Sample collection

Whole eggs were collected from nine breeding pairs of Humboldt Penguins under managed care at the Brookfield Zoo, Chicago, Illinois, USA, from September 2020 to April 2021. Eggs were refrigerated in plastic specimen containers (4 C) until shipment to Nova Southeastern University (Fort Lauderdale, Florida, USA) for analysis. Analyzed samples included eggshell (n=18), yolk (n=14), and albumen (n=14). Internal components for four eggs were omitted from analyses because of structural degradation.

Sample preparation and analysis

Whole eggs were rinsed with ultrapure deionized water (18.2 megohms) from a Barnstead water (Lake Balboa, California, USA) purification system, and gently scrubbed with a nylon cleaning brush. Eggs were then cracked along the equatorial axis and separated into eggshell, albumen, and yolk. Individual egg components were placed in acid-washed glass beakers and dried for a minimum of 24 h in an isotemp vacuum oven (Thermo Fisher Scientific, Waltham, Massachusetts, USA) at 100 C, with pressure reduced below 10^-2 torr using a vacuum pump (14008-01 model Welch 1400 DuoSeal, Welch, Mt. Prospect, Illinois, USA). Approximately 0.2 g (dry mass) of each sample was placed in a microwave acid digestion vessel (Parr Instrument Company, Moline, Illinois, USA) with 4 mL of 30% hydrogen peroxide (Sigma-Aldrich, St. Louis, Missouri, USA) and 1 mL of concentrated trace-metal-grade nitric acid (Sigma-Aldrich). The vessel was heated in a conventional microwave oven at 400 W for 35 s to achieve internal device pressure up to 1200 psi and reactant temperatures up to 250 C, ensuring total sample digestion, then cooled and depressurized for 45 min.

Digested samples were shipped to the University of Southern Mississippi Center for Trace Analysis (Hattiesburg, Mississippi, USA) for analysis of 16 elements (15 metals plus selenium) via inductively coupled plasma-mass spectrometry (ICP-MS): aluminum (Al), arsenic (As), cadmium (Cd), chromium (Cr), cobalt (Co), copper (Cu), iron (Fe), lead (Pb), manganese (Mn), mercury (Hg), molybdenum (Mo), nickel (Ni), selenium (Se), tin (Sn), vanadium (V), and zinc (Zn). Before analysis, digested samples were diluted fivefold in 0.64 M ultrapure nitric acid containing 2 ppb indium as an internal standard and held in acid-washed Teflon autosampler vials (Thermo Fisher). Mass spectrometer scans were performed at low (Cd-111, Hg-199, 200, 201, 202, Pb-208), medium (Al-27, V-51, Cr-52, Mn-55, Fe-56, Co-59, Ni-60, Cu-63, Zn-66), and high (As-75, Se-77,82) resolution, depending on the isotope. To correct for molybdenum oxide interference on Cd, Mo-98 was monitored. External standards were used with a high standard and a blank rerun every eight samples. Two US Geological Survey reference water concentrations were also assessed as part of each analytical run to verify standardization. In several cases, sample calibration was also verified by standard additions. Blanks of ultrapure deionized water and trace metal basis nitric acid (3%, 4%, 5%) were used for quality control purposes. The ICP-MS detection limits of each element are provided in  Supplementary Materials Table S1 (JWD-D-22-00176R2_Suppl_Final.docx). No certified reference materials were used because none are available for heavy metal concentrations in egg components.

Statistical analyses were performed using Excel (version 16.60; Microsoft Corporation, Redmond, Washington, USA) and R statistical software (R 4.2.0 GUI 1.78 High Sierra build [8075]; R Core Team 2022). Geometric mean was used rather than arithmetic mean, to mitigate the impact of data outliers. A Shapiro-Wilk test was used to determine if the data were normally distributed, and a Bartlett's test was used to determine homogeneity of variances. Nonparametric analyses were used because of variable data distributions and lack of homogeneity. Kruskal-Wallis and post-hoc tests were performed to determine if metal concentrations varied among egg components. Potential correlations in heavy metal concentrations within egg components were assessed via Kendall's tau correlation (for n,30) and plotted as correlation matrices ( Supplementary Materials Figs. S1–S3 (JWD-D-22-00176R2_Suppl_Final.docx)). Values below the detection limit (DL) were proxied as one-half of the DL. Statistical significance was identified when P,0.05. Selenium binds to organic Hg at a molar ratio of 1:1; therefore the molar ratio was calculated for each sample using

where 78.96 g/mol and 200.59 g/mol represent the atomic masses of Se and Hg, respectively.

