Mercury stable isotopes reveal sources of methylmercury and prey in giant Pacific bluefin tuna from the western North Pacific Ocean

Sources of methylmercury (MeHg) in adult Pacific bluefin tuna (Thunnus orientalis, PBT) from the western North Pacific Ocean (WPO) were examined using mercury stable isotopes. Significant increases in δ202Hg and Δ199Hg values with PBT size and age, along with those of potential prey, indicate a shift in the source of MeHg accumulated by PBT as they age. Among adults from the WPO, this shift likely involves greater accumulation of MeHg from epipelagic prey in the Kuroshio extension in large vs. small and medium‐sized PBT. For all adults, little MeHg is accumulated in the spawning grounds near Taiwan. Significantly lower Δ199Hg/Δ201Hg ratios in adult PBT and their prey from the WPO than from the central and eastern North Pacific indicate different sources or transformations of MeHg prior to accumulation in the WPO food web than further east. Our results show that MeHg sources to oceanic food webs vary across the North Pacific Ocean and regionally within the WPO.

Bluefin tuna are a long-lived species with broad migration ranges and a potentially useful bio-integrator of ocean basin-wide food web accumulation of monomethylmercury (MeHg), a neurotoxin that biomagnifies in marine food webs (Morel et al. 1998;Lee et al. 2016;Schartup et al. 2019;Tseng et al. 2021). Levels of total mercury (THg, present as almost 100% MeHg) in bluefin tuna may be affected by spatially variable prey abundances (Polovina 1996;Portner et al. 2022), prey MeHg concentrations (Choy et al. 2009;Ferriss and Essington 2011;Madigan et al. 2018;Houssard et al. 2019), and oceanographic conditions (Kitagawa et al. 2002;Kitagawa et al. 2004;Kitagawa et al. 2007;Furukawa et al. 2014). In addition, different life-histories across age classes may also influence THg levels within bluefin tuna populations (Itoh 2006). Although all Pacific bluefin tuna (Thunnus orientalis, PBT) are spawned in the western Pacific Ocean (WPO), a fraction of PBT migrate to the eastern Pacific Ocean (EPO) at 1-3 yr old, and return to the WPO by age 7 presumably to spawn (Boustany et al. 2010;Madigan et al. 2014;Tawa et al. 2017;Shiao et al. 2020). While the exact oceanic areas where adult PBT feed and accumulate MeHg are still unknown, proximity to eastern Asia likely plays a role in the elevated THg concentrations found in commercially harvested adult PBT and other top predators from the WPO (Colman et al. 2015;Tseng et al. 2021;Médieu et al. 2022).
Naturally occurring mercury stable isotope ratios in fish tissue provide useful information about the sources and transformations of MeHg prior to entering marine food webs. In the sunlit surface ocean, large increases of mass dependent fractionation (MDF; denoted by δ 202 Hg) and odd mass independent fractionation (MIF; denoted by Δ 199 Hg) of dissolved MeHg in seawater result from the production of volatile elemental Hg during abiotic photochemical reactions , as well as photo-microbial demethylation in phytoplankton (Kritee et al. 2018). As a result of the vertical attenuation of photochemically-driven MIF in the ocean, marine fish from epipelagic depths (0-200 m) have higher Δ 199 Hg values than fish from mesopelagic (200-1000 m) depths Sackett et al. 2017;Madigan et al. 2018). Divergent δ 202 Hg values in oceanic species with similar Δ 199 Hg values reflect variations in the types or extents of non-photochemical transformations (e.g., dark microbial reduction and demethylation) that control Hg MDF in different ocean ecosystems (Madigan et al. 2018;Motta et al. 2020a). In addition, the MIF ratio of Δ 199 Hg/Δ 201 Hg is diagnostic of the dominant photochemical Hg reaction pathways (photoreduction of inorganic Hg, photodegradation of MeHg) prior to incorporation at the base of the marine food web (Zheng et al. 2015;Blum et al. 2020). To examine the sources of Hg in PBT from the WPO, we analyzed Hg stable isotope ratios in adult (≥ 5-yr old) PBT from the WPO with respect to PBT age/size, Hg stable isotope ratios in potential PBT prey, and pathways of Hg transformation before entering the marine food web.

