Uptake, Efflux, and Sequestration of Mercury in the Asian Clam, Corbicula fluminea, at Environmentally Relevant Concentrations, and the Implications for Mercury Remediation

Abstract

(1) Mercury (Hg) is a persistent, ubiquitous contaminant that readily biomagnifies into higher trophic level species in aquatic environments across the globe. It is crucial to understand the movement of environmentally relevant concentrations of Hg in impacted freshwater streams to minimize risks to ecological and human health. (2) The bioconcentration kinetics of aqueous Hg exposure (20, 100, and 200 ng/L) in the invasive Asian Clam, Corbicula fluminea, were measured. A toxicokinetic model, the first parameterized for Hg accumulation in freshwater clams, was developed to estimate uptake and efflux parameters and compared to previous parameter values estimated for other mollusk species. (3) Results demonstrated that even at low Hg concentrations, Corbicula record signals of contamination through bioconcentration, and both direct measurement and toxicokinetic models demonstrate large Hg bioconcentration factors (as high as 1.34 × 10⁵ mL/g dry tissue), similar to partitioning coefficients seen in engineered Hg sorbents. (4) Our study found that Corbicula accumulated Hg at aqueous concentrations relevant to impacted streams, but well below regulatory drinking water limits, demonstrating their utility as a sensitive sentinel species and potential bioremediator.

1. Introduction

Freshwater mollusks are foundational species in many freshwater ecosystems, providing a wide variety of important ecosystem services and influencing ecosystem health, structure, and function [1,2,3]. Some of the important ecosystem services provided by mollusks include improving water quality, influencing nutrient cycling, algal communities, and structural habitat, as well as acting as biomonitors of pollution because bivalves accumulate waterborne contaminants as filter feeders [1,4,5,6,7]. Despite their ecological and environmental value, anthropogenic impacts have resulted in large declines in native mollusk populations, with dozens of species going extinct and hundreds more imperiled [3,5,8,9,10]. Corbicula fluminea (hereafter Corbicula), an invasive species of freshwater clam, often fill the niche left by the absence of native species. They are highly successful at outcompeting native species of freshwater bivalves due to biological characteristics that make them resilient in environments not well suited to less hardy bivalve species, such as rapid growth rates, early sexual maturity, short life spans, and spatial dispersion often driven by human activity [11].

Post invasion, Corbicula quickly become a foundational species in ecosystems, supporting or replacing the ecosystem services of water filtration previously supplied by native species of filter feeders [12,13,14]. Corbicula filter significant volumes of water, on the order of 100–1000 mL/h/g of body mass, through their digestive and respiratory systems, which removes particulate matter, including bacteria and algae, improving water quality [15,16]. Previous work has shown that Corbicula have a number of different filtration behaviors compared to native mussels, such as a lower filtration rate per individual than the native species Utterbackia imbecillis, but a higher filtration rate per gram of body mass [17,18], negatively correlating with size (smaller individuals had higher specific clearance rates) [19]. Additionally, Corbicula filtration rates increase at night and are 4.6× higher than Utterbackia imbecillis [20]. This filtration often has a profound impact on nutrient cycling and plankton populations: one study in a river in Germany found that freshwater mussel filtration drove changes in plankton community profiles [1,21]. This sequestration of nutrients is a major transfer vector from the water column to the benthic environment, and provides important nutrients for deposit feeders, enriching the nutritional value of benthic matter and increasing the number of benthic macroinvertebrates [22,23,24].

Corbicula also play a role in aqueous contaminant sequestration, with high bioaccumulation metrics for a variety of compounds of concern, such as emerging contaminants of concern, PCBs, and a variety of metals [16,25,26,27,28]. Due to this ability, Corbicula have been proposed as a possible remediation method for reducing industrial metal contamination in surface waters [29]. However, the invasive capacity of these clams must be considered, and ref. [29] emphasizes that their proposal to use Corbicula as biofilters is to suggest measuring their impact in systems that are already invaded by this species (C. fluminea is distributed across all continents except Antarctica) or to develop a controlled configuration to hold these organisms and restrict their movement. Corbicula serve a similar role in contaminated ecosystems, as additional work has shown a negative correlation between mercury (Hg) concentrations in redbreast sunfish and the number of collector-filterers in a community [17]. Corbicula exhibit an affinity for sequestering Hg, accumulating high concentrations in their tissue relative to aqueous concentrations [11,26,30,31]. Additionally, the removal of particulate matter from the water column may reduce the rate of Hg methylation [32] and improve the effectiveness of engineered sorbents that have been proposed as a remediation tool for Hg removal in contaminated streams [33]. The ability of Corbicula to filter large volumes of water, sequester metals, and reduce particulate matter concentrations, combined with their rapid growth and rapid proliferation suggests they may play an important role in maintaining water quality and influencing Hg cycling in contaminated streams [17].