Yolk

Of all egg components, yolk exhibited the highest concentrations for 10/16 elements; Mo was detected only in yolk and not in albumen or eggshell. The range of elements concentrated in yolk can be largely explained by the metabolomics of avian vitellogenesis. Lipid is the largest yolk constituent by dry mass (Nys and Guyot 2011). Certain metals, such as Mn and Zn, are able to form lipophilic metal complexes that promote absorption and tissue penetration, increasing contaminant retention in yolk (Tjälve and Gottofrey 1991). These metals also tend to accumulate in organs responsible for toxin elimination, such as the liver (Das et al. 2003). The lipids vitellogenins and yolk-targeted very-low-density lipoproteins (VLDLy) are synthesized in the liver, secreted into the circulatory system, and taken up within the ovary's follicles for generation of yolk, the nutrient source for the developing embryo (Walzem 1996; Vezina et al. 2003). Once transported to the oocyte, vitellogenin is catabolized to phosvitin and lipovitellin, the predominant proteins found in yolk (Hermier et al. 1989; Speake et al 1998). Phosvitin exhibits a high capacity to bind essential metals (e.g., Ca and Fe), hence their enrichment in yolk (Strixner and Kulozik 2011).

During yolk development, ∼17 d in Humboldt Penguins (Ancel et al. 2013), hepatic lipid production increases as lipoprotein synthesis shifts to generate the smaller, lipid-rich VLDLy molecules required to support an embryo. As plasma vitellogenin concentrations increase, a corresponding rise in trace mineral concentrations occurs due to vitellogenin-metal binding (Hill 1974; Panic et al. 1974; Richards 1989). Increases in VLDLy and vitellogenin production by the liver necessitate a corresponding increase in hepatocyte metabolism (Vezina et al. 2003). Richards (1997) attributes yolk heavy metal content almost entirely to vitellogenesis-driven increases in hepatic metabolism, which in turn facilitate maternal transfer of contaminants to the developing egg via the circulatory system.

The trends in yolk metal accumulation that we found have been seen in other birds, as vitellogenesis mechanisms are conserved across avian species. Duck eggs from commercial farms in Thailand had yolk concentrations of Fe, Mn, and Zn an order of magnitude greater than those in albumen (Aendo et al. 2018), a trend reflected in our Humboldt Penguin eggs (Table 3). Humboldt Penguin yolk Cu, Hg, and Pb concentrations (Table 1) exceeded wet weight concentrations in Taiwanese duck eggs (Jeng and Yang 1994) by an order of magnitude of 19. However, concentrations in commercial chicken eggs (Demirulus 2013; Sarkar et al. 2018), were consistently higher than those we observed in Humboldt Penguin eggs; this might be attributable to herbicide or pesticide contamination of feed residues (Sarker et al. 2017; Table 3). Yolk Pb concentrations (0.322 mg/g; Table 1) in our study were similar to those found in the homogenized internal egg components of Great Cormorants (Phalacrocorax carbo) in Greece (Goutner et al. 2001). Although values for yolk of Pied Avocets (Recurvirostra avosetta), Yellow-legged Gulls (Larus cachinnans michahellis), and Mediterranean Gulls (Larus melanocephalus) from the Evros and Axios deltas in Greece were an order of magnitude greater than those penguin yolk, these birds also occupy habitats with considerable pollutant input from freshwater discharge generated by local agricultural complexes and commercial aquaculture ponds (Goutner et al. 2001). Compared to homogenized egg contents collected from Yellow-breasted Chats (Icteria virens) and Willow Flycatchers (Empidonax traillii) from Arizona, USA (Mora 2003), Humboldt Penguin yolk values exceeded wet weight concentrations of Cu, Mn, and Zn but were lower in Se (Table 3). This may be attributable to the managed-care Humboldt Penguins occupying a higher trophic niche than either shorebird species, with contaminant biomagnification subsequently resulting in greater metal accumulation in the penguins' egg components.

Table.

Mean metal concentrations (mg/g) in the egg components of various bird species compared to those in egg components of Humboldt Penguins (Spheniscus humboldti) from a managed-care population at the Brookfield Zoo, Chicago, Illinois, USA.

Table 3.

Continued.

Mean As (0.20 mg/g), Mn (2.71 mg/g), Pb (0.322 mg/g), and Zn (64.03 mg/g; Table 1) concentrations in Humboldt Penguin yolk fell within the range of mean concentrations reported for feathers from the Antarctic penguins Adelie Penguins (Pygoscelis adeliae; Dos Santos et al. 2006), Chinstrap Penguins (Pygoscelis antarctica; Jerez et al. 2013b), and Gentoo Penguins (Pygoscelis papua; Metcheva et al. 2006; Jerez et al. 2011, 2013a;). Although As, Mn, and Zn concentrations in kidney, liver, and muscle of these species exceed managed-care Humboldt Penguin yolk concentrations by up to an order of magnitude (Metcheva et al. 2010; Jerez et al. 2013a, b), the same tissues are significantly lower than yolk for Pb. Yolk Pb values were also similar to those in feathers of Magellanic Penguins (Spheniscus magellanicus) from the coast of southern Brazil (Kehrig et al. 2015), and yolk As and Zn concentrations were similar to those in feathers of Little Penguins (Eudyptula minor) from three island rookeries near Melbourne, Australia (Finger et al. 2015).

