Maternal body burdens of methylmercury impair survival skills of offspring in Atlantic croaker (Micropogonias undulatus)

Aquatic Toxicology 80 (2006) 329–337

Maternal body burdens of methylmercury impair survival skills of offspring in Atlantic croaker (Micropogonias undulatus) Mar´ıa del Carmen Alvarez a,∗ , Cheryl A. Murphy b,1 , Kenneth A. Rose b , Ian D. McCarthy a,2 , Lee A. Fuiman a b

a The University of Texas Marine Science Institute, 750 Channel View Drive, Port Aransas, TX 78373, USA Louisiana State University, Department of Oceanography & Coastal Sciences/Coastal Fisheries Institute, Energy, Coast and Environment Building, Baton Rouge, LA 70803, USA

Received 28 June 2006; received in revised form 29 September 2006; accepted 30 September 2006

Abstract Methylmercury (MeHg), the organic form of mercury, bioaccumulates easily through the food chain. Fish in high trophic levels can accumulate substantial levels of MeHg and transfer it to their developing eggs. Here, the effects of maternally derived MeHg on the planktonic larval stage of Atlantic croaker were investigated. Adult Atlantic croaker were fed MeHg-contaminated food at three levels: 0, 0.05, and 0.1 mg kg−1 day−1 for 1 month. Fish were then induced to spawn and MeHg levels in the eggs were measured (0.04–4.6 ng g−1 ). Behavioral performance of exposed and control larvae was measured at four developmental stages: end of yolk absorption (yolk), end of oil absorption (oil), and 4 and 11 days after oil absorption (oil + 4 and oil + 11). Behaviors analyzed included survival skills related to foraging and predator evasion: routine behavior (rate of travel, active swimming speed, net-to-gross displacement ratio, and activity) and startle response to a visual and a vibratory stimulus (responsiveness, reactive distance, response distance, response duration, average response speed, and maximum response speed). Maternally transferred MeHg induced concentration-dependent effects on survival skills. Statistical and simulation models applied to predict the ecological consequences of the behavioral effects suggested that maternal transfer of MeHg may substantially lower survival of planktonic stage larvae compared to unexposed larvae. These results also imply that larvae of top predatory fish species, such as blue marlin, may suffer mortality through maternal transfer of MeHg. © 2006 Elsevier B.V. All rights reserved. Keywords: Methylmercury; Fish larvae; Endocrine disrupting chemical; Behavior; Sublethal; Individual-based model

1. Introduction Mercury (Hg) is a naturally occurring heavy metal. Natural processes and anthropogenic activities release Hg into the atmosphere as elemental and inorganic Hg where it is readily transported (Downs et al., 1998). Anthropogenic sources of Hg are many; however, the largest emissions are from coal and fossil-fuel burning (Moore, 2000). Non-point-source deposition occurs mainly in the form of rain, therefore, Hg pollution is widespread (Downs et al., 1998). Hg is transformed into methylmercury (CH3 Hg2+ , MeHg) mainly by bacterial action ∗

Corresponding author. Tel.: +44 1482 887 417. E-mail address: [email protected] (M.d.C. Alvarez). 1 Current address: Deptartment of Zoology, University of Toronto, 25 Harbord Street, Toronto, Ontario M5S 3G5, Canada. 2 Current address: School of Ocean Sciences, University of Wales Bangor, Askew Street, Menai Bridge, Anglesey LL59 5AB, UK. 0166-445X/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.aquatox.2006.09.010

within the sediment (Benoit et al., 1998). Organisms readily take up MeHg by direct exposure through the body surface or by ingestion, and it can then bioaccumulate and biomagnify through the food chain (Lawrence et al., 1999). Concentrations of MeHg in water or sediments are generally less than 30% of total Hg, while more than 90% of total Hg in fish muscle tissue is in the form of MeHg (Watras and Bloom, 1992). Adult Atlantic croaker (Micropogonias undulatus) are demersal and feed mainly on benthic organisms (worms, clams). Spawning occurs along the continental shelf from mid-summer to spring (Miller et al., 2003). Pelagic larvae stay in the open ocean for approximately 30–60 days (Nixon and Jones, 1997), arriving at estuarine nursery areas at an average length of 11 mm (Diamond et al., 1999). Benthic feeding organisms are particularly susceptible to MeHg exposure through ingestion of contaminated food and sediments or through their gills as they sift MeHg-contaminated sediments and water (Hall et al., 1997; Hanson and Zdanowicz, 1999). However, approximately 90%

