procellariiformes.

Body condition and plastic ingestion
FA PC scores and linear morphometrics showed no relationship with the mass of plastic ingested in either species, except in the muscle of short-tailed shearwaters (P < 0.019). Analysis of flesh-footed shearwater data from 2011 suggested a strong link between plastic ingestion and decreased body size at fledging, and potentially a poorer likelihood of survivability due to high plastic ingestion (Lavers et al., 2014). The same analysis conducted on birds from 2017 failed to show these relationships, serving as a reminder of the annual variability in plastic ingestion (much lower in 2017 compared to 2011), in populations' responses to ingestion, and the difficulty of identifying sub-lethal impacts of plastic in seabirds. Finally, our inability to detect a relationship between ingested plastic, linear morphometrics and FA composition across both species may have been hampered by the small sample size (n = 18 in 2017 compared to n = 37 in 2011) and variability in the number and mass of plastic ingested by birds in 2017. The mean mass of ingested plastic was unrelated to body size of short-tailed shearwaters collected in 2017, a pattern that was also reported by Carey (2011) and Cousin et al. (2015) who failed to find statistically significant relationship between plastic and fat condition scores. Plastic ingestion was, however, related to short-tailed shearwater FA composition in the breast muscle only, indicating plastic at this concentration either had few measurable impacts on this species, or that impacts affected different physiological pathways.
It is possible this study may have overlooked a number of dietary-derived FAs in tissues with fast metabolic turn over, as the fledgling birds collected at the time of this study were undergoing a period of parental neglect and may not have fed in 1-2 weeks (Schultz andKlomp, 2000, Williams et al., 2007). This period of short starvation may also influence the variability in these results. Allmann et al. (1965) observed spikes in 16:1 and 18:1 FAs in liver samples after re-feeding starved rats (Rattus norvegicus) and suggested a relationship with linoleic acid (18:2n6) deficiencies. Our birds showed high concentrations of 16:1 FA, which may align with our knowledge of their life history (i.e. periods of parental neglect). Few data are available on FA profiles in relation to starvation in seabirds (Phillips and Hamer, 1999). FA metabolism in seabirds differs significantly from other species, including mammals, thereby limiting the application of this literature (Williams and Buck, 2010). More seabirdfocused studies are required to develop a comprehensive understanding of these relationships. condition in free-living birds (Williams et al., 2009). Connors and Smith (1982) was one of the earliest studies to identify a negative correlation between ingested plastic and fat deposition in seabirds. Auman et al. (1997) continued to explore fat scoring as an indicator for condition, a method that has been criticized for being subjective ( van Franeker, 2004). This presents an opportunity to identify alternate techniques to determine body condition and justify morphometric measurements in this group of seabirds. Schamber et al. (2009) argued that a quantitative index of condition is required to limit similar criticisms and provide confidence when using condition indices to assess health. Using PCA to explore FA composition, we determined PC1 and PC3 in the adipose and muscle of flesh-footed shearwaters and liver and muscle of short-tailed shearwaters were correlated with bird body mass (Table 1). This suggests FAs can accurately reflect certain morphometric measurements when sampled from the correct tissues, information that can be used to justify these measurements in future studies.

Conclusions
Effective conservation and management of seabirds relies on a comprehensive and robust understanding of the pressures faced by populations and individuals. Current knowledge of the impacts of marine plastic pollution is limited to superficial health assessments and documentation of debris ingestion by species. FA analysis was one tool to explore how plastic may disrupt nutritional pathways, though our findings suggest that at least with our sample of birds from a single year, there was no effect. Other similar tools may play an important role in uncovering the sub-lethal impacts of plastic, and should be explored as current conservation and management strategies are not able to account for morbidity and mortality associated with the less visible impacts of plastic. As plastic production and waste increases, we expect to see these impacts on wildlife increase.