Yolk was an order of magnitude higher in Hg and two orders of magnitude higher in Zn than serum from wild Humboldt Penguins at PSJ but comparable in Al concentrations. Wild Humboldt Penguin feathers at PSJ had significantly higher concentrations of Al, As, Cu, Fe, Hg, and Mn than yolk from the eggs that we tested (Adkesson 2019). Humboldt Penguin excreta from three Chilean rookeries had mean As and Pb concentrations 40 times greater and mean Zn eight times higher (Celis et al. 2014) than in the yolks we tested (Table 4). Heightened metal concentrations in wild birds are to be expected because of the prevalence of mines and generalized anthropogenic activity throughout coastal Peru and Chile (Adkesson 2019).

Table 4.

Mean metal concentrations (mg/g) in the eggs and tissues of various penguin species compared to those in egg components of Humboldt Penguins (Spheniscus humboldti) from a managed-care population at the Brookfield Zoo, Chicago, Illinois, USA. For comparisons where an element was not tested, a dash is used in lieu of a concentration value.

Table 4.

Continued.

Data for other heavy metals that we found to be highly concentrated in the yolk (Co [0.01 mg/g], Cr [0.086 mg/g], Mo [0.57 mg/g], and Sn [3.29 mg/g]; Table 1) are not well established in the literature for penguins.

Albumen

We found the highest concentrations of Al, Cd, Cu, Hg, and Se in albumen, a lipid-free fluid comprising primarily water and proteins, as well as vitamins and minerals required for embryonic development (Strixner and Kulozik 2011). Unlike yolk constituents, which are derived from components synthesized remotely (i.e., in the liver), albumen proteins are synthesized and secreted locally by the magnum of the oviduct (Nys and Guyot 2011). Ovalbumin and ovotransferrin, the two principal albumen proteins, form metal complexes that offer protection from bacterial infections (Schiavone and Barroeta 2011). Ovalbumin binds both Se and Hg (Magat and Sell 1979; Latshaw and Biggert 1981), as well as Cu, Mn, and Zn (Goux and Venkatasubramanian 1986). Ovotransferrin is principally responsible for Fe binding (Schiavone and Barroeta 2011), but Al, Cd, Cr, Co, Cu, Mn, and Zn bind in vitro (Burley and Vadehra 1989). The highly positive correlation between Mn and Fe in Humboldt Penguin albumen may result from Mn interacting with Fe cofactors on enzymes and impinging on transferrin receptor expression, limiting Fe endocytosis into red blood cell precursors, hepatocytes, and endothelial cells (Gupta et al. 2007).

Mercury levels in birds are well established due to the toxicity associated with its bioaccumulation and biomagnification. Humboldt Penguin mean Hg concentrations in albumen (0.62 mg/g; Table 1) were five times higher than levels found in wild Gentoo Penguin albumen (Brasso et al. 2012). This likely results from managed-care birds receiving higher trophic prey items than those foraged in the wild, because Hg levels in albumen from eggs laid by managed-care Gentoo Penguins are higher than albumen from wild birds (Brasso et al. 2012).

Notably, Hg in Humboldt Penguin albumen (0.62 mg/g; Table 1) far exceeded values in wild Magellanic Penguin feathers (0.033 mg/g; Frias et al. 2012) despite the species' nearly identical trophic dynamics and ecological niche (Simeone et al. 2009). Additionally, Humboldt Penguin albumen is comparable to values found in Adelie (Carravieri et al. 2016) and Gentoo Penguin feathers (Dos Santos et al. 2006; Table 4).

Concentrations of Cu (4.02 mg/g), Hg (0.618 mg/g), and Pb (0.0541 mg/g) in Humboldt Penguin albumen were greater than those found in Thai (3.38 mg/g for Cu, 2.21 mg/g for Pb; Aendo et al. 2018), and Taiwanese duck eggs (0.00136 mg/g, 0.0178 mg/g, and 0.0146 mg/g, respectively; Jeng and Yang 1994). Additionally, penguin albumen had higher Cd levels than yolk, a trend also observed in ducks (Aendo et al. 2018) that is likely attributable to ovotransferrin binding (Burley and Vadehra 1989). However, much like for yolk, heavy metal concentrations in commercial chicken egg albumen have been found to be significantly greater than those in Humboldt Penguin egg albumen (Demirulus 2013; Sarkar et al 2018; Table 3).