330

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

of the MeHg accumulated in fishes is of dietary origin (Hall et al., 1997; Spry and Weiner, 1991). Ingested MeHg is quickly absorbed, circulated, and deposited in tissues (Mason et al., 1995). In fishes, MeHg mainly resides in proteins (Mason et al., 1995) and can be transferred from the female to the eggs during oogenesis (Latif et al., 2001; Hammerschmidt and Sandheinrich, 2005). Maternal transfer of MeHg to oocytes is believed to occur through the mechanism of molecular mimicry. MeHg readily binds with the amino acid cysteine to form MeHg-l-cysteine. This complex is structurally similar to methionine and is transported across the cell membrane by methionine transporters (Ballatori, 2002; Simmons-Willis et al., 2002) during oogenesis (Fyhn, 1989). Therefore, female Atlantic croaker inhabiting contaminated areas could transfer part of their body burden of MeHg to their offspring through the yolk. MeHg is a known endocrine disrupting chemical (EDC) and neurotoxicant (Colborn et al., 1993; Rice, 1995; Zhou et al., 1999; Myers et al., 2000). Endocrine and neural systems control many physiological and developmental processes in animals (Randall et al., 1997). Any disturbance to either of these systems during development could have profound and permanent effects on the organism, both in terms of its development and its performance in ecological interactions. Therefore, it is likely that fish larvae, as rapidly developing organisms, will be highly susceptible to MeHg burdens of maternal origin. Faulk et al. (1999) and McCarthy et al. (2003) have shown that exposure to EDCs  (o,p -DDT and Aroclor 1254, respectively) through maternal deposition in the egg can impair the performance of Atlantic croaker larvae in behaviors relevant to foraging and predator evasion. Here, we evaluate the effects of maternal MeHg exposure on the ecological performance of Atlantic croaker larvae by analysis of routine and predator evasion behaviors. Statistical and simulation models are then applied to the behavioral data to evaluate the potential effects of MeHg exposure to planktonic stage larval survival.

MeHg. All groups were fed ad libitum. Fish were maintained on these diets for 1 month after which a pair (one female and one spermiating male) was removed from each dose level tank in the afternoon for spawning. The female received a single injection of gonadotropin releasing hormone analog (GnRHa) in fish saline at a concentration of 50 ng g−1 of body weight to induce spawning (Zohar, 1989). The next morning, spawned eggs were collected and a sample was frozen for later analysis of MeHg content, using a modification of the digestion procedure in the method described by Dusci and Hackett (1976). In this case the tissue was left to dissolve for 1 week in 50 ml of HNO3 solution at room temperature prior to analysis. Eggs were collected from 6 control, 5 low, and 4 high dose spawns. However, only spawns that survived to 4 days after absorption of the oil globule were used. Therefore, larvae from 3 control, 2 low, and 3 high spawns were analyzed. Upon collection, eggs were disinfected to remove possible parasites by immersing the eggs in a solution of 1 ppm of formalin for 45 min then placing them in glass watch bowls filled with 1.5 l of sea water at a density of 2 eggs ml−1 to hatch. The following morning, hatched larvae from each spawn were transferred to two 150 l conical rearing tanks at a density of 20 larvae l−1 . Larvae were reared in constant conditions of photoperiod (12L:12D), salinity (approximately 29.5 PSU), and temperature (23 ◦ C). For the experimental period, larvae were fed rotifers (Brachionus plicatilis) enriched with algae (Isochrysis galbana). Larvae were sampled for behavioral assays at developmental stages identified by the completion of yolk absorption (referred to as “yolk”), completion of oil absorption (“oil”), and on days 4 (“oil + 4”) and 11 (“oil + 11”) after complete oil absorption. The first two stages were chosen to help identify the primary source of MeHg in the larvae (yolk or oil) and the latter two stages were chosen to determine whether the effect of MeHg was a temporary physiological impairment or a persistent developmental damage.

