Temporal Variability of Polycyclic Aromatic Hydrocarbons in Deep-Sea Cephalopods of the Northern Gulf of Mexico

As part of the effort to understand the effects of the Deepwater Horizon Oil Spill (DWHOS), we analyzed tissue from five species of midwater oceanic cephalopods in the northern Gulf of Mexico (GoM) during three time periods, including one period sampled fortuitously just before the spill (2010), and two periods sampled after the spill (2011 and 2015–2016). The species, Japetella diaphana, Abralia redfieldi, Histioteuthis corona, Leachia atlantica, and Onychoteuthis banksii were collected in three geographic areas in the GoM (east, south and southeast of the Macondo wellhead). Results indicate a shift in the composition of polycyclic aromatic hydrocarbons (PAHs) in the tissue of all cephalopods after 2010, with a more petrogenic source in 2011 that weathered and mixed with other sources in 2015–2016. Overall, PAH concentrations, as well as lipid content, were lower in 2011 relative to 2010 and 2015–2016, suggesting secondary effects to oil-residues exposure from the DWHOS. Collectively, PAHs in the tissues of deep-sea cephalopods indicate an episodic exposure to petrogenic PAHs that occurred between 2010 (pre-spill) and 2011, and continued through 2015–2016.


INTRODUCTION
Ecosystem-based management of the oceans relies heavily on continuous observations, essential for distinguishing natural variability from anthropogenic changes. However, epi-, meso-, and bathypelagic environments generally lack such studies. This was highlighted in 2010 by the absence of baseline data during the Deepwater Horizon oil spill (DWHOS) in the Gulf of Mexico (GoM). The DWHOS occurred at a depth of 1500 m, releasing approximately 4000000 barrels of oil that contaminated vast areas of the water column (e.g., oil residues rose to form surface slicks contaminating the water column, subsurface plumes concentrated at 900-1300 m) and coastal environments (e.g., marshes, beaches), as well as an extensive area of the seafloor (about 76000 km 2 by sinking of oil residues from the water column) (Dietrich et al., 2012;Brooks et al., 2015;MacDonald et al., 2015;Romero et al., 2015Romero et al., , 2017Daly et al., 2016;Harding et al., 2016;Murawski et al., 2016;Yan et al., 2016). The deep-pelagic habitat in the GoM was among the environments most affected by the DWHOS, as indicated by multiple studies detecting high concentration of contaminants and shifts in the composition of microbial communities (Hazen et al., 2010;Fisher et al., 2016;Wade et al., 2016;Yan et al., 2016). Effects to marine communities were observed as carbon from the DWHOS entered the planktonic and mesopelagic food webs, and oilderived toxic compounds were found to correlate with elevated skin lesions in bottom-dwelling offshore fishes (Graham et al., 2010;Chanton et al., 2012;Murawski et al., 2014;Quintana-Rizzo et al., 2015). Oil residues from the DWHOS have persisted in the deep-pelagic habitat of the GoM longer than anticipated. For example, 4-6 years after the DWHOS as found in the water column and in deep-pelagic fishes (Walker et al., 2017;Chanton et al., 2018;Romero et al., 2018). The potential consequences of the DWHOS on deep-pelagic invertebrates are yet to be determined, as exposure to oil-derived toxic compounds has not been previously studied in invertebrates from the pelagic domain of the GoM.
Many deep-sea pelagic organisms undergo diel vertical migrations, playing an important role in the flux of organic carbon between the surface and bathypelagic depths, including the water column and benthic ecosystems (Steinberg et al., 2008;Davison et al., 2013;Hudson et al., 2014;Irigoien et al., 2014;Steinberg and Landry, 2017;Taucher et al., 2018). Deep-sea cephalopod species are of particular interest due to their role in the diets of large fishes and marine mammals (Davis et al., 2007;Young et al., 2010;Romeo et al., 2012;Logan et al., 2013;Lalas and Webster, 2014;Salman and Karakulak, 2019;Southall et al., 2019). They are also voracious predators (Stewart et al., 2014;Corrales et al., 2015;Hoving and Robison, 2017) and are therefore potential vectors for contaminants through the water column (Unger et al., 2008).
Exposure to toxic compounds, such as polycyclic aromatic hydrocarbons (PAHs), is of specific concern in deep-pelagic cephalopods. These toxic compounds are highly lipid-soluble and therefore can cross lipid membranes and bioaccumulate in marine organisms. Mollusks, including cephalopods, have a reduced capacity to metabolize PAHs with higher bioaccumulation potential than other aquatic organisms (Lacoue-Labarthe et al., 2016;Rodrigo and Costa, 2017). In addition, cephalopods have high growth rates, short life spans, and high sensitivity to environmental changes (Rodrigo and Costa, 2017). Cephalopods have a high capacity to bioaccumulate and concentrate contaminants at higher levels than other aquatic groups, thus are potentially useful bioindicators of aquatic contamination over short periods (Gomes et al., 2013;Semedo et al., 2014). During the DWHOS, an unprecedented amount of PAHs was released, with PAHs remaining mostly offshore and contaminating the deep-pelagic domain of the northern GoM (U.S. District Court, 2015;Wade et al., 2016;Romero et al., 2017Romero et al., , 2018. However, comprehensive studies of PAHs in cephalopod tissues are scarce, with little understanding of the influence of natural and biological factors (e.g., diet, migratory behavior, habitat) affecting their storage after major events like the DWHOS.
This study investigated, for the first time, the composition and concentration of PAHs in deep-sea cephalopods of the GoM over a long period, to better understand levels of contamination, interspecies variability, exposure to natural sources, and potential consequences to the deep-pelagic environment. Specifically, samples were collected before and up to 6 years after the DWHOS (pre-spill: 2010, post-spill: 2011, long-term period: 2015-2016). The samples collected before the DWHOS (January to March, 2010), provided an exceptional opportunity to study uptake and persistence of petrogenic PAHs under multiple natural and anthropogenic sources present in the GoM (Mitra et al., 2002;Mitra and Bianchi, 2003;Romero et al., 2016). Overall, this study provides a time-series analysis of PAH composition and concentrations generating trends in PAH signatures for baseline and chronic conditions in deep-sea cephalopods.