Sample information and preparation
Detailed sampling locations and analytical methods of PBT are described elsewhere (Tseng et al. 2021). In brief, we analyzed 91 adult (≥ 5-yr old) PBT caught in their primary spawning grounds of the Philippine Sea, southeast of Taiwan from April to July 2017. Based on prior bluefin tuna feeding studies (Shimose and Wells 2015;Madigan et al. 2016), potential epi-and mesopelagic prey fish were collected from Taiwan fish markets in 2019. Samples of dorsal muscle tissue were collected from all fish. Information about depth range, sex, and fork length were recorded upon collection and can be found online (He et al. 2022). Ages of PBT were estimated using a fork length based empirical model (Shiao et al. 2017). All samples were freeze dried for 96 h and stored in 10 mL borosilicate glass vials with screw caps at À20 C.
Acid digest solutions used for THg determination were subsequently used for mercury isotope analysis. Based on the concentration of THg, analytes were diluted to 0.5-2 ng mL À1 . A multiple collector inductively coupled plasma mass spectrometer (MC-ICP-MS; Thermo Neptune) in the Department of Earth and Planetary Sciences at Rutgers University was used for Hg isotope ratio measurement. Isotopic compositions were reported as per mil (‰) deviations from the average isotope ratios of NIST SRM 3133 bracketing standards using delta notation following Eq. 1: where xxx is the mass of each mercury isotope between 199 Hg and 204 Hg. We use δ 202 Hg to represent isotopic differences due to MDF. MIF is reported in Δ xxx Hg, which quantifies the difference between measured δ xxx Hg and theoretical MDF predicted based on δ 202 Hg: where xxx is the mass of each Hg isotope 199, 200, and 201 and β is a constant (0.2520, 0.5024, and 0.7520, respectively) . Accuracy of measurement was monitored by running several UM-Almadén secondary Hg standards (δ 202 Hg = À0.53 AE 0.08‰ 2SD, Δ 199 Hg = À0.03 AE 0.07‰ 2SD, n = 21) and IAEA-436 reference material (δ 202 Hg = 0.63 AE 0.19‰ 2SD, Δ 199 Hg = 1.51 AE 0.11‰ 2SD, n = 11).

Statistical analysis
Adult PBT were grouped into small, medium, and large size classes by arbitrarily dividing the total range of fork lengths (84 cm) into three groups spanning 28 cm each. This resulted in groups of small (182-210 cm, 5-9 yr), medium (210-238 cm, 9-16 yr), and large (238-266 cm, > 16 yr) PBT with sufficient numbers of individuals (n = 40, 23, and 28, respectively) to support statistically meaningful comparisons among size classes (Table S1). To identify the ecological and environmental factors that may account for differences in Hg stable isotope ratios among PBT size classes, we compared Hg stable isotope values for PBT and their prey from the EPO and WPO (Madigan et al. 2018), and with those for other oceanic fish from the North Pacific Subtropical Gyre of the central North Pacific Ocean (CPO) . We also compared these to new measurements of Hg stable isotope ratios in potential prey fish from waters east of Taiwan. Significance of difference of THg, δ 202 Hg, and Δ 199 Hg between size groups was assessed using Welch ANOVA and tested with Games-Howell post hoc test. Slopes of Δ 199 Hg/Δ 201 Hg and Δ 199 Hg/ δ 202 Hg were calculated by York Regression and Deming regression, respectively (York 1968). All statistical tests were performed on R version 4.0.2 (R Core Team 2020).