2. Materials and methods

2.2. Growth

2.1. Experimental fish

Five larvae were sampled from each rearing tank on days 1, 3, 6, 11, and 17 after hatching. Larvae were anesthetized with tricaine methane sulfonate (MS-222, 1%, v/v) and their total length (TL, mm) to the nearest ␮m was measured with the aid of a microscope and computer-assisted measuring system (Measurement TV, Data Crunch software). Growth rates (G) were computed as the slope of the exponential growth model:

Sexually mature Atlantic croaker were fed rations contaminated with different levels of MeHg and the larvae they produced were studied for effects on growth and behavior that would forecast reduced probability of survival in their natural environment. Male and female Atlantic croaker were collected with gill nets in early winter 2000 in the Aransas Pass Ship Channel (near Port Aransas, Texas). Fish were maintained in 4000 l tanks with recirculating water at constant temperature (22 ◦ C), 12L:12D photoperiod, and fed a diet of shrimp (3% body weight per day). From 1 December onward, three groups of fish (average body weight 370 g) consisting of 16 females and 8 males were placed into tanks of approximately 10,000 l capacity with recirculating water. The control group was fed only shrimp. The low dose group was fed a diet of blue marlin (Makaira nigricans) muscle tissue having a concentration of 0.05 mg kg−1 of MeHg. The high dose group was fed marlin supplemented with contaminated shrimp to a final concentration of 0.1 mg kg−1 of

TL = a eGt where t is age in days after hatching. 2.3. Routine behavior assays Routine behavioral assays were conducted on larvae from the eight successful spawns (3 control, 2 low, and 3 high dose) when they reached the experimental stages. Twenty larvae for yolk or oil stages and 15 larvae for oil + 4 and oil + 11 stages were placed in experimental chambers (glass, 2.5 cm × 7.5 cm × 2.5 cm) and allowed to acclimate and recover from handling for 2 h (as

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

recommended by Fuiman and Ottey, 1993) in a temperaturecontrolled room at 27 ◦ C. Then the chambers were carefully placed above an infrared sensitive video camera (Cohu, model 3315-2000/0000) and left undisturbed for 5 min to allow the larvae to recover from handling. Larval routine behavior was then recorded from below using a video recorder (Panasonic, model AG-1960). Experiments were conducted under visible light which was supplemented with indirect infrared light. Video segments were digitized as AVI files and movements of the larvae were analyzed with the aid of a computerized tracking system (WinAnalyze 2D Software, Version 1.5, Mikromak, Germany). To characterize routine swimming, measurements of rate of travel (mm s−1 ), active swimming speed (mm s−1 ), net-to-gross displacement ratio (NGDR), and activity (% of time) were made from randomly selected 30 s video segments. Atlantic croaker larvae swim in alternating episodes of active swimming and resting. Rate of travel is a measurement of the time-averaged swimming speed during the measurement period, including resting periods. Active swimming speed describes the average swimming speed while the fish was actively swimming (excluding the resting periods). NGDR is a measurement of the linearity of the swimming path traveled by a larva, where net displacement is the linear distance between the beginning and the ending points of the measurement period, and gross displacement is the actual distance covered by the larva along its swimming path. The closer NGDR is to 1, the straighter the path swum by the larva. 2.4. Predator evasion behavior 2.4.1. Visual startle stimulus The visual startle assay was conducted immediately after the routine behavior assay. It elicited an evasive response in larvae similar to that used by a larva to escape from a predator (Fuiman and Magurran, 1994). The stimulus was a black oval on a white card, simulating the cross-sectional silhouette of a predatory fish, held at the end of a remotely controlled pendulum. The pendulum was held away from the larvae by an electromagnet. When the pendulum was released the predatory stimulus accelerated toward the larvae but a tether stopped it just before hitting the chamber containing the fish (depicted by Fuiman and Cowan, 2003). The whole procedure was videotaped for later video analysis. A total of 100 video frames were analyzed, beginning 50 frames before the maximum extent of the stimulus. Frame-by-frame analysis of these video recordings provided data on responsiveness (percentage of larvae responding), reactive distance (distance from the pendulum to a larva when the response started, in mm), response distance (total distance covered by a larva during the response, in mm), response duration (total duration of the response, in ms), average response speed (response distance divided by response duration, in mm s−1 ), and maximum response speed (greatest speed observed between two consecutive video frames, in mm s−1 ). 2.4.2. Vibratory startle stimulus The vibratory startle stimulus consisted of a remotely controlled metal hammer placed at about 5 mm from a metal post