Hg is a widespread contaminant of concern at a global scale, with unique chemical properties that simultaneously make it valuable to industry, toxic to humans, and both ubiquitous and persistent in the environment [34,35]. Hg is released into the environment by a combination of natural and anthropogenic activity and is distributed globally through both atmospheric deposition and point-source contamination [34,36,37]. Once released, Hg undergoes complex chemical, physical, and biological processes, which cause it to partition into many phases: adsorbed to sediments, accumulated in animal tissue, dissolved in water, and dispersed in the atmosphere [34,36]. Inorganic Hg may also be transformed, through the actions of microbes, into methylmercury (MeHg) which is more bioaccumulative and toxic than inorganic Hg. Efforts to control Hg contamination in the environment generally focus on limiting anthropogenic releases, such as mining, industrial waste, and the burning of biomass and fossil fuels [36,38,39]. Remediation technologies not focused on treating or reducing discharges include phytoremediation by algae and aquatic plants, adsorption to solids such as functionalized clays and carbon products, stabilization by the addition of chelating agents, and thermal treatment to volatilize Hg in highly contaminated soils [33,38,40]. Despite these efforts, Hg contamination in fish, the major pathway of human exposure, continues to present regulatory challenges and raises considerable health concerns [37,38,41,42,43,44,45]. Ecological and biological methods for Hg sequestration focus on Hg uptake by organisms with a high affinity for Hg and may present a viable solution for reducing Hg concentrations in complex natural systems at large scales [46,47]. Due to their high accumulation of metals, the toxicokinetics of Hg accumulation in Corbicula were measured in this paper and compared to published values for engineered Hg sorbents to estimate the impact these invasive clams have on Hg movement through impacted systems.

Past work examining Hg uptake in Corbicula was conducted on aqueous concentrations in excess of 1 μg/L, dozens or hundreds of times higher than Hg concentrations regularly measured in many Hg-contaminated streams [26,30,37,48,49]. While Hg concentrations vary considerably in surface waters, experiments conducted using concentrations that are orders of magnitude higher than are commonly seen in streams may not predict uptake and depuration behavior at much lower aqueous concentrations. Additionally, it is important to understand at what Hg concentrations clams will exhibit a measurable bioaccumulation response, as well as other physiological responses. To better understand Hg dynamics in Corbicula at concentrations relevant to natural waters, a laboratory study was performed exposing the clams to three concentrations of inorganic Hg (20, 100, and 200 ng/L) with the intention of quantifying aqueous Hg uptake and efflux. The chosen treatment levels cover a range of Hg concentrations seen in waterways with legacy Hg contamination and are much lower than those previously empirically examined with this species. These data were then used to parameterize a toxicokinetic model of Hg uptake and depuration in Corbicula that is directly relevant to streams with Hg concentrations common in contaminated surface waters. The combination of this data set and successfully parameterized model allows for the prediction of the accumulation of Hg by Corbicula in impacted systems.

2. Materials and Methods

2.1. Organism Collection and Care

Approximately 300 C. fluminea between 0.6 and 4.6 g whole body weight were collected via sediment disturbance and netting from Sewee Creek, Meigs County, TN (35°35′38.8″ N, 84°42′46.9″ W) in the spring of 2021. The clams were maintained for two months in a 110 L plastic flow-through tank at room temperature in water pumped from First Creek, a stream in East Tennessee with no history of Hg contamination; samples taken during the holding period were below the detection limit of the Ohio Lumex Portable Mercury Vapor Analyzer RA-915 Plus (Ohio Lumex Company, Inc., Twinsburg, OH, USA) instrument, 2 ng/L. Water was renewed at a rate of 0.53 L/min, corresponding to a residence time of approximately 3.4 h. Clams were continuously fed live algal cells (an equal mixture of Chlamydomonas reinhardtii and Raphidocelis subcapitata cells) at a volume of 1.7 mL of algal solution per minute via a peristaltic pump, with a target cell concentration of between 10⁵ and 10⁶ cells/mL. Algal batch cultures were generated weekly to maintain food quality and consistency. Clam tanks were provided with a combination of cobbles and sand as a substrate. Aeration stones were used to maintain oxygen levels in the tank. Low mortality was observed, with 0–4 deaths per week in a population of approximately 300 individuals in the 2 months prior to the start of the experiment. Clams were not fed during the 5-day uptake phase as previous work demonstrated drops in aqueous Hg associated with the addition of algae, likely due to sorption to or uptake by the algae. The purpose of this experiment was to quantify the aqueous uptake of Hg rather than the dietary uptake. During the depuration phase, clams were fed with the same algal mixture that they had been fed for several months prior to the start of the experiment, once daily, to an initial cell density of 10⁵ cells/mL. This approximates the feeding regimes described in several other publications designed to provide adequate cell density without promoting excessive pseudofeces production [50,51,52,53].

2.2. Experimental Setup

Eight glass aquaria, each with a total capacity of 12 L, were rinsed twice with dechlorinated tap water and then filled with 8 L of dechlorinated tap water. Aeration stones were prepared for each tank to maintain oxygen levels. All treatments were prepared in duplicate, including the control tank. Three Hg treatment levels were selected to represent aqueous concentrations observed at Hg-contaminated sites (20, 100, and 200 ng/L). Between 18 and 20 Corbicula were added to each tank to provide at least 2 individuals for each sampling event. Corbicula were roughly separated into groups such that no single tank contained a disproportionate number of large or small individuals. This resulted in no significant difference in the whole weight of Corbicula between tanks (Figure S6, oneway ANOVA, F(7142) = 1.862; p = 0.08).