Collection of Samples
Samples were collected during three different studies conducted between 2010 and 2016. In all studies, trawl samples were collected within the upper 1500 m of the water column. In 2010, sampling was conducted from January to March aboard the NOAA Ship R/V Pisces for the Sperm Whale Acoustic Prey Survey (SWAPS; see Judkins et al., 2015) using a large midwater trawl. In 2011, sampling was conducted from March to September aboard the R/V Meg Skansi as part of the Offshore Nekton Sampling and Analysis Program (ONSAP) using a 10-m 2 MOCNESS (MOC10) with 3-mm mesh (Judkins et al., 2017). In 2015 and 2016, sampling was conducted in May and August aboard the R/V Point Sur as part of the Deep Pelagic Nekton Dynamics of the Gulf of Mexico (DEEPEND), also using a MOC10. All studies covered similar sites in the northern GoM (Figure 1). The spatial range of these studies extended from 24.4 to 29.3 • N latitude and 86.5 to 90.9 • W longitude (Figure 1). Sampling stations were located in three areas in the northern GoM relative to the Macondo wellhead, similar to previous studies in the region (Judkins et al., 2017;Romero et al., 2018). In the east area, four stations were sampled, located between ∼41 and ∼ 123 km east of the wellhead. In the south area, ten stations were sampled, located between ∼88 and ∼312 km south of the wellhead. In the southeast area, eight stations were sampled, located between ∼120 and ∼297 km southeast of the wellhead. All areas were sampled between 2010 and 2016, except for the southeast area (only during 2010 and 2015-2016).
Taxonomic identification of all samples was highly consistent among studies. All specimens were identified to the lowest taxonomic level possible, by either H. Judkins or M. Vecchione while at sea. Deep-sea cephalopod species that were available for chemical analysis and included in this study are Japetella diaphana (n = 18), Abralia redfieldi (n = 11), Histioteuthis corona (n = 11), Leachia atlantica (n = 21), and Onychoteuthis banksii (n = 13) (Figure 2). PAH exposure and accumulation was studied as PAH concentrations in the mantle tissue. A small piece of mantle tissue was taken from the anterior portion of the mantle and samples were kept frozen at −20 • C until freeze-dried.
Total lipid content (TLC; %lipids) in tissue was calculated gravimetrically from ASE extracts using 100% dichloromethane (Sloan et al., 2004) in samples with sufficient mass available (∼0.5 g) after PAH analysis. For samples with limited mass after PAH extraction, we followed modified gas chromatography method (Hooper and Parrish, 2009) to calculate the %lipid as the total lipid-equivalent fraction by calculating the sum of individual lipids to generate an estimate of total lipids (see Romero et al., 2018, for details).