331

upon which a plastic dish containing larvae rested, as described by Faulk et al. (1999) and McCarthy et al. (2003). Groups of 10 larvae were placed in transparent plastic dishes (50 mm diameter × 12 mm) and allowed to acclimate from handling for 2 h. After this period the dish containing the larvae was placed on top of the metal post and surrounded by a ring of infrared lightemitting diodes. After allowing 5 min for the larvae to recover, the hammer was remotely triggered to strike the metal post and the responses of larvae were recorded with an infrared sensitive video camera (Cohu, model 3315-2000/0000) placed above the dish. This was repeated two to four times for each spawn and developmental stage. Video recordings were then analyzed using a computerized video measurement system (Measurement TV, Data Crunch software). The percentage of larvae responding to the stimulus (responsiveness) was calculated from the number of visible larvae. Mean response speed of each larva (mm s−1 ) was calculated from the response distance (mm) and the response duration (ms). 2.5. Statistical analyses Statistical analyses of the behavioral and growth data were conducted using SYSTAT software (Version 10.0). Mean values for each variable were computed for each spawn, treatment, and developmental stage, and used for the statistical comparisons. Variables were screened for normality, and logarithmic or angular (arcsine) transformations were applied when necessary (Zar, 1999). Measured concentrations of MeHg in the eggs were separated roughly by a factor of 10, therefore, log-transformed MeHg concentrations (larval body burdens) were used in all statistical analyses. Concentrations of MeHg in control eggs were below detection limits (0.001 ng g−1 ), so a value of 0.0002 or 0.0004 was assigned to the samples in which measured MeHg concentrations were below the detection limit to allow logarithmic transformation without loss of data (Table 1). All variables were first examined using analysis of covariance (ANCOVA) with MeHg concentrations in eggs and age as covariates. In the case of a significant concentration × age interaction or a significant concentration effect, linear regressions were computed for each age to detect any concentration-dependent trends.

Table 1 Mean concentration of MeHg (in ng g−1 of spawned eggs) from each of three treatment groups Control (0 mg kg−1 day−1 )

Low (0.05 mg kg−1 day−1 )

High (0.1 mg kg−1 day−1 )

0.0010 – 0.0012 0.0014 –

– – – 0.639 0.294

0.567 4.574 – 3.874 –

Values in brackets represent the nominal dietary exposure given to the broodstock. Only spawns with successful survival through the oil + 4 stage were used. Values below detection limits were assigned a value of 0.0004.

332

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

2.6. Statistical and simulation models For the purpose of simulations, experimental results were grouped by exposure level into control (MeHg in eggs < 0.05 ng g−1 ), low dose (0.05–1.0 ng g−1 ), and high dose (>1.0 ng g−1 ) groups. Maternally derived MeHg effects on croaker larvae were analyzed using a regression tree (RT), and the RT outputs were used as inputs to a larval stage individualbased model (IBM). The IBM was employed to predict how the MeHg effects measured in the lab would translate to effects on planktonic stage larval survival, mortality, duration, and growth. Details of the methods were provided by Murphy (2006) and Murphy et al. (in prep.) and are summarized here. 2.6.1. Regression tree (RT) An RT relationship developed for red drum larvae (Fuiman et al., 2006) was adapted for use with croaker larvae. The objective was to identify the survival skills that are best predictors of the probability of croaker larvae escaping the attack of a predatory fish (the presumed primary source of mortality at this stage). Fuiman et al. (2006) used RT analysis to find that a few survival skills – including rate of travel, visual reactive distance, and acoustic response distance – measured on red drum larvae (a close relative of Atlantic croaker), could be used as a rough predictor of the probability of escaping a real predator attack. A standardization of the RT was done in order to apply the survival skills data collected from croaker larvae to the RT estimated for red drum larvae, and from that predict how contaminants would affect a larva’s probability of escaping an attack by a predatory fish (Murphy, 2006; Murphy et al., in prep.). Standardization involved three steps: (1) variables measured on red drum were transformed to satisfy normal probability distributions, (2) the RT was re-fit to the transformed red drum data, and (3) splitting values at the nodes from the re-fit RT were converted to z-scores. Probabilities of croaker larvae escaping a predatory fish attack were predicted using the standardized RT and the measured survival skills of the croaker larvae for each developmental stage (yolk, oil, oil + 4, and oil + 11) separately for control, low, and high MeHg exposures. The measured croaker rates of travel and visual reactive distances were transformed to conform to normal probability distributions. Rate of travel was square-root transformed, and visual reactive distance was transformed using a Box–Cox transformation with λ of 0.4 (SAS Language Reference, Version 9, SAS Institute Inc., USA). Then z-scores for each of the survival skills used in the RT were calculated using the mean and standard deviation from the transformed control data. The measured survival skills of each croaker larva, now expressed as z-scores relative to control means and standard deviations, were inputted to the RT and the probability of escaping a predatory attack was predicted. Predicted probabilities of escaping a real predator’s attack were accumulated for all larvae measured in each developmental stage for the control, low, and high MeHg exposures. The original values of rate of travel and the probability of escaping a predator predicted by the standardized RT were converted into multipliers for use in the individual-based model (IBM) of the larval stage. Each rate-of-travel value and prob-

