Methylmercury (MeHg) is a global pollutant that can cause metabolic disruptions in animals and thereby potentially compromise the energetic capacity of birds for long-distance migration, but its effects on avian lipid metabolism pathways that support endurance flight and stopover refueling have never been studied. We tested the effects of short-term (14-d), environmentally relevant (0.5 ppm) MeHg exposure on multiple lipid metabolism markers in the pectoralis and livers of yellow-rumped warblers (Setophaga coronata) that were previously found to have poorer flight endurance in a wind tunnel than untreated conspecifics. Compared to controls, MeHg-exposed birds displayed lower muscle aerobic and fatty acid oxidation capacity, but similar muscle glycolytic capacity, fatty acid transporter expression, and PPAR expression. Livers of dosed birds indicated elevated energy costs, lower fatty acid uptake capacity, and lower PPAR-γ expression. The lower muscle oxidative enzyme capacity of dosed birds may have caused or contributed to their weaker endurance in the prior study, while the metabolic changes observed in the liver have potential to inhibit lipogenesis and stopover refueling. Our findings provide concerning evidence that fatty acid catabolism, synthesis, and storage pathways in birds can be dysregulated by only brief exposure to MeHg, with potentially significant consequences for migratory performance.

We analyzed tissues that were collected from 24 yellow-rumped warblers used by Ma et al. [17] to study the effects of MeHg on flight performance in a wind tunnel. The reader is directed to Ma et al. [17] for detailed methods regarding capture, husbandry, dosing, and wind tunnel testing. Briefly, the birds were captured during migratory stopovers at Long Point, Ontario, Canada in the autumn of 2014 and transported to indoor aviaries at the University of Western Ontario in London, Canada. There they were maintained on an ad libitum mercury-free, synthetic diet under light cycles that simulated autumn (12L:12D) and then winter (9L:15D) photoperiods until March of 2015, when the photoperiod was lengthened (16L:8D) to stimulate the birds into spring migratory condition. After 14 d of the lengthened photoperiod, 12 randomly selected birds (4 male, 8 female) were switched to an ad libitum synthetic diet containing 0.5 ppm ww MeHg for 14 d while the other 12 birds (3 male, 9 female) continued to feed on the same mercury-free diet to serve as a control group. The dosing concentration of 0.5 ppm ww was intended to represent the middle to upper range of MeHg levels commonly found in the arthropod prey of songbirds in eastern North America [4,27,28].

Blood samples were collected by brachial venipuncture on days 0 and 14 of the experiment, and stored at -80º C. Total body mass was measured to 0.01 g on a digital balance, and fat mass and lean mass were measured to 0.001 g with quantitative magnetic resonance analysis, also on days 0 and 14. On day 14, 22 of the 24 birds were flown for up to 2 h in a wind tunnel, after which they were immediately euthanized. Two individuals that could not be flown in the wind tunnel because of missing tail feathers were also euthanized on day 14 for inclusion in the present study. Pectoralis muscles and livers were then collected from all 24 birds immediately following euthanasia and stored at -80º C. Dosed and control birds did not differ in body size (wing length) upon capture, or body mass or body composition at the start or end of the 14-d experiment [17].

Laboratory analyses

Tissue mercury concentrations

We measured blood total mercury (THg; ww) as a proxy for MeHg using a direct mercury analyzer (DMA-80, Milestone Inc., Shelton, USA), as described in Ma et al. [17]. The instrument was calibrated with a certified reference material (Caprine Blood SRM 955c) and quality control was assessed with a method blank, certified concentration standard (CCS), and sample duplicates. Mean percent recoveries were 104.33 ± 3.37% (SRM 955c; n = 8), 92.15 ± 0.98% (DORM-2; N = 3), and 97.85 ± 0.81% (CCS; N = 36). CVs of duplicates averaged 6.84% ± 1.40%. We estimated pectoralis and liver THg concentrations based on blood THg concentrations, using regression equations from Ma et al. [17].