2.3. Dosing and Sampling

Clams were added to experimental tanks and allowed to acclimate for 2 days. Individuals were screened for valve movement prior to the start of the test as a health indicator, and all individuals displayed valve movement at the time of sampling. After 2 days of acclimation, the uptake phase of the experiment began and continued for 5 days. Hg sorption to container walls is a significant and known issue in Hg batch experiments [26,54]. To maintain consistent Hg concentrations in each tank, a 90% water change was performed daily, and tanks were dosed to the appropriate Hg concentration. This was done using a 40 mL stock solution of HgCl at a concentration of 24.4 ppm (confirmed by analytical measurement). The stock solution was stabilized with 200 μL BrCl. Five water samples were taken on day 1 to quantify any Hg loss to sorption or volatilization and to allow more precise modeling of aqueous Hg concentrations. Subsequent water samples were taken at regular intervals to provide daily measures of aqueous Hg. Water samples were taken by immersing borosilicate glass vials in each tank and quickly capping them. Clams were sampled at seven time points from each tank during the uptake phase, with two individuals taken at each sampling event. Three sampling events occurred on day 1, two on day 2, one on day 3, and one on day 5. During the two-week depuration phase, clams were sampled at one week and at two weeks, with two individuals taken during the first sampling event, and any remaining clams taken at the final sampling event. Additional clams were included in each tank to account for potential mortality.

2.4. Hg Analysis

Prior to the start of the experiment, aliquots of the stock solution were tested using an Ohio Lumex Portable Mercury Vapor Analyzer RA-915 Plus (Ohio Lumex Company, Inc., Twinsburg, OH, USA), and this concentration was used to accurately dose each tank. Water samples were prepared for analysis by the addition of 200 μL of BrCl shortly after sampling to stabilize Hg, maintain adequately low pH, and prevent sorption to the walls of the sample vial. Clam samples were collected from each tank using long metal forceps, starting in the control tank and progressing to increasingly high Hg concentration tanks to avoid cross-contamination of Hg. After each tank was sampled, the tweezers were rinsed with MilliQ water and dried thoroughly with Kimtech Kimwipes (Kimberly-Clark Worldwide, Inc., Roswell, GA, USA). Tanks were also covered with foil when sampling was not occurring to prevent contamination from splashing due to aeration.

Clam samples were then placed into plastic centrifuge tubes, tightly capped, and frozen at −20 °C to aid in tissue extraction at least overnight. At the end of the experiment (after approximately 5 days), clams were removed from the freezer, all external water was removed, and whole organism weights were recorded. Next, the soft tissue was dissected from the shells using tweezers and a scalpel, placed in 20 mL borosilicate glass vials, and freeze-dried for 48 h using a Labconco FreeZone freeze dryer (Labconco Corporation Kansas City, MO, USA). Both dry and wet weights were recorded but the discussion will focus on results on a dry weight basis unless specifically stated. Dry tissue was then digested using a 70:30 mixture of HNO₃ and H₂SO₄, followed by refluxing on a hot plate at 100 °C for 2 h. After refluxing, 100 μL of BrCl was added, the lid tightened, and the solution was given 24 h to fully digest at room temperature. Then, 5 mL of MilliQ was added to dilute the solution for analysis. This method is used for all tissue digestions for Hg analysis, as per ref. [37]. The tissue digested easily into a clear liquid with little visible particulate matter. Three Certified Reference Materials were digested in parallel with the clam tissue, IAEA 436A, IAEA 407, and IAEA 407 with 100 μL of Brooks Rand 1 ppm Hg standard (lot: 27365, p/n 03600), along with 2 method blanks.

All samples (water, clam tissue, and standards) were analyzed following EPA method 1631, revision E, or slight modifications of this method [55] similar to previous work [56]. Samples were analyzed via Lumex (cold vapor atomic absorption spectroscopy). Calibration curves were generated daily using a 5-point linear calibration by the addition of known volumes of a 10 ppm dilution of the same Brooks Rand total Hg standard. R² values of the regression were all greater than 0.9997. All reported Hg values for clams and reference tissue fell within the calibration curve range. Some water concentrations were below the calibration curve but still produced a visible peak. These values were reported as positive values, while any sample that did not produce a visible peak was reported as 0. The lower limit of detection for the Lumex as operated was approximately 25 pg of total Hg in a sample for a reproducible, visible peak, equating to 1.25 ng/L Hg in a 20 mL water sample. IAEA 407 was measured at 0.203 ± 0.002 mg/kg (n = 2, min/max: 0.203/0.2050, 0.222 mg/kg IAEA reported, equating to 91.4% recovery), IAEA 436A was measured at 3.91 mg/kg ± 0.06 (n = 2, min/max: 3.871/3.955, 4.26 mg/kg IAEA reported equating to 91.8% recovery). Mercury was below the detection limit in the method blanks. A 50 ppb Hg standard (Brooks Rand (Seattle, WA, USA), stabilized with 200 μL BrCl) was run at regular intervals to check for instrument drift and measured 51.17 ppm ± 1.03 (1 SD).