Analytical Method and Quality Control
We followed modified EPA methods 8270D and 8015C, and QA/QC protocols for the analysis of PAHs. Analyses were carried out on an Agilent 7680B gas chromatograph interfaced with an Agilent 7010 triple quadrupole mass spectrometer (GC/MS/MS). A 30 m RXi-5sil (Restek Corporation, PA, United States) was used with GC liners in the injector in splitless injection mode. UHP helium was used as the carrier gas, while UHP argon gas was used to facilitate the dissociation of the precursor ions (CID) in the collision cell (at 1 mTorr pressure). Inlet temperature was 295 • C, constant flow rate was 1 ml/min, and MS detector temperature was 250 • C. GC oven temperature program was 60 • C for 2 min, 60 • C to 200 • C at a rate of 8 • C/min, 200 • C to 300 • C at a rate of 4 • C/min and held for 4 min, and 300 • C to 325 • C at a rate of 10 • C/min and held for 5 min.
The GC/MS/MS was operated in Multiple Reaction Monitoring mode (MRM) to characterize parent and alkyl PAHs at a high resolution without interferences in the chromatograms. Molecular ion masses for PAHs (precursor and product ions) were selected based on previous studies using GC-MS/MS-MRM (Sorensen et al., 2016;Adhikari et al., 2017;Romero et al., 2018;van Eenennaam et al., 2018). For accuracy and precision of analyses we included laboratory blanks for every 10-14 samples, spiked controls for every 14-18 samples, tuned MS/MS to PFTBA (perfluorotributylamine) daily, checked samples with a standard reference material (NIST 2779) daily, and reanalyzed sample batches when replicated standards exceeded ±20% of relative standard deviation (RSD), and/or when recoveries were low. Recovery of individual PAHs ranged within QA/QC criteria of 50-120%. PAH concentrations are reported as recovery corrected. Each PAH analyte was identified using certified standards (Chiron S-4083-K-T, Chiron S-4406-200-T) and performance was checked using a 5-point calibration curve (0.04, 0.08, 0.31, 1.0 ppm). Quantitative determination of PAHs was conducted using response factors (RFs) calculated from the certified standard NIST2779. The limits of quantification and detection (N = 10) ranged from 0.01 to 0.9 ng/g and 0.001 to 0.3 ng/g, respectively.

Data Analysis
Polycyclic aromatic hydrocarbons concentrations reported in this study are expressed as tissue lipid weight concentrations, and average values are shown as arithmetic mean ± standard error. Total PAH concentrations were calculated as the sum of 2-ring to 6-ring PAHs and their alkylated homologs. Low molecular weight (LMW) PAHs are the sum of all 2-3 ring PAHs (including alkylated homologs), and high molecular weight (HMW) PAHs are the sum of 4-6 ring PAHs (including alkylated homologs). To assess sources of PAHs in the tissue samples, selected diagnostic ratios were calculated: (1) FL/Pyr ratio; (2) % Retene; and (3) pyrogenic index [PI; (other 3-6 ring EPA priority PAHs)/ (5 alkylated PAHs)]. For the pyrogenic index, the five alkylated PAHs are the alkylated compounds of: N, P, D, F, and C. The "other" 3-6 ring EPA priority PAHs are: ACL, ACE, AN, FL, PY, BAA, BBF, BKF, BAP, ID, DA, and BGP (Wang and Fingas, 2003;Fitzgerald and Gohlke, 2014;Romero et al., 2015Romero et al., , 2018. These PAH ratios include isomer pairs with similar adsorption and dissolution properties abundant in different sources.
Statistical analyses were conducted on log-transformed data to approach normal distribution. Differences in mean concentrations with respect to years and species were tested by one-way ANOVA followed by Tukey's HSD test. Significance was set at p < 0.05. We used JMP 12.1 for Mac (SAS Institute Inc., Cary, NC, United Sates, 1989-2007.