Metabolic markers

We measured citrate synthase (CS) and carnitine palmitoyltransferase (CPT) activity to assess aerobic capacity and fatty acid oxidation capacity, respectively [29-31]. We measured the gene expression of fatty acid transport proteins to assess muscular and hepatic fatty acid uptake capacity. They included fatty acid translocase (FAT/CD36) and plasma membrane fatty acid binding protein (FABPpm), to evaluate membrane fatty acid uptake capacity, and heart-type fatty acid binding protein (H-FABP), to evaluate intracellular fatty acid transport capacity (pectoralis only) [20,21,30]. We then measured expression of peroxisome proliferator-activated receptors (PPAR-α, PPAR-ß, PPAR-γ), which are nuclear receptors that regulate lipid metabolism in birds [32]. PPAR-α and PPAR-β are highly expressed in heart and skeletal muscle where they mainly target genes for proteins involved in lipid oxidation, including CPT and FAT/CD36 [33,34]. PPAR-γ, in contrast, primarily controls expression of FABPpm, FAT/CD36, and lipogenic enzymes in the liver and adipocytes to regulate lipid synthesis and storage [33,34]. Lastly, we measured lactate dehydrogenase (LDH) activity to assess effects of MeHg on glycolytic capacity given the broad ability of MeHg to inactivate enzymes [35] and the effect this could have on migrating birds early in flight, before they switch over from catabolizing primarily carbohydrates to primarily fatty acids [36].

Enzyme assays

We homogenized 50-100 mg of pectoralis and liver using beadmill homogenization (NextAdvance, Troy, NY; speed 8, time 3 for liver and time 4 for pectoralis, 4°C) in 9 volumes of 20 mM Na2PO4, 0.5 mM EDTA, 0.2% BSA, 0.1% Triton x-100, and 50% Glycerol, pH = 7.4. We then assayed CS, CPT, and LDH in duplicate or triplicate in a microplate reader (BioTek, Winooski, VT) at 39°C to represent a typical avian body temperature. We measured LDH at 340 nm using 0.4 mM pyruvate, 0.66 mM NADH, 5 mM DTT, 50 mM Tris, pH = 7.4. For CS and CPT, we measured absorbance at 412 nm to detect the appearance of CoA-TNB as in Price et al. [37].

qPCR of transport proteins and PPARs

We homogenized ~50 mg of pectoralis and liver in 1 mL of TriZol (Invitrogen, Carlsbad, CA) and separated RNA into an aqueous phase using a chloroform ethanol procedure. We purified total RNA using a PureLink RNA kit (Invitrogen, Carlsbad, CA) and quantified RNA using a plate adapter (BioTek, Winooski, VT). We treated 14 µg of RNA using TURBO I DNase (Invitrogen, Carlsbad, CA) and used Nanodrop 2000 (Thermo Fisher Scientific, Waltham, MA) to quantify and check quality before reverse transcription of 1 μg of DNase-treated RNA using NEB Luna RT (New England Biolabs, Ipswich, MA).

We selected Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as a housekeeping gene for qPCR using a published primer sequence for yellow-rumped warblers [38] and used published primer sequences for FAT/CD36, FABPpm, H-FABP, and PPARs [39-41]. Each cDNA sample was diluted 1:5 with nuclease-free water before analysis. We confirmed primer specificity on a 2% agarose gel and used a serial dilution series of pooled cDNA to obtain amplification efficiency (between 97.1% and 99.9%). Reactions were run in duplicate in a StepOnePlus qPCR machine (Applied Biosystems, Foster City, CA; 60s at 95°C, followed by 40 cycles at 95°C for 15s and 60°C for 30s) in 10 µL volumes with 0.2 µM primers and 1 µL cDNA in Luna Universal qPCR master mix (New England Biolabs, Ipswich, MA). We used comparative cycle threshold (C_T) analysis following Schmittgen and Livak [42], subtracting the average C_T for the corresponding housekeeping gene reading (GAPDH) from the average C_T from duplicate target gene wells to obtain the ΔC_T for each individual and gene, then subtracting the control group mean ΔCT from the MeHg-dosed group mean to obtain ΔΔC_T. Fold change was calculated for the mean and upper and lower limits of the SEM using 2-^(ΔΔCT).

Statistical analyses

We compared blood THg concentrations between treatment groups with a two-tailed t-test. We tested the effects of treatment on each enzyme, transport protein, and PPAR isoform using general linear models. We included lean mass, fat mass, the lean mass by treatment interaction, and fat mass by treatment interaction in full models which we then compared to nested models with likelihood ratio tests to determine whether any of these terms could be dropped. We retained a term for inclusion in a final model along with treatment when a likelihood ratio test of the full model against the nested model from which it was removed had a P value of < 0.1. We used lean mass also as a proxy for body size (wing length) and sex, as all three variables were significantly related to each other.

We performed all statistical analyses in R (v 4.0.4, R Foundation for Statistical Computing, Vienna, Austria) and accepted significance when P < 0.05. We log-transformed non-normal variables prior to analyses and removed outliers that were likely measurement errors, such as unrealistically high or low values, an error in the housekeeping gene, or a consistent error for a set of tissue samples run together.