Mercury concentrations and stable isotope ratios in adult PBT
THg concentrations in muscle tissues of adult (5 to more than 27 yr old) PBT from the WPO ranged from 0.80 to 4.96 mg g À1 (reported on a wet weight basis throughout) and were positively correlated with fork length (R 2 = 0.44, Pearson's p < 0.001, n = 91). This range of concentration is 2.7-17 times the U.S. EPA human consumption advisory level (0.3 mg g À1 ). Male adult PBT had significantly higher fork lengths and body weights than females (Mann-Whitney U test, p < 0.001). However, since there were no significant differences in THg, δ 202 Hg or Δ 199 Hg between male and female PBT of similar size in our samples, results for fish of both sexes were pooled for further analysis. The observed increase in Hg concentration with age in PBT (Fig. 1A) was reported previously (Tseng et al. 2021), and is consistent with observations for other marine predatory fish (Storelli et al. 2005;Lee et al. 2016), and the slow turnover of Hg in PBT (Kwon et al. 2016) since for MeHg in fish, dietary assimilation exceeds growth dilution (Amlund et al. 2007;Cossa et al. 2012). This increase continues a trend of increasing THg from low levels in juvenile (< 2 yr) PBT from coastal California waters of the EPO (Madigan et al. 2018) to higher concentrations in adults from the west. Similar to the age-dependent increase in THg concentration, δ 202 Hg and Δ 199 Hg values increased from juvenile to small PBT, and from small and medium to large adult PBT (Figs. 1B,C, S1), indicating a shift in the source of Hg accumulated by PBT as they age.

Sources of Hg to adult PBT
Assuming negligible fractionation during trophic transfer, Hg stable isotope ratios in PBT should correspond to those in their prey (Kwon et al. 2016). Our δ 202 Hg and Δ 199 Hg results for adult PBT suggest that prey from the Kuroshio extension east of Japan is the most important source of Hg for adult PBT from the WPO. Prey from the Kuroshio extension have δ 202 Hg values that are similar to those of adult PBT, but much higher than δ 202 Hg values of prey with similar extents of odd MIF (Δ 199 Hg) from the CPO, EPO and east of Taiwan (Fig. 2). Preliminary C and N stable isotope mixing model results indicate that the adult PBT from our study fed mostly ($ 80%) on prey from east of Japan and the Kuroshio-Oyashio transition region (F.-J. Lee and Y.-K. Yeh, pers. comm.), consistent with previous results (Kitagawa et al. 2002(Kitagawa et al. , 2004Madigan et al. 2016). We could not run a Hg isotope mixing model because many adult PBT from the WPO had higher δ 202 Hg values than any prey item.
Large positive Δ 199 Hg values in marine biota only result from photochemical transformations of MeHg in the surface mixed layer (SML) Motta et al. 2019). As in previous studies across the North Pacific Ocean Sackett et al. 2017;Madigan et al. 2018;Motta et al. 2020b), Δ 199 Hg values in prey fish from east of Taiwan decreased with depth ( Fig. S2), reflecting a depth-dependent decline in the photochemical transformation of Hg entering the food web. However, much lower δ 202 Hg values in prey from east of Taiwan than in adult PBT from the WPO indicate that adult PBT accumulate very little Hg from their breeding grounds east of Taiwan, consistent with diet tracking studies, which showed minimal feeding by adult PBT in waters east of Taiwan .
Higher Δ 199 Hg values in large vs. small and medium adult PBT from the WPO indicates an increase in the accumulation of MeHg from specifically epipelagic prey by adult PBT as they grow. Such a shift in Hg source could result from an ontogenetic change in diet or feeding location. Adult PBT from the WPO apparently switch from consuming lower trophic level prey such as sardine and anchovy from east of Japan as smallsized adults to consuming higher trophic level prey such as pomfret, mackerels, and squids from further east in the Kuroshio-Oyashio transition as large adults . Alternatively, environmentally-driven changes in prey availability could lead to a change in Hg sources to adult PBT as they age. For example, the shift from warm to cool phase in the Pacific Ocean between 2013 and 2014 resulted in reduced anchovy and sardine stocks in the WPO after 2014 (Kuroda et al. 2020;Watanuki et al. 2022). A lack of preferred epipelagic prey after 2014 may have resulted in a decreased accumulation of Hg from epipelagic forage fish, as indicated by lower δ 202 Hg and Δ 199 Hg values in small and mediumsized compared to large adult PBT. Gut content studies of variously sized adult PBT collected during this period would be needed to evaluate this hypothesis.