ability of escaping a predator attack was divided by the mean value in the control group for each developmental stage. This resulted in frequency distributions for rate-of-travel multipliers and probabilities of escape multipliers for each developmental stage for control, low, and high MeHg exposures. 2.6.2. Individual-based model The multipliers for rate of travel and probabilities of escaping a predator were used as inputs to an individual-based model (IBM) that then converted these to larval stage duration and survival. The IBM tracked the daily growth and mortality of croaker larvae as larvae grew from 2.5 to 11 mm. We treat survival to 11 mm as larvae surviving the planktonic larva life stage (Diamond et al., 1999). In our simulations, growth depended on encounters with zooplankton prey, and mortality depended on encounters with individual sea nettle, ctenophore, and fish predators. Rate of travel influenced the growth and mortality rates by affecting encounters with prey and with all three predators, while probability of escape only influenced mortality by fish predation. Growth of larvae was simulated using a bioenergetics model that incorporated consumption, egestion, metabolism, and specific dynamic action. All were expressed as functions of larva weight. Consumption was determined by simulating encounter rates and capture of each of four zooplankton prey types (invertebrate eggs, copepod nauplii, copepodites, and adult copepods) by each larva. The growth component of the IBM was adapted from a generic larval model developed by Letcher et al. (1996). Starvation and predation mortality were simulated daily. If the condition of a larva (simulated weight relative to weight expected from its length) was below the starvation threshold value, the larva was simulated to starve to death. Mortality due to predation was determined by simulating larva vulnerability to predators based on their encounter and capture by three predator types (sea nettles, ctenophores, and fishes). Information on predators used to determine encounter rates with the larval fish, such as predator lengths, rates of travel, encounter radii, and capture success, were obtained from Cowan et al. (1996). A multiplier for rate of travel and probability of escape was calculated for every simulated larva in the initial number of 10,000 and as each larva entered each new developmental stage. Multipliers were randomly assigned to larvae in proportion to their values from the frequency histograms, and when each larva entered each developmental stage, the larva received multiplier values from the new developmental stage independently of its previous multiplier values (i.e., no memory in the multipliers from one developmental stage to the next). 2.6.3. Model simulations The RT and IBM were used to simulate planktonic larval stage duration (which includes the four developmental stages: yolk, oil, oil + 4, and oil + 11 as the individual larva grows from 2.5 to 11 mm TL) and survival for the control, low, and high MeHg exposure groups. Each IBM simulation started with 10,000 larvae, and simulations were repeated three times using different random number sequences that affect realized encounter rates and other stochastic aspects of the model. The difference among