2.5. Statistical Analysis

All statistical analyses were conducted in R (R version 4.2.1). Cohen’s d values and confidence intervals were calculated according to ref. [57] and using the “compute.es” package. Linear models and ANOVA analyses were calculated using the “stats” package.

2.6. Toxicokinetic Modeling

A simple toxicokinetic model was fitted to these data to predict both Hg uptake and retention by Corbicula in environments beyond those explicitly tested here and compare our results to other studies of Hg uptake in bivalves. The model is a simple biodynamic model used extensively in the literature (see ref. [58] for a good review, Equation (1)) [58,59]:

$${\displaystyle \frac{d{C}_i}{dt}}=k_u{C}_W-k_e{C}_i$$

(1)

in which C_i represents the internal Hg concentration (ng Hg/mg dry weight clam tissue), k_u is the uptake rate (L mg dry weight clam⁻¹ min⁻¹), C_W is the aqueous concentration Hg (ng Hg/L), and k_e is efflux rate (min⁻¹). Since our experiment only exposed Corbicula through aqueous exposure and in the absence of food, dietary uptake can be ignored.

All fitting was performed using a likelihood method coded in BYOM (“Bring Your Own Model”) platform for parameter estimation, developed by Tjalling Jager http://debtox.info (accessed on 30 September 2023) for Matlab (Mathworks, ver. R2022a). Fitting routine details can be found in ref. [60]. Confidence intervals were estimated from likelihood profiles. Fitting of this model was performed in three steps:

(1)

The efflux rate parameter was fitted first only using the depuration data, assuming an exponential decline in internal Hg concentrations when the Corbicula were moved to clean water. The initial value of this phase was set by the concentration of each treatment level at the end of the uptake phase (Day 5) and we fitted the depuration data to Equation (1); in clean water, the concentration of Hg in the water was 0 (C_W = 0 ng Hg/L), so Equation (1) simplifies to an exponential decline function at a rate k_e.

(2)

During the exposure and uptake phase, the aqueous concentrations of Hg decreased with time due to sorption onto the tank walls and uptake into the Corbicula, so a constant value of C_W was not assumed in each treatment tank. To account for this decline and better estimate the value of k_u, an exponential decay function was fit to the measured aqueous Hg concentrations as the input for C_W (Equation (2)):

$$C_W={C_{0}e}^{-\lambda t}$$

(2)

with C₀ representing the initial Hg concentration (the nominal, dosed concentration was assumed in ng Hg/L) and λ representing the exponential decay rate parameter (min⁻¹). The same decay parameter was assumed across all treatments.

(3)

Once this equation was fitted, the value of k_u was fitted to Equation (1) using measured C_W Hg aqueous concentrations with estimated values connecting these data points assuming the fit exponential decay function (Equation (2)). The empirically measured and simulated data points were combined for the C_W input such that there was a data point every 200 min. The value of k_e was held constant at the value estimated using the depuration data (Step 1 above), assuming efflux was occurring simultaneously with uptake during the 5 day uptake period.

Bioconcentration factors (BCF) were calculated to compare to results from previous work looking at Hg uptake into bivalves using the following equation:

$$BCF={\displaystyle \frac{C_i}{C_W}}$$

(3)

where C_i is the concentration of Hg in tissue (ng/g) and C_W is the aqueous concentration of Hg (ng/L). This value was calculated using the average Hg clam tissue concentration of the two final organisms sampled during the uptake phase and the average aqueous Hg concentration during the uptake phase.

The kinetic rate constants from this experiment were used to calculate a steady-state tissue concentration in clams as a function of the aqueous Hg concentration using Equation (4) [58,59]

$$C_{i,ss}={\displaystyle \frac{k_u{C}_w}{k_e+g}}$$

(4)

where C_i,ss is the steady-state clam tissue Hg concentration, C_W is the aqueous Hg concentration, k_e and k_u are the efflux and uptake rates, respectively, and g is the growth rate. This value describes the equilibrium concentration of Hg in clam tissue as a function of aqueous concentrations.

3. Results and Discussion

3.1. Aqueous Concentrations

Hg concentrations in the three treatment levels varied throughout the day likely as a function of Hg sorption to the walls of the tanks and uptake into the clams (Figure 1). Maintaining low concentrations of Hg in solution is a recognized and common issue in tank experiments [26], as Hg readily adsorbs to many materials, causing samples to lose Hg from the aqueous phase if not stabilized appropriately, often with BrCl [26,54]. While this method is effective for sample preservation of inorganic Hg, it is not feasible for tests with live organisms, as the concentration at which BrCl maintains Hg in solution will likely result in high mortality due to low pH levels. Instead, Hg must be dosed at regular intervals to maintain a relatively constant aqueous concentration and sampled regularly to monitor Hg loss (Figure 1). Daily addition of Hg after water changes allowed concentrations to remain within 35–50% of the target concentration. All dosed tanks reached the appropriate concentrations shortly after dosing, and aqueous Hg concentrations in the tanks were modeled well by an exponential decay curve (Figure 1). The exponential decay rate constant was estimated to be 1.09 × 10⁻³ min⁻¹ (

Table 1. Toxicokinetic parameters.