Across the North Pacific Ocean, Δ 199 Hg/δ 202 Hg slopes decline from 2.16 AE 0.13 (1SE) in juvenile PBT from the EPO to 0.90 AE 0.07 (1SE) in adult PBT from the WPO. Since the Δ 199 Hg/δ 202 Hg slope for adult PBT is lower than that of available prey from the WPO (1.50 AE 0.02, Kuroshio east of Japan; 2.13 AE 0.16, east of Taiwan), adult PBT likely accumulate MeHg from other prey with lower Δ 199 Hg/δ 202 Hg slopes. In vivo demethylation in fish may raise δ 202 Hg values without affecting Δ 199 Hg (Kwon et al. 2013). However, no increase in δ 202 Hg was observed in 1-9 yr old PBT (Kwon et al. 2016), so unless in vivo demethylation increases significantly as PBT age, it would not account for the observed decrease in Δ 199 Hg/δ 202 Hg slope.
We further examined sources of MeHg to adult PBT using Δ 199 Hg/Δ 201 Hg slopes. The overall Δ 199 Hg /Δ 201 Hg slope for adult PBT (1.09 AE 0.01 1SE, York regression; Figs. 3, S3) did not vary among size classes (Table S2) and is similar to that for vertically grouped prey from the Kuroshio extension (1.12 AE 0.04) (Madigan et al. 2018) and the Kuroshio current  (Shiao et al. 2020). N = 24, 40, 23, and 28 for each group, respectively. Letters indicate significant differences between groups tested using Welch ANOVA followed by Games-Howell post hoc test (p < 0.05). Gray dashed line corresponds to EPA safety limit guidelines of 0.3 mg kg À1 of wet weight. Analytical uncertainties (2SD) of δ 202 Hg and Δ 199 Hg were 0.08‰ and 0.07‰, respectively, based on the analysis of UM-Almadén secondary Hg standards (n = 21), and 0.19‰ and 0.11‰, respectively, based on the analysis of IAEA-436 Tuna Fish Homogenate reference material (n = 11).  (Madigan et al. 2018), adult PBT from the western North Pacific Ocean (pink, orange and red ellipses; this study, Fig. S1), and various prey and predator fish species from across the North Pacific Ocean (symbols). Prey and predator fish were captured from the coast of California (Madigan et al. 2018) (EPO), the central North Pacific Subtropical Gyre (CPO) , the Kuroshio Current east of Japan (KEJ, WPO) (Madigan et al. 2018), and east of Taiwan (EOT, WPO, this study). Ellipses corresponding to PBT were drawn at a confidence level of 0.95, with colors corresponding to the size classes in Fig.1. Symbol shading represents epipelagic (non-filled) or mesopelagic (filled) species. Prey values are means AE 1SD for each species (n = 2-4). Blue and red lines show Δ 199 Hg/δ 202 Hg linear regression slopes for juvenile PBT from the EPO (2.16 AE 0.13 1SE) and adult PBT from the WPO (0.90 AE 0.07 1SE, all size classes combined), respectively. east of Taiwan (1.13 AE 0.04) (Fig. 3). In contrast, Δ 199 Hg/ Δ 201 Hg slopes are significantly higher in vertically grouped prey fish from the CPO (1.20 AE 0.01) and the EPO (1.22 AE 0.09). Yellowfin and bigeye tuna from the CPO have an even higher Δ 199 Hg/Δ 201 Hg slope (1.34 AE 0.09, n = 6) compared to lower trophic level fish from the North Pacific as do juvenile PBT from the EPO (1.37 AE 0.06) (Madigan et al. 2018). Such high Δ 199 Hg/Δ 201 Hg slopes, which have been observed in marine birds from polar regions (Point et al. 2011;Zheng et al. 2015) and in some freshwater fish Perrot et al. 2012;Sherman and Blum 2013;Lepak et al. 2018), are consistent with the overwhelming dominance of the photodegradation of MeHg in controlling odd-MIF of Hg accumulated in the oceanic food webs of the eastern and central North Pacific.
The increase in Δ 199 Hg/Δ 201 Hg slopes in pelagic fish from west to east across the North Pacific indicates that Hg in fish from the WPO originates from a different source or is transformed by a different set of processes than Hg in fish from the CPO and EPO. A possible explanation of this observation is a shift in the dominance of photoreduction of iHg with a Δ 199 Hg/Δ 201 Hg slope ≤ 1.1 (Zheng and Hintelmann 2010;Motta et al. 2020b) in the west to photodegradation of MeHg with a Δ 199 Hg/Δ 201 Hg slope = 1.2 Kritee et al. 2018) in the east as the primary factor setting Hg isotope signatures at the base of the marine food web. Low Δ 199 Hg/Δ 201 Hg slopes ($ 1.0) and relatively low Δ 199 Hg values (< 1.5‰) in fecal deposits and tissues of marine animals (Zheng et al. 2015;Masbou et al. 2018;Meng et al. 2020) have been attributed to the accumulation of MeHg produced below the SML where photochemical degradation of MeHg is limited by low irradiance Motta et al. 2019).