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

333

the control, low, and high exposure simulations was the RTpredicted values for rate-of-travel multipliers and probabilityof-escaping multipliers for each developmental stage. 3. Results Since the feeding rate of each adult croaker varied, MeHg concentrations in the eggs varied among spawns within nominal treatment groups. Analysis of the egg samples showed six distinct MeHg levels (Table 1). For analysis of effects on behavior, dose groups were assigned based on MeHg levels measured in the eggs rather than nominal concentrations in the diet. 3.1. Growth There was no significant effect of MeHg exposure on larval growth over the duration of this study (P = 0.267). Instantaneous growth rates for control, low and high exposure groups were 0.034, 0.032, and 0.033 day−1 , respectively. Overall Atlantic croaker larva growth followed the equation: TL = 2.21 e0.033t where TL represents the total length in mm and t is the number of days (R2 = 0.82; P < 0.001; n = 270). 3.2. Routine behavior assay There was not a significant concentration × stage interaction nor developmental stage effect on the rate of travel of exposed and control larvae (P = 0.763 and 0.145, respectively). However, the rate of travel varied significantly with the MeHg concentration in the eggs (P = 0.028) reflecting the decreasing rate of travel with an increasing MeHg exposure level (Fig. 1). Unexposed larvae exhibited an average rate of travel of 0.865 mm s−1 ± 0.135 (± S.E.) while MeHg exposed larvae showed an average rate of 0.552 ± 0.052 (± S.E.). Active swimming speed did not vary

Fig. 1. Effect of maternally derived MeHg exposure on routine behavioral trait, rate of travel, of Atlantic croaker larvae. Data points represent mean values for each spawn and developmental stage. Data points flowed significant regression (P = 0.037) described by the equation y = −0.058x − 0.295 (R2 = 0.142, N = 31).

Fig. 2. Effect of maternally derived MeHg exposure on routine activity of Atlantic croaker larvae. Data points represent mean values for each spawn and developmental stage. Angular transformation was performed to attain normality. Data points flowed significant regression (P = 0.042) described by the equation y = −0.030x + 0.459 (R2 = 0.136, N = 31).

significantly with MeHg level (P = 0.061), developmental stage (P = 0.478) or concentration × stage interaction (P = 0.418). Activity levels were significantly related to MeHg concentration (P = 0.044, Fig. 2), although no significant effect of development (P = 0.306) or interaction term (P = 0.445) was observed. On average control larvae were active 28.27 ± 3.81 % of the time while MeHg treated larvae were active only 19.88 ± 1.64 % of the time (± S.E.). 3.3. Predator evasion behavior assays 3.3.1. Visual startle stimulus No significant relationship with MeHg concentration was observed for any of the seven variables computed from the visual startle assay. MeHg concentration in the eggs did not affect the proportion of larvae responding to the stimulus (P = 0.436). Larvae from all treatments reacted at similar distance from the stimulus (P = 0.518), and their responses lasted for comparable amounts of time (P = 0.718) at all developmental stages. MeHg concentration in the eggs did not affect larval response distance (P = 0.130) or the average (P = 0.134) or maximum (P = 0.154) response speeds. 3.3.2. Vibratory startle stimulus Control and MeHg-treated larvae were similarly responsive to the vibratory startle stimulus (P = 0.620). A significant effect of MeHg concentration (P = 0.013) was observed for the response duration to the vibratory stimulus (Fig. 3). The analysis showed that the duration of responses by larvae to the vibratory stimulus were positively related to the MeHg body burden, where the average response duration for control larvae was 50.14 ± 4.81 ms compared to 63.09 ± 4.31 ms for MeHgtreated larvae. A significant effect of MeHg concentration on response speed (P = 0.023) was also observed. Response speed decreased significantly with increasing MeHg burden in the eggs

334

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

Fig. 3. Effect of maternally derived MeHg on response duration to a vibratory startle stimulus of Atlantic croaker larvae. Behavioral traits were logarithmically transformed to attain normality. Data points represent mean values for each spawn and developmental age. The equation y = 0.039x + 1.77 describes the significant regression (P = 0.026) between the response duration and MeHg concentration in the eggs (R2 = 0.159, N = 31).

Fig. 4. Effect of maternally derived MeHg on response speed to a vibratory startle stimulus of Atlantic croaker larvae. Behavioral traits were logarithmically transformed to attain normality. Data points represent mean values for each spawn and developmental age. The significant regression (P = 0.0.19) is described by the equation y = −0.04x + 1.619 (R2 = 0.177, N = 31).

3.4. Statistical and simulation models

ing an attack did not follow a typical dose–response relationship of decreasing with increasing MeHg exposure levels. For most stages except for oil + 11, low exposure to MeHg caused the greatest reduction in the probability of escape; high exposure to MeHg resulted in probabilities of escape that were intermediate between control and low exposure values. The oil + 11 larvae were the only group that showed the lowest probabilities of escape at the highest MeHg exposure.