Parameter	Description	Units	Parameter Estimate	Confidence Interval
λ	Exponential decay rate of aqueous Hg adsorbing to sides of container	min−1	1.09 × 10−3	[8.14 × 10−4–1.37 × 10−3]
ku	Uptake rate	L·g dry weight clam−1·min−1	7.04 × 10−3	[5.21 × 10−3–8.17 × 10−3]
ke	Efflux rate	min−1	5.27 × 10−5	[3.55 × 10−5–8.31 × 10−5]

) and fit the water data reasonably well (Figure 1). In control tank 1, a spike in Hg occurred at the start of the test that was not intentionally introduced as part of the experiment, resulting in elevated water concentrations for the first day of the uptake phase, but not seen thereafter. This may be due to residual Hg contamination on the sides of the aquarium that was not removed by the initial rinsing prior to dosing (Figure 1). Because of this, Control 1 was excluded from calculations in which treatments were compared to the control, and only Control 2 was used for these comparisons.

No mortality was observed in any of the tanks during the experiment, suggesting tolerance to acute Hg concentrations at or above 200 ng/L Hg in Corbicula. Previous short-term tests with Corbicula fluminea led to similar results, with no mortality at Hg concentrations as high as 500 ug/L [26,61].

3.2. Effect Sizes and Statistical Significance

By the end of the uptake phase, Hg concentrations measured in dry Corbicula tissue showed large effect sizes (Cohen’s d > 0.8) [62] between all three Hg concentrations and Control 2 (

Table 2. Hg concentration and statistical data for all treatments. Cohen’s D was calculated by comparing each treatment to Control 2. The last data points from the uptake phase (under the “Uptake on Day 5” section of the table) and the last data points from the depuration phase (the “Depuration after 2 weeks” section of the table) were compared to the Control 2 results for the same time point. BCFs were calculated based on the average Hg measured in the water during uptake (“Avg. Hg (Aq) (ng/L)”), and the average Hg concentration in the clam tissue at the end of the uptake phase. Confidence intervals (CIs) for Cohen’s D that include 0 indicate nonstatistical significance (p > 0.05), and CIs that do not include 0 indicate statistical significance (p < 0.05, marked with a *).

	Uptake on Day 5						Depuration after 2 Weeks
Treatment	Avg. Hg (Aq) (ng/L)	Hg (µg/g)	Std. Dev	BCF (L/g)	Cohen’s D	CI Cohen’s D	Hg (µg/g)	Std. Dev	Cohen’s D	CI Cohen’s D
Control 2	2.09	0.5	0.18				0.51	0.243
Control 1	2.62	1.51	1.64		0.87	[−0.15, 5.02]	0.545	0.367	0.11	[−0.71, 1.87]
20 ng/L (Rep 1)	14.93	1.06	0.28	71.1	2.43	[−0.91, 3.37]	0.653	0.244	0.58	[−0.99, 1.66]
20 ng/L (Rep 2)	14.31	0.77	0.26	53.8	1.23	[−0.87, 3.43]	0.611	0.339	0.33	[1.25, 5.25] *
100 ng/L (Rep 1)	68.45	0.92	0.43	13.4	1.28	[2.35, 14.97] *	1.424	0.306	3.25	[2.15, 7.21] *
100 ng/L (Rep 2)	72.7	3.63	0.48	49.9	8.66	[−0.23, 4.81]	1.539	1.539	4.68	[1.36, 5.19] *
200 ng/L (Rep 1)	140	5.52	3.1	39.5	2.29	[−0.54, 4.08]	1.795	0.459	3.28	[1.31, 5.75] *
200 ng/L (Rep 2)	140.1	3.96	2.77	28.3	1.77	[−1.18, 2.92]	1.392	0.256	3.53	[−1.16, 1.37]

), although the standard error of the effect size intersects with 0, suggesting the effect is not significant, except for the second 100 ng/L tank. This effect size was calculated for each individual treatment on the last sampling event of the uptake phase and compared to Control 2 at the same sampling event. Control 1 was excluded because of the Hg spike due to desorption on day 1, meaning both aqueous and clam concentrations are not representative of a Hg-free control treatment.

The confidence intervals of the Cohen’s D estimates were calculated to determine the magnitude and significance of the difference in mean Hg concentrations in clam tissue between the replicate tanks at the end of the uptake phase and the end of the depuration phase compared to the average tissue concentration in Control 2 from the same sampling time points. All treatments had positive Cohen’s D values, indicating that all treatments, even the lowest exposure concentration (20 ng/L), accumulated more Hg than the control (Table 2). This demonstrates a measurable Hg response at all Hg concentrations in this study. The assumption to compare each treatment to the entire sample set from Control 2 is supported by the regression of Hg concentration in dry clam tissue vs. time, which resulted in a slope of −0.007 µg/g/day and a p-value of 0.23, suggesting no change in Hg concentration with time in Control 2 (Figure S4).

