Methylmercury produced in upper oceans accumulates in deep Mariana Trench fauna

MMHg is a neurotoxic compound that accumulates in marine food webs, exposing humans primarily through seafood consumption. Mercury (Hg) enters oceans via atmospheric deposition and terrestrial discharge. Within the ocean, inorganic Hg(II) can be reduced to gaseous Hg(0) or converted by microbes to methylated forms (MMHg and DMHg). The biological pump transports particulate organic matter (POM) and scavenged Hg from the surface to depth, producing macronutrient-like vertical Hg distributions with low surface concentrations and elevated levels in mesopelagic and deep waters. Current paradigms suggest MMHg is produced in situ primarily in oxygen-depleted mesopelagic waters during remineralization of sinking POM, with some production also occurring in oxygenated surface waters. Because MMHg in surface waters is readily photodegraded, concentrations are typically depleted in the epipelagic. Despite limited downward transport via downwelling and particles inferred by field data and models, deep oceans contain abundant MMHg and DMHg, leaving the origin and cycling of deep-ocean methylated Hg uncertain. Hg isotopes display mass-dependent fractionation (MDF) and notable mass-independent fractionation (MIF), particularly odd-MIF induced by photochemical transformations, providing tracers of MMHg sources and processing that are largely preserved through bioaccumulation. The study aims to determine the origin of MMHg in hadal ecosystems of the Mariana and Yap Trenches by analyzing Hg concentrations and isotope signatures in endemic amphipods and sediments, testing whether deep-ocean MMHg is produced in situ or derived from upper-ocean sources transported to depth.

Prior work shows: (1) MMHg formation in mesopelagic oxygen minimum zones during microbial remineralization of sinking POM, with additional documented production in epipelagic waters; (2) strong photodegradation of MMHg in surface waters leading to low epipelagic concentrations; (3) limited efficiency of downward transport of mesopelagic MMHg by downwelling or sinking POM inferred from observations and models, despite abundant MMHg/DMHg observed in deep waters; (4) Hg stable isotopes exhibit MDF and odd-MIF during photochemical reactions, while trophic transfer introduces little isotopic fractionation, making biota isotopes robust tracers; (5) even-MIF (Δ200Hg, Δ204Hg) typically reflects atmospheric processes and can track atmospheric Hg inputs. These insights motivate using isotope fingerprints in deep fauna to resolve whether deep MMHg is produced locally or supplied from upper oceans.