3.4.1. Regression tree Analyses suggested that maternal exposure to MeHg caused a small, but consistent, decline in the probability of escaping a predatory fish attack and that the effect varied by developmental stage of the larva (Table 2). However, the probability of escap-

3.4.2. Individual-based model Simulations suggested that MeHg could have a significant effect on the planktonic larva stage survival and stage duration (Table 3). Low exposure MeHg multipliers for rate of travel and probabilities of escape resulted in an 86% reduction in predicted

(Fig. 4). On average, unexposed larvae responded at a speed of 58.63 ± 4.02 mm s−1 , while MeHg-treated larvae responded at a speed of 43.98 ± 4.43 mm s−1 (±S.E.). Response distance to the vibratory stimulus was not significantly affected by MeHg concentrations in the eggs (P = 0.881).

Table 2 Mean transformed rate of travel and visual reactive distance values (±1S.E.) used in the regression-tree analysis Rate of travel (mm s−1 )

Visual reactive distance (mm)

Sample size

Probability of escape (% of control)

Control dose Yolk Oil Oil + 4 Oil + 11

0.76 1.17 1.39 0.69

± ± ± ±

0.09 0.14 0.24 0.09

249.07 164.01 304.22 229.45

± ± ± ±

21.27 22.37 20.94 20.65

82 50 37 52

0.56 0.54 0.55 0.55

± ± ± ±

0.02 0.03 0.03 0.03

Low dose Yolk Oil Oil + 4 Oil + 11

0.58 0.58 0.52 0.50

± ± ± ±

0.09 0.15 0.15 0.1

167.77 239.61 258.40 269.26

± ± ± ±

19.89 29.04 25.15 31.66

63 29 24 32

0.50 0.45 0.43 0.58

± ± ± ±

0.02 (−11%) 0.04 (−17%) 0.04 (−22%) 0.03 (+5%)

High dose Yolk Oil Oil + 4 Oil + 11

0.30 0.80 0.57 0.40

± ± ± ±

0.05 0.14 0.11 0.06

236.72 268.41 238.61 216.52

± ± ± ±

26.91 32.39 29.67 21.64

38 34 39 48

0.52 0.50 0.48 0.49

± ± ± ±

0.03 (−7%) 0.04 (−7%) 0.03 (−13%) 0.03 (−11%)

Probabilities of escaping a predatory attack were calculated in control, low and high MeHg-exposed larvae at each of the four developmental stages studied, using a regression tree developed for red drum (Fuiman et al., 2006; Murphy, 2006; Murphy et al., in prep.). Values in brackets represent the percent difference from the control value.

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337 Table 3 Calculated growth and mortality rates, percent survival through the four developmental stages to the end of the planktonic larval stage (11 mm), and planktonic stage duration obtained by the IBM Treatment

Growth (mm day−1 )

Mortality (day−1 )

Survival (%)

Stage duration (days)

Control Low High

0.31 ± 0.00 0.28 ± 0.01 0.26 ± 0.01

0.15 ± 0.00 0.19 ± 0.00 0.21 ± 0.01

1.02 ± 0.04 0.19 ± 0.01 0.07 ± 0.02

29.8 ± 0.48 32.7 ± 0.63 34.1 ± 1.23

Values represent the average (±1S.E.) of three simulations each for control, low and high exposure groups.