Hg concentrations in Control 1 clam tissue were on average 3.04 times higher than Control 2 at the end of uptake, likely due to the spike in Hg seen in Control 1 at the start of the experiment, as discussed previously (Figure 1). This suggests that the Hg spike in Control 1, which was measured as less than 20 ng/L and lasted less than 12 h, resulted in measurable Hg uptake into clam tissue (Table 2).

3.3. Bioconcentration

Clam samples from the 20 ng/L tanks had between 1.55 and 2.14 times higher average tissue Hg than in Control 2 at the end of the uptake phase (Table 2). At the end of the 2-week depuration phase, Hg concentration in the clam tissue in the 20 ng/L tanks was still on average 23% higher than in Control 2 but was not statistically different than Control 2. These responses to very low levels of Hg demonstrate the sensitivity of the clam community to small deviations in Hg concentration and the persistence of those signals.

Bioconcentration was statistically significant for both uptake and depuration in the 100 ng/L and 200 ng/L tanks but with high individual variability in clam tissue concentrations (Figure 2). By the end of the uptake phase, clam tissue Hg concentrations were on average 4.6 times higher than Control 2 in the 100 ng/L tanks and 9.6 times higher in the 200 ng/L tanks. At the end of depuration, clam tissue Hg in the 100 and 200 ng/L tanks was on average 3 times higher than the control tanks. This rapid increase in tissue Hg concentrations followed by a gradual decrease demonstrates their value as a sentinel species, reacting quickly to a wide variety of Hg concentrations and retaining some portion of that signal through two weeks of depuration.

3.4. Bioconcentration Factors

Bioconcentration factors of Hg can vary widely between individuals and species but were relatively similar in all dosed tanks, with an average value of 42.7 L/g. Estimating the BCF from a previous study performed at 3 μg/L aqueous Hg yields a value of approximately 4.12 L/g at 5 days and 12.3 L/g at 14 days [26]. Despite nearly identical aqueous concentrations, the 100 ng/L tanks had very different BCFs, although the difference shrinks considerably if the highest individual in tank 100 ng/L, replicate 2 (7.7 µg/g) is not considered. This individual was the smallest clam by whole body mass and 1.97 standard deviations below the mean clam weight but within 10% of the next smallest individual. This low body mass may have influenced its Hg concentration, but clams of similar size did not exhibit large Hg concentrations relative to other individuals in their same treatment level, and overall no consistent, clear correlation was seen between body mass and Hg concentration (Figures S1–S3). While both tanks showed a significantly higher mean clam tissue Hg concentration than the control, the 100 ng/L tanks had significantly different means, while tanks Control 1 and Control 2, the 20 ng/L tanks, and the 200 ng/L tanks did not differ significantly. This suggests tank 100 ng/L, replicate 1, may have had some confounding factor that contributed to the closure of the clams in that tank.

3.5. Concentration and Biomagnification Variability

The variation in reported Hg concentration in bivalves suggests that many factors such as environmental chemistry, temperature, salinity, pH, and the species of bivalve being studied play a role in this metric. Studies looking at Hg uptake in a variety of bivalves in both freshwater and marine environments at a variety of temperatures found values on the order of 10² to 10⁵ L/kg (dry tissue basis) [63,64,65,66,67,68]. BCFs decreased with increasing aqueous Hg concentrations, suggesting perhaps a decrease in filtration behavior above 10 µg/L as observed by ref. [66], or other sublethal physiological factors as Hg levels approach toxic concentrations observed in juveniles [66,69]. BCFs in the lowest Hg systems (0.1–0.6 ng/L Hg) were between 10,000 and 240,000 L/kg and gradually increased as the length of exposure increased before reaching a steady state [63,64]. Lab studies were conducted at higher aqueous Hg concentrations, in the 10–100 µg/L range, and found concentration factors between 80 and 4300 L/kg, again decreasing with aqueous Hg concentrations [65,66,67,68]. These literature values agree with our study, with aqueous concentrations in the range of 20–200 ng/L, BCFs measured in the lab between 13 and 71 L/kg, and an estimated steady state BCF based on the toxicokinetic model of 134 L/kg.

High individual variability in tissue Hg was seen in all treatment levels. Prior to the start of this experiment, nine clams were randomly selected from the group used in this experiment, weighed, and analyzed as discussed previously for total Hg. These clams were living in a tank with a constant flow of dechlorinated tap water for two months with Hg concentrations below detection limits (less than 1.25 ng/L). The results suggested that among clams living in very low Hg water, some individual concentration differences can be explained by the total mass of the individual (which is also a reliable proxy for age in Corbicula [70]).