Samples: Endemic amphipods (mostly Hirondellea gigas; one Alicella gigantea) were captured with baited funnel traps on deep-sea landers (@Tianya, @Yuanwei, @Wanquan) deployed to 7000–11,000 m below surface in the Mariana (11.5°N, 142.5°E) and Yap (9.5°N, 138.5°E) Trenches during multiple cruises (July 2016–March 2017). A snailfish (primary diet: amphipods) was also collected in Yap (~8000 m). Amphipods were rinsed, heads removed to avoid bait contamination, freeze-dried, homogenized, and in several cases separated into muscle, lipid, and gut contents. To enhance representativeness, individuals of similar sizes were composited. Seafloor surface sediments (top 0–6 cm) were collected by box corer at 5500–9200 m, subsampled into short cores, frozen, freeze-dried, and homogenized. Analytical measurements: Total Hg (THg) in solid samples was measured by Lumex RA-915F (PYRO-915+ combustion; Zeeman AAS). THg in digests for isotope analysis was measured by Tekran 2600 CV-AFS (US EPA 1631E). MMHg in amphipods/snailfish was quantified by Tekran 2700 GC–CVAFS after KOH/MeOH extraction and NaBEt ethylation in closed purge vessels. QA/QC used CRMs (GBW07405, DORM-4); 2 SD uncertainties: <5% for THg (Lumex and Tekran 2600), <15% for MMHg; duplicates RSD <5% (THg), <8% (MMHg). Isotope ratios: Amphipods (~0.2–0.6 g) were microwave-digested (HNO3 + H2O). Sediment Hg was preconcentrated by combustion-trapping into acid (2HNO3:1HCl). Blanks <1% of sample Hg; recoveries 88–110%. Hg isotope ratios were measured by MC-ICPMS (Nu Plasma 3D, Tianjin University) with customized cold vapor introduction. Mass bias corrections used internal Tl (NIST 997) and standard-sample bracketing with NIST 3133, matched within 5% in matrix and Hg concentration. All seven Hg and two Tl isotopes were collected on Faraday cups; typical sensitivity ~2 V per ng g−1 202Hg; acquisition 5 blocks × 20 cycles (4.2 s). Washout between samples limited carryover to <1%. Isotope notation: δ202Hg relative to NIST 3133; MIF as ΔxxxHg = δxxxHg − αxxx × δ202Hg, with α199=0.2520, α200=0.5024, α201=0.7520, α204=1.4930. Secondary standards (NIST 3177) and CRMs (GBW07310; DORM-4) verified accuracy; typical 2 SD uncertainties ~0.08‰ (δ202Hg; fauna/sediment) and 0.10‰ (Δ199Hg; fauna) or 0.04‰ (Δ199Hg; sediments). Modeling and statistics: Weighted upper-ocean MMHg isotope values (0–1000 m) were estimated via binary mixing of surface fish (0–100 m; photodemethylated MMHg) and intermediate particles (IHg proxy adjusted by −0.5‰ δ202Hg for microbial methylation), weighted by mean dissolved MMHg concentrations (surface 18±7 pM; intermediate 38±19 pM) using Monte Carlo (n=10,000). Contributions of surface vs intermediate MMHg to trench fauna used Δ199Hg binary mixing with Monte Carlo (n=10,000). Statistical analyses employed OriginPro 9 (nonlinear exponential fits for depth regressions; linear fits otherwise; 95% confidence; ANOVA for slope significance).

• Hg burdens: Amphipod THg averaged 547 ± 230 ng g−1 dw (range 235–1070; n per Supplementary Tables), comparable to abyssal Arctic amphipods and higher than many benthic coastal/freshwater organisms. MMHg concentrations 8–403 ng g−1; MMHg% 2–59% with a significant negative correlation with body length (R²=0.48, P=0.02, n=13). Snailfish THg 970 ng g−1; MMHg 809 ng g−1; MMHg% 83%.
• Isotope signatures in fauna: Amphipods show δ202Hg from −0.05 to 0.54‰ (n=28); odd-MIF Δ199Hg 1.26–1.70‰ and Δ201Hg 1.01–1.37‰ (n=28), closely matching North Pacific fishes feeding at 300–600 m. Positive correlations: δ202Hg vs THg (R²=0.27, P=0.01), Δ199Hg vs body length (R²=0.18, P=0.04). Limited variability in δ202Hg and Δ199Hg despite wide MMHg% range implies internal demethylation rather than external IHg(II) uptake controls IHg fraction.
• Sediments: δ202Hg −0.96 ± 0.27‰ (1 SD, n=5); Δ199Hg 0.20 ± 0.07‰; Δ201Hg 0.18 ± 0.04‰, consistent with IHg(II) in sinking particles and low-MIF sources.
• Photochemical diagnostic slopes: Trench fauna Δ199Hg vs Δ201Hg slope 1.15 ± 0.08 (1SE), matching MMHg photodegradation (~1.2) rather than IHg(II) photoreduction (~1.0). Most fauna align with fish Δ199Hg–δ202Hg slope (2.13 ± 0.19), supporting photochemical imprint from surface waters.
• Provenance: Deep fauna Δ199Hg (mean 1.47 ± 0.13‰) matches weighted upper NPO MMHg Δ199Hg (1.44 ± 0.75‰), indicating supply from upper ocean. Binary mixing of Δ199Hg attributes 37–48% of fauna MMHg to surface waters and 52–63% to intermediate waters.
• Process in transit: Fauna δ202Hg (0.27 ± 0.14‰) is ~0.3‰ higher than weighted upper-ocean MMHg (−0.05 ± 0.30‰), consistent with microbial demethylation enriching δ202Hg during transport.
• Minimal in situ deep production: Elevated odd-MIF in fauna, contrasted with low-MIF sediments and expected near-zero MIF for hydrothermal/geologic Hg, indicates little to no in situ deep-ocean MMHg production; instead, MMHg carrying surface photochemical odd-MIF is transported downward, likely adsorbed to or incorporated in fast-sinking POM.
• Broader implications: Findings imply anthropogenic Hg introduced at the surface pervades to the deepest ocean via particle transport; warming-induced increases in upper-ocean MMHg could rapidly propagate to hadal ecosystems.