Species Variability
Large variability among cephalopod species was found for %lipid content and PAH concentrations over the studied period (Figure 3). For example, %lipid content varied significantly among the species studied, although a similar temporal trend was observed with lower %lipid content in 2011 (post-spill) for all species sampled in this year (p < 0.05; J. diaphana: 4.0 ± 0.8%; A. redfieldi: 7.2 ± 1.6%; H. corona: 11.8 ± 1.3%; L. atlantica: 5.6 ± 0.6%; and O. banksii: 6.1 ± 0.5%; Figure 3). Lipid loss in 2011 relative to 2010 (pre-spill) was about 50% in J. diaphana and L. atlantica, and about 25% in A. redfieldi and H. corona. An increase in %lipid content was observed in 2015-2016 for all species sampled in the long-term period (Figure 3).
Due to the differences found in %lipid content among the species studied, PAH concentrations were calculated as lipid-normalized concentrations (µg PAH/g lipid weight). PAH concentrations were lower post-spill for J. diaphana, A. redfieldi, and L. atlantica, with 16%, 56%, 34% decrease in PAH concentration in 2011 relative to 2010, respectively (Figure 3). Histioteuthis corona showed lower PAH concentrations in 2010 (pre-spill) and 2011 (post-spill), with a ∼84% increase in concentration during 2015-2016 (long-term period; Figure 3). In contrast, O. banksii showed similar PAH concentrations between the pre-spill and long-term periods (Figure 3). The temporal trend in PAH concentration observed in most species is not explained by %lipid content because lipid-normalized PAH data removes the influence of interspecies %lipid content. Also, we found that PAH concentrations were not correlated with mantle length in all cephalopod species (p > 0.05), except H. corona (R 2 = 0.011; p < 0.01; Table 1). However, significant differences of PAH concentrations between cephalopod life stages were not found in any studied year for any species, including for H. corona (p > 0.05; Table 1). Therefore, the temporal trend in PAH concentration for H. corona is not explained by the different lengths of animals collected among the years.
A similar abundance of 3-ring (76 -81% phenanthrene compounds, including alkyl homologs) and 4-ring (11 -17% pyrene and fluoranthene compounds, including alkyl homologs) PAHs were observed among the species studied (Figure 4). However, small changes in the composition of PAHs were detected between the periods covered in this study (Figure 4 and

Spatial Variability of PAHs
In 2010 (pre-spill), significantly lower PAH concentrations were observed in the south area relative to the other areas (p < 0.001; east: 15.9 ± 6.4 µg/g lipid; south: 4.5 ± 3.7 µg/g lipid; southeast: 18.9 ± 4.7 µg/g lipid). In contrast, in 2015-2016 (long-term period), PAH concentrations were not significantly different among the areas studied (p > 0.05). Only the south area showed significant differences in PAH concentrations in the tissue samples among years (p < 0.001), with an increase of PAHs over time (Supplementary Figure S2). The composition of PAHs changed post-spill (2011) and returned during the longterm period to conditions similar to pre-spill (Supplementary Figure S3). These results indicate that the temporal trend observed in the concentration and composition of PAHs is independent of the area sampled in each time period.

General Temporal Trends in PAH Levels and Composition
Even though a large variability was observed among the cephalopod species studied, all species indicate a higher exposure to 2-3 ring PAHs after 2010 (Figure 4). By integrating the PAH data from all species together, the general temporal trend observed indicates a ∼2-fold increase in PAH concentrations during the long-term period relative to samples collected prespill for the deep-sea cephalopod assemblage in the northern GoM (2010: 13.5 ± 3.0 µg/g lipid weight; 2011: 11.5 ± 1.7 µg/g lipid weight; and 2015-2016: 29.0 ± 6.6 µg/g lipid weight). Significantly higher concentrations were consistently observed in the long-term period relative to the pre-spill and post-spill periods (p < 0.01; Figure 5). Collectively, the trends observed in PAH composition and calculated ratios suggest an episodic exposure to petrogenic PAHs that occurred after the post-spill period (2011) and continue through 2016 (p < 0.01; Figures 6, 7).
In 2010 (pre-spill), higher ratios denote a more pyrogenic origin for PAHs (