However, large extents of odd MIF associated with the photoreduction of iHg are unlikely to occur in the open ocean, where dissolved iHg is dominated by photochemically less reactive chloro-complexes (Yin et al. 2018). In addition, since photochemical reduction of iHg complexed by thiols or inside phytoplankton cells results in a Δ 199 Hg/Δ 201 Hg slope of 1.2 (Zheng and Hintelmann 2010;Kritee et al. 2018;Motta et al. 2020b), this could not explain the observed Δ 199 Hg/ Δ 201 Hg slopes in PBT from the WPO. Other sources of iHg previously imprinted with positive odd MIF (Δ 199 Hg) may serve as the substrate for Hg methylation in the WPO. Wet deposition of iHg through atmospheric precipitation could be an important source of positive odd Hg MIF, which has an overall global mean Δ 199 Hg of 0.41‰ (Yin et al. 2018;Jiskra et al. 2021). Particle bound Hg from coastal marine boundary layers in the Northern Hemisphere, which has Δ 199 Hg values of 0.35-0.4‰ (Rolison et al. 2013;Fu et al. 2019;Yu et al. 2020;Qiu et al. 2021;Sun et al. 2021), may also be an important proximal source of iHg with positive odd MIF to the WPO. Note that the Δ 199 Hg/Δ 201 Hg slopes of both oceanic precipitation and particle bound Hg in the atmosphere from coastal regions are $ 1.0, similar to that in adult PBT from the WPO. While the odd MIF of Hg entering marine ecosystems could be subject to transformations in the lower atmosphere (Estrade et al. 2009;Zheng and Hintelmann 2010;van Groos et al. 2014;Sun et al. 2016 Recent studies show that high concentrations of THg in tunas from the WPO may be linked to anthropogenic inputs (Colman et al. 2015;Médieu et al. 2021;Tseng et al. 2021). Higher δ 202 Hg values than expected with respect to Δ 199 Hg as we observed in PBT may also be related to an anthropogenic source of Hg to the WPO. East Asia, including mainland China, has been widely implicated as a major contributor of Hg to oceans globally (Liu et al. 2016;Kim et al. 2017Kim et al. , 2019. Anthropogenic Hg inputs are particularly important in coastal zones. Industrial and riverine inputs of Hg in sediments of the marginal seas of China and Korea have δ 202 Hg values that are 1-2‰ higher than that of atmospherically-sourced Hg (Yin et al. 2018;Meng et al. 2019;Jung et al. 2022). However, dark microbial transformations could also enrich 202 Hg in the residual pool of atmospherically deposited Hg in coastal marine environments (Kritee et al. 2007(Kritee et al. , 2009). While it is clear that anthropogenic Hg enters coastal waters of east Asia, the physical and/or biological processes by which it may be entrained into oceanic marine food webs leading to PBT in the WPO, have yet to be examined.
Our analysis of Hg stable isotopes clearly shows that the source of MeHg in adult PBT from the WPO is distinct from that in PBT and other tunas in the EPO and CPO. Epipelagic prey from the Kuroshio extension east of Japan is the most important source of MeHg for adult PBT from the WPO and becomes increasingly so, as fish age. However, very limited MeHg is accumulated by adult PBT from prey consumed in their spawning grounds east of Taiwan. Despite these findings, our ability to trace MeHg through the oceanic food web of the WPO to adult PBT is limited by the availability of representative prey from the entire migration range of PBT, including the Kuroshio-Oyashio transition region, an important feeding ground for these fish (Kitagawa et al. 2002(Kitagawa et al. , 2004Madigan et al. 2016). Connecting MeHg accumulated in adult PBT from the WPO to specific anthropogenic sources and prey will require higher resolution observations of Hg stable isotopes in coastal and oceanic air, water and biota.