The mean albumen Al (4.19 mg/g) that we found resembled concentrations in the kidneys and livers of Adelie, Chinstrap, and Gentoo Penguins (Jerez et al. 2013a), while mean Cd levels (0.0009 mg/g) were lower than values in these penguins across tissue types (Jerez et al. 2013a; Table 4). While mean Al concentrations in feathers from Humboldt Penguin at PSJ far exceeded mean egg albumen levels in the managed-care population, serum Al levels in the wild birds were approximately half that seen in albumen (Adkesson et al. 2019). Albumen also contained up to two orders of magnitude higher Hg and Zn than serum collected from wild Humboldt Penguins at PSJ. Concentrations of Se (1.52 mg/g) were highest in albumen; however, much like Co, Cr, Mo, Sn, and V in yolk, Se concentrations are not reported for penguins (Table 4).

Eggshell

Eggshell had the lowest concentration of all analyzed metals except Ni (0.31 mg/g). This was unexpected, as certain divalent metals, specifically Cd and Pb, mimic and replace Ca during shell biomineralization (Strixner and Kulozik 2011). The replacement of calcium by its mimics results in the inhibition of Ca precipitation in calcified tissues (Wada et al. 1995; Bridges and Zalups 2005), resulting in metal increases in eggshells (O'Flaherty 1991; Teodorova et al. 2003; Scheuhammer 1996). Accordingly, Cd and Pb are expected to concentrate in the eggshell but were found in greater concentrations in albumen and yolk.

Humboldt Penguin heavy metal concentrations in eggshell were comparable to Gentoo Penguin eggshells for Al, Cu, Mn and Zn; Gentoo Penguin eggshell had higher metal concentrations (Metcheva et al. 2011). Likewise, all analyzed metals were found in higher concentrations in the feathers and serum of wild penguins at PSJ than in our Humboldt Penguin eggs (Adkesson et al. 2019; Table 4).

However, our Humboldt Penguin eggshells were higher in As (3.61 mg/g) than eggshells collected from Bridled Terns (Sterna anaethetus; 0.397 mg/g) inhabiting the Pearl River Delta manufacturing complex near Hong Kong (Lam et al. 2005). Eggshells of Little Egrets (Egretta garcetta) and Black-crowned Night Herons (Nycticorax nycticorax) inhabiting the same industrial complex had higher quantities of Cd, Mn, and Pb, while Zn concentrations were within the same range (2.351 mg/g to 9.636 mg/g) as our Humboldt Penguin eggshells (3.00 mg/g; Table 3). Managed-care birds represent a population with a significantly reduced risk of environmental contaminant exposure, yet the penguins assessed in this study nevertheless had accumulated metals in their eggshells to an extent comparable to findings in the wild birds assessed by Lam et al. (2005). Thus, when considering Humboldt Penguin populations inhabiting rookeries near coastal mines in Peru and Chile, we conjecture that there is an immense potential for significant contaminant accumulation in wild penguin eggshells, because of the introduction of metals and other environmental contaminants from local ore generation.

Implications for wild Humboldt Penguins

A recent ecosystem health assessment of Peruvian marine protected areas (MPAs), Loaiza et al. (2022) designated Punta San Juan, home to the largest Peruvian Humboldt Penguin rookery, as having the highest incidence of adverse biological effects and highest metal concentrations for As, Cd, Cu, Ni, Pb, and Zn in sediment and seston (particles suspended in water) out of all surveyed MPAs, because of its proximity to mines. Heavy metal exposure during embryonic development has been heavily implicated in teratogenesis and developmental deformities in birds (Gilani and Alibhai 1990). Accordingly, the potential for reductions in offspring viability due to heavy metal contamination in eggs is a concern for species exhibiting multidecadal population declines, such as the Humboldt Penguins. Our findings serve to substantiate the value of assessing eggs when conducting assessments of environmental contamination in these birds.

It is important to note that the penguins in this study remained in a controlled environment (i.e., consistent diet and environment) throughout oogenesis and egg laying. Variance in habitat, diet, and environmental exposure risk experienced by wild birds significantly influences their assimilation of inorganic contaminants, and this must be taken into consideration. Additionally, most published studies assessing heavy metal contamination in eggs have presented data in terms of eggshell and homogenized internal components (i.e., combined yolk and albumen), limiting the extent to which some comparisons could be made. In the future, quantification of heavy metal concentrations in wild Humboldt Penguin egg components should be pursued to fully assess concerns regarding heavy metal exposure and contamination to the chick embryo. By determining the extent of heavy metal offloading that occurs via egg production, linkages among contaminants, embryo viability, and the likelihood of post-hatch chick mortality may be established. Such data would significantly inform existing strategies for the management of wild Humboldt Penguins, facilitating the conservation of this species.