DISCUSSION
Interspecies differences were observed in lipid content and PAH concentrations, suggesting biological factors (e.g., diet, habitat, life span, behavior, and lipid metabolism) play an important role in the bioaccumulation of PAHs in the mantle tissues ( Table 1).
The species examined in this study are different from one another in many physical and ecological aspects (Supplementary Table S1). Abralia redfieldi is a small muscular squid (collected sizes: 9-35 mm ML), living primarily in the mesopelagic and migrating to the epipelagic zone nightly (Judkins and Vecchione, this vol). Onychoteuthis banksii is also a muscular squid, but larger (collected sizes: 14-326 mm ML) and found at depth during the day and migrating to the upper mesopelagic and epipelagic zones nightly. In contrast, J. diaphana is a gelatinous octopod (collected sizes: 9-105 mm ML), found throughout the water column from the epi-to the bathy-pelagic zones; it exhibits ontogenic descent (i.e., larger individuals tend to live deeper in the water column than smaller ones). Leachia atlantica is a glass squid species (collected sizes: 36-111 mm ML) that uses an ammonia-filled chamber to achieve neutral buoyancy and it is found throughout the water column with no evidence of ontogenic descent (Judkins and Vecchione, this vol). Very different from the other species studied, H. corona is a large squid with ammoniacal muscle tissue (collected sizes: 17-71 mm ML; max known size: ∼190 mm ML) that does not vertically migrate daily, primarily inhabiting the mesopelagic zone. Altogether, these multiple ecological differences might affect lipid content and PAH concentrations among the species studied (Figure 3), as species are exposed to different environments in the water column. For example, H. corona and J. diaphana do not migrate vertically every day to the surface (∼50 m). Therefore, they may have been exposed to different hydrocarbon mixtures formed in the water column after released at ∼1500 m depth containing distinct chemical compositions (Ryerson et al., 2012). Also, biological parameters such as lipid metabolism may have an essential role in the variability of PAH bioaccumulation among cephalopod species. It seems that in some shallowerliving cephalopod species, the digestive gland has an important role in detoxification of organic pollutants, including PAHs (Rodrigo and Costa, 2017). The physiological and molecular mechanisms involved in detoxification processes for organic pollutants in deep-sea cephalopod species are not known. It seems likely that the pathways for detoxification of organic pollutants may be different among the deep-sea cephalopod species studied as shown by the PAH concentrations in the mantle tissues (Figure 3). Interestingly, the temporal trends observed in this study were all similar among the deep-sea cephalopod species studied. For example, our results indicate a shift in the composition of PAHs in all cephalopod species in 2011 (post-spill) relative to 2010 (January to March; pre-spill) that continue to 2015-2016 (long-term period; Figures 4, 6), supporting previous studies that suggest exposure and uptake of oil-derived PAHs from the DWHOS occurred in offshore organisms (Murawski et al., 2014;Snyder et al., 2015;Romero et al., 2018). There are multiple sources of PAHs in the northern GoM (seeps, riverine discharges, continental runoff, coastal erosion, atmospheric deposition, and oil and gas exploration) (Park et al., 2001(Park et al., , 2002Mitra and Bianchi, 2003;Tansel et al., 2011), with seeps as the most abundant among those sources (Ocean Studies Board and Marine Board, 2003). However, the DWHOS released an amount over seven times the average annual input of oil into the GoM, becoming the most important source of PAHs in summer to fall of 2010 (Murawski et al., 2014). As expected, most of the change in the composition of PAHs occurred for the low-molecular weight compounds (2-3 rings), abundant in oil that weathered and mixed over time with other sources (e.g., Mississippi river, natural seeps), as observed in 2015-2016. This inference is supported by similar abundances of retene between pre-pill and post-spill periods (Figure 7) indicating that natural sources of PAHs in the GoM do not explain the temporal trend observed in the composition of PAHs in the deep-sea cephalopod species studied.