This trend was not seen in the clams sampled during this experiment, with much lower correlation coefficients between total body weight and tissue concentration in both control tanks, and variable strength and direction of correlation in dosed tanks, suggesting no consistent control in acute tests (Figure S1). Similarly, no clear, consistent correlation exists between Hg concentration and total mass, wet tissue mass, or dry tissue mass (Figures S2 and S3 show correlations for dry tissue). This suggests that the effect of body size seen in Figure S5 may be less relevant in acute exposures. We hypothesize that size could be important when clams are kept at relatively constant Hg concentrations for long periods of time, as they are allowed to all reach equilibrium with their environment, and larger clams have lived longer and have had the opportunity to be exposed to more total Hg. Because the goal of this experiment was to develop a toxicokinetic model applicable at a range of aqueous Hg concentrations, and no consistent significant correlations were seen, the total body mass of the clam was not considered in this model.

Corbicula have been observed to maintain long periods of shell closure in the presence of environmental contaminants [71,72]. Shell closure reduces base metabolic activity by 90%, drastically reducing interaction with the environment [71]. Tran et al. found that Corbicula had an average time to closure in the presence of 20 μg/L Hg of 3 min, as compared to 257.2 min for no added Hg, but did not explore concentrations below 20 μg/L [73]. Our tests were 2–3 orders of magnitude below this concentration, so the effect of shell closure on Hg uptake may or may not be relevant to this study. Even though shell closure was not observed during the entire study, this could still account for some individual variation, which leads to larger standard deviations and lower statistical certainty. Despite this, results showed that at aqueous concentrations well below the EPA drinking water maximum contaminant level (2 μg/L), Corbicula exhibit significant bioconcentration, with elevated tissue concentrations even after 2 weeks of depuration (Figure 2).

3.6. Toxicokinetic Model

The simple toxicokinetic model fit the Hg bioconcentration data reported in this paper well (Figure 3).

As an additional attempt at model verification, the toxicokinetic model (Equation (1)) was simulated using parameters estimated from the data reported here (parameters in Table 1) and this prediction was compared to data from ref. [30] that also measured the uptake of Hg into Corbicula fluminea. Internal concentrations of Hg in Corbicula were reported on a wet weight basis in ref. [30], so these data were first converted to g⁻¹ dry weight for comparison using the dry wet to wet weight conversion measured in the data set reported in this paper (dry wet/wet weight of all Corbicula across all treatments = 0.023 ± 0.008 [mean ± standard deviation, n = 150]). Ref. [30] also only reported nominal aqueous Hg concentrations of 1.45 and 5 µg/L, but the authors noted in the results of this paper that the average decrease in concentration across the daily dosing cycles for Hg was 50%, so concentrations of 0.725 and 2.5 µg/L were assumed as the aqueous Hg concentrations (C_w model input). The fit of the model simulation is shown in Figure 3—the model fit to the data presented in this paper predicts a naïve data set [30], especially for the long-term Hg concentrations in the Corbicula. The model overestimates the transient Hg concentrations at the higher Hg dose, however, it fits the final measurements well (Figure 4).

The parameter estimates reported here (Table 1) also compare well with previous work on other bivalve species. Based on our literature review, no prior work quantifies Hg uptake and depuration kinetic parameters in freshwater mussels or clams. Ref. [64] defines kinetic parameters for Hg in a variety of marine scallop, clam, and mussel species. Efflux rates (k_e) for the marine species varied between 0.021 and 0.060 day⁻¹. Ref. [68] calculated a value of 0.02 to 0.074 day⁻¹ in Crassostrea virginica (American Oyster), a brackish water species. Both ranges are slightly lower than the 0.076 day⁻¹ calculated for Corbicula in this study. These values are quite close considering the differences between the studies, including the species being tested, marine/brackish vs. fresh water, and Hg concentration both in the water and in the mollusks. Ref. [64] found uptake rates (k_u) varied between 3.5 and 32.8 L g⁻¹ d⁻¹, ref. [74] found values between 3.8 and 11.3 L g⁻¹ d⁻¹, ref. [53] found a value of 32.8 L g⁻¹ d⁻¹, and ref. [65] found a value of 35 L g⁻¹ d⁻¹, while our study estimated a k_u of 10 L g⁻¹ d⁻¹, falling within the range seen in the marine species. The values from ref. [74] were calculated in terms of dry tissue mass by assuming 97.7% moisture, the mean moisture content calculated in this study. Several of the studies measured Hg uptake by diet, which results in higher Hg uptake rates compared to aqueous exposure alone, which our current study focused on. This suggests that uptake rates would be higher if both aqueous and dietary exposure were considered. This comparison shows considerable agreement between kinetic uptake and efflux parameters for Hg among the variety of mollusks, in a wide variety of environmental conditions across multiple studies. These values suggest a degree of consistency of Hg uptake in mollusks, as well as high steady-state Hg concentrations.