The study resolves the origin of methylmercury in the deepest marine food webs: isotope evidence (positive odd-MIF and diagnostic Δ199Hg/Δ201Hg slopes) in trench amphipods and a snailfish matches surface-ocean photochemical processing, ruling out significant deep in situ MMHg production. The mismatch between high odd-MIF in fauna and low odd-MIF in sediments and particles indicates that MMHg incorporated into hadal food webs must inherit photochemical signatures from the euphotic zone. Agreement between trench fauna Δ199Hg and a concentration-weighted upper-ocean MMHg Δ199Hg, alongside mixing model results, supports a mechanism where surface and mesopelagic-produced MMHg is transported to depth primarily via sinking particles (the biological pump). During descent and residence, microbial demethylation enriches residual MMHg in δ202Hg without altering odd-MIF, explaining observed δ202Hg shifts. These results suggest deep-ocean microbial methylation is limited under cold, oxygen-rich, nutrient-poor conditions, while microbial demethylation is more widespread (including in amphipod guts). Given the linkage to upper-ocean sources, anthropogenic Hg perturbations at the surface, and potential climate-driven changes that elevate MMHg in upper waters, are likely to transmit rapidly to hadal food webs. The findings also inform DMHg cycling: if MMHg production in deep waters is minimal, then subsequent microbial methylation of MMHg to DMHg is likely limited; deep-ocean DMHg may form abiotically, potentially on particle surfaces.

Mercury isotope fingerprints in hadal amphipods and a snailfish from the Mariana and Yap Trenches demonstrate that bioavailable MMHg in the deepest ocean is derived from upper-ocean production and delivered to depth by sinking particles, with negligible in situ deep-ocean methylation. A binary mixing analysis indicates roughly 37–48% surface and 52–63% intermediate contributions to trench fauna MMHg. δ202Hg enrichment relative to weighted upper-ocean MMHg implicates microbial demethylation during transit. These results imply that anthropogenic Hg introduced at the surface permeates to hadal ecosystems and that future changes elevating upper-ocean MMHg (e.g., warming) could rapidly affect deep food webs. Future research should quantify rates of particle-associated MMHg transport, constrain regional and temporal variability across trenches and basins, directly measure deep-water microbial methylation/demethylation rates, and elucidate mechanisms of DMHg formation at depth, including potential abiotic pathways on particle surfaces.

• Geographic and ecological scope: Samples are from the Mariana and Yap Trenches in the North Pacific; findings may not capture variability across other ocean basins or trench systems.
• Sample representation: Isotope analyses were performed on a limited number of fauna (n≈28 for isotopes; fewer for MMHg%), with compositing of individuals; life-history differences and physiological effects could introduce variability not fully resolved.
• Isotopic diagnostic constraints: Limited variation in Δ199Hg and δ202Hg within trench fauna precluded independently confirming the Δ199Hg/δ202Hg photochemical slope; small physiological correlations (R² ~0.2–0.3) suggest minor internal effects on isotope composition.
• Assumptions in proxies and mixing: Use of fish and particle isotope values as proxies for surface and intermediate MMHg, and adjustment for methylation MDF, introduce uncertainties; Monte Carlo mixing relies on literature MMHg concentrations.
• Lack of direct deep-rate measurements: The study infers limited deep in situ methylation without direct incubation rate measurements of deep waters; hydrothermal contributions are inferred from isotope patterns but not directly quantified at vents near sampling sites.