The temporal trend for lipid content and PAH concentrations were also similar among the cephalopod species with significant changes over the 6-years of this study (Figure 3). However, the temporal trends observed in Figure 3, were not as expected for a one point-source contamination event, like the DWHOS   pre-spill: 2010, post-spill: 2011, and long-term: 2015-2016. Graph shows shaded boxes as the interquartile ranges, with horizontal lines indicating median values and whiskers representing 10th and 90th percentile. Blue circles denote mean values, and white circles denote outliers. Data are shown as mean ±SE.
2011 PAH composition shows a shift in the source of PAHs but concentrations decreased, as well as lipid content, indicates that the temporal trends observed are secondary effects of the exposure to oil-residues from the DWHOS. We hypothesize that the lower lipid content in the cephalopod species observed in 2011 (post-spill) may be due to a change in dietary intake (decrease in prey availability) and/or dietary quality (e.g., fatty acids decrease in prey). Stable-isotope analysis of J. diaphana, H. corona, O. banksii and L. atlantica (Staudinger et al., in press, Richards et al., unpublished data) indicate that these species occupy multiple trophic positions, inhabiting a variety of ecological niches within the water column, all potentially affected by the DWHOS. Several studies demonstrated carbon from the DWHOS entered the planktonic and mesopelagic food web after the spill (Chanton et al., 2012;Quintana-Rizzo et al., 2015) supporting the idea that potential prey species from different trophic levels were affected by oilresidues, which may have influenced the dietary intake of cephalopods. Moreover, several studies in shallower species, showed that dietary inputs can severely affect lipids, which are critical in cephalopod development (Navarro and Villanueva, 2003;Almansa et al., 2006;Guinot et al., 2013;Reis et al., 2016). Due to the hydrophobic properties of PAHs, significant changes in lipid content may also affect bioaccumulation of PAHs. The exception observed in the temporal trend for PAH concentrations in H. corona suggests other factors are important for bioaccumulation of PAHs in cephalopods. In this species, lower lipid content was observed post-spill (2011), albeit similar PAH concentrations pre-and post-spill (2010 and 2011), and 84% higher concentrations in the long-term period (2015)(2016). Of all the species studied, H. corona is the only one that lives in the mesopelagic zone and does not migrate to the surface to feed (Supplementary Table S1). Oil-residues have been observed in the water column years after the spill (Walker et al., 2017), probably from resuspension of contaminated sediments (Romero et al., 2017;Diercks et al., 2018). It is possible that the dietary intake for H. corona was affected like the other species studied (Figure 3), but with an extended period of exposure to oilresidues persistent at depth. This is supported by elevated PAH concentrations observed in the long-term period (2015)(2016).
Collectively, PAH composition trends in the deep-sea cephalopod species studied suggest an episodic exposure to petrogenic PAHs that occurred after the pre-spill sampling campaign in 2010 and continued through 2015-2016. Exposure to oil-derived PAHs occurred after the DWHOS, but secondary effects, including potential trophic-web effects, may have influenced the lipid content and PAH levels post-spill (2011). It appears that exposure to oil-derived residues is longer than expected in the pelagic deep-sea environment, as observed from the same event in mesopelagic fishes (Romero et al., 2018). This phenomenon may have disturbed the deep-sea environment via the food web, as indicated by the temporal trend observed in cephalopod species. Due to the important role of lipids on the physiology of cephalopods (Reis et al., 2016), future studies should evaluate lipid composition in relation to dietary intake and organic contaminants. Also, studies should cover longer periods for a better understating of the persistence of organic pollutants at depth and impacts to pelagic communities from different trophic levels.

ETHICS STATEMENT
The animal study was reviewed and approved by the IACUC USF.

AUTHOR CONTRIBUTIONS
IR carried out the laboratory and data analyses, and wrote the manuscript including figures and tables. HJ and MV collected the samples and identified specimens to species levels, as well as contributing to the manuscript by revising, editing, and adding references and comments.

ACKNOWLEDGMENTS
The authors would like to thank the crew of the R/V Pisces, R/V Meg Skansi, and R/V Point Sur for their help during the field program, T. Sutton for support and information about ONSAP and DEEPEND programs, D. Fenolio for images, and the technicians Thea Barlett, Hannah Hamontree, and Olivia Traenkle for laboratory support.