3.7. Implications for Metal Sequestration

Sequestration of metals is one of many ecosystem services provided by Corbicula [26,28,29,30,50]. Because of their rapid proliferation, ubiquity, tolerance for contaminants, and high affinity for a wide variety of metals, Corbicula have been proposed as a mechanism for remediating high metal, low pH acid mine drainage [29]. Other studies have considered bivalves more generally as a remediation technique paired with freshwater pearl aquaculture and found that a moderately sized oyster farm (100 tons total oyster mass per year) could remove as much as 300 kg of metals and 24 kg of organic contaminants per year from the water simply by harvesting the oysters grown that year [75]. This suggests that Corbicula or other freshwater mussels may be valuable as remediators for Hg-contaminated streams.

Engineered sorbents have been proposed as a remediation technique on Hg-contaminated streams with some promising results, but sorption partitioning coefficients are greatly impacted by the presence of organic matter (60 mg/L DOC), by as much as 3 orders of magnitude among the most efficient sorbents [33]. Along with Hg, Corbicula also filter considerable volumes of water, algae, organic material, and other particulate matter from the water column [15,16,21,22,23,24]. Corbicula have high rates of growth, large population densities, and high bodyweight-specific filtration rates of Corbicula which can exceed native populations [8,18]. This suggests that Corbicula populations already established in a system or deployed in a contained manner alongside engineered sorbents may greatly enhance the effectiveness of a sorbent-based engineering solution by lowering the concentration of organic compounds and particulate matter that may impede the optimal function of the deployed sorbents.

Partitioning coefficients (K_d) are used to describe the potential of the sorbent to adsorb or absorb a given contaminant and are calculated by dividing the concentration of the contaminant in the sorbent by the concentration of the contaminant in water with units of volume/mass [33,76,77]. Multiple sorbent materials based on clay, carbon, and mesoporous silica have been studied for Hg uptake potential to possibly be deployed in impacted streams [33]. Testing resulted in K_d values between 10² and 10⁵ mL/g, strongly influenced by the presence of organic matter and varying between sorbents.

To calculate an equivalent parameter for Corbicula, the steady-state equilibrium concentration equation was re-arranged to solve for the concentration of Hg in clam tissue over the concentration of Hg in water at a steady state. Assuming a low and thus negligible growth rate [58,59] (lab studies measuring Corbicula somatic growth rates find values ranging from 0.01–0.03 day⁻¹ which changes estimated internal concentrations by less than 1%) [78], this value becomes the ratio of uptake (k_u) to efflux (k_e), which is equal to the concentration of Hg in the clam over the concentration of Hg in the water. This value was calculated to be 1.34·10⁵ mL/g when considering only the mass of dry tissue and 1.29·10³ mL/g when considering the entire mass of the clam, including shell material at a circum-neutral pH (measured by pH strip to be around 7) and 20 °C. This places Corbicula within the range of values seen in the engineered sorbents (10²–10⁵ mL/g [33]), even when only considering aqueous Hg uptake and ignoring dietary assimilation. This suggests that direct Hg uptake is another valuable ecosystem service provided by Corbicula and may be similar in effectiveness to other sorbent materials, although additional testing under identical conditions would be necessary to confirm this. Additionally, the uptake of organic material by Corbicula may be synergistic with other sorbent materials, improving the total Hg uptake performance in engineered systems. It must again be emphasized, however, that Corbicula’s potential as a bioremediator of Hg does not outweigh the potential disastrous results of introducing more of an invasive species into a system. The comparison of Corbicula to engineered sorbents for Hg bioconcentration is performed here to demonstrate the impact these clams are already having on impacted systems they have invaded (which are most freshwater systems) and to suggest the controlled use of these clams to remove aqueous Hg (as discussed in ref. [29]).

Further, Hg bioconcentration by Corbicula fluminea may have a positive impact on the trophic system in which they are present as they sequester aqueous inorganic Hg, removing it from the water column before it can be methylated and turn into the much more bioaccumulative and toxic form of Hg, MeHg. This has been observed in marine systems with saltwater mussels [79] and freshwater systems with Corbicula fluminea (Figure S7).

4. Conclusions

This paper provides the first instance of toxicokinetic modeling of Hg uptake and efflux rates in Corbicula. Elevated Hg concentrations were evident in tissue at aqueous concentrations relevant to Hg-contaminated streams, but well below regulatory drinking water limits, demonstrating their utility as a sensitive sentinel species and potential bioremediator. At aqueous concentrations as low as 20 ng/L, Hg uptake was evident by the end of the experiment and remained elevated through 2 weeks of depuration, although this effect was small in the lowest treatment level. Kinetic constants were similar to values found by previous studies in other mollusks, despite differences in species, temperature, and salinity, suggesting other freshwater mollusk species may have similar uptake and efflux parameters to their marine counterparts. The Hg uptake capacity of Corbicula calculated from the toxicokinetic model was determined to be similar to engineered sorbents proposed as possible remediation technologies for Hg on a contaminated stream if introduced in a controlled manner that restricts the movement of this invasive species. This study demonstrates the value of Corbicula as a sentinel species, as a vector for Hg sequestration, and suggests their potential importance for water quality and possible utility as a tool for future remediation efforts involving natural and engineered sorbents.