stage survival, but did not affect predicted stage duration. Not surprisingly, high exposure to MeHg showed a stronger effect than the low exposure; high MeHg exposure reduced survival by 93%. In this case an increase of 26% in stage duration was also observed (Table 3). 4. Discussion Mercury contamination in aquatic systems is ubiquitous, and Texas coastal waters are no exception. For example, Sager (2004) reported total mercury levels for three important sport fishes from five minimally impacted bays in Texas: southern flounder (Paralichthys lethostigma), spotted seatrout (Cynoscion nebulosus), and red drum (Sciaenops ocellatus). Muscle levels of total mercury ranged between 50 and 250 ng g−1 . In another study, Edwards et al. (1999) reported total mercury levels in eel (Anguilla anguilla) muscle tissue ranging from 15 to 501 ng g−1 , and muscle tissue ranging from 19 to 121 ng g−1 in roach (Rutilus rutilus), depending on the river of origin. At least 81% of this total mercury in the muscle was in the form of MeHg (Edwards et al., 1999). Latif et al. (2001) estimated maternal transfer of MeHg to eggs to be 2% and 11% of the total mercury in muscle of walleye (Stizostedion vitreum) for uncontaminated and contaminated lakes, respectively. Conservatively assuming a 2% transfer, eggs produced by the fishes in Texas bays (Sager, 2004) could have MeHg loads between 1 and 5 ng g−1 , while fish studied by Edwards et al. (1999) could produce eggs with 0.3–2.42 ng g−1 . In the current study, Atlantic croaker fed MeHg-contaminated food at levels encountered by benthic feeders in the wild (Locarnini and Presley, 1996), produced eggs with MeHg concentrations of 0.04–4.6 ng g−1 . The levels in this study are, therefore, environmentally realistic and comparable to levels expected in wild fish based on Sager (2004) and Edwards et al. (1999). However, in the present study, the control diet was not uncontaminated marlin muscle, which would have been the most suitable. Therefore, the contribution of differences in the experimental diet separate from MeHg content of Atlantic croaker broodstock to the observed behavioral effects between control and experimental larvae cannot be determined at this time. The developmental stages chosen for this study were designed to help determine the role of mobilization of the yolk and oil globule in producing sublethal effects and whether any observed effects were physiological (temporary) or developmental (permanent) in nature. None of the effects observed

335

varied significantly with developmental stage, nor was there a statistically significant interaction term. All observed effects of MeHg exposure were associated with MeHg concentration in the eggs, suggesting that MeHg effects on behavior are developmental. Behavioral effects of maternally derived MeHg exposure were variable. Impairment of routine behavior occurred as a concentration-dependent decrease in the rate of travel and activity (Figs. 1 and 2). Startle response traits also exhibited a concentration-dependent effect, in which the response to a vibratory stimulus was longer and slower with increasing MeHg concentration in the eggs (Figs. 3 and 4). The behavioral skills analyzed here are presumed to be ecologically relevant and related to larval survival. A decrease in the rate of travel (Fig. 1) and activity (Fig. 2) suggest an impairment of foraging performance. In fact, Fuiman et al. (2006) showed that reduced rate of travel is associated with reduced success in escaping from a predator. A slower response to a vibratory stimulus (Fig. 3) suggests that larvae will be more vulnerable to a predatory attack. The likely impact of maternal exposure to MeHg in croaker larvae on the planktonic stage was evaluated using published and purpose-developed statistical models. RT analyses were performed using a subset of the behavioral traits we analyzed (specifically, the rate of travel and the visual reactive distance) that were shown to be predictive of escape probabilities for red drum larvae (Fuiman et al., 2006). These RT analyses suggest that MeHg reduces the probability of surviving a predatory attack in three of the four developmental stages (Table 2). The effects of maternally transferred MeHg on the rate of travel and the visual reactive distance predicted higher mortality rates (Z) compared to control larvae. Larvae from eggs in the low exposure group exhibited a 27% higher predicted mortality rate than the control group, and larvae from the high dose group had a 40% higher predicted mortality rate (Table 3). Growth rates of the low and high exposure groups derived from the simulations were also reduced compared to the control group by 10% and 16%, respectively (Table 3), because of the behavioral effects of MeHg on prey encounter rates. The combination of a higher mortality rate at this planktonic stage (2.5–11 mm) and a slower growth rate (longer stage duration) resulted in a mean reduction in larval survival to the end of the planktonic larval stage of 81% and 93% for low and high exposure groups, respectively, compared to control larvae (Table 3). However, the limitations of these types of simulation approaches must be considered. Although the RT analysis and the individual-based model used here are comprehensive methods, the models are essentially an abstraction and conclusions have to be interpreted cautiously. Therefore, maternal exposure to MeHg is expected to significantly impact Atlantic croaker populations in the wild but perhaps not exactly an 86–93% reduction in survival of a larval cohort. Maternally derived MeHg induced no direct effects on growth rates of croaker larvae in the laboratory. However, simulated growth rates (i.e., model derived) were affected by MeHg level and were faster than the rates measured in the laboratory. This discrepancy arises from model assumptions for natural prey densities (artificially elevated in the laboratory), observed stage

336

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337

durations for field-collected larvae, and foraging ability of larvae in the field estimated from observed behavioral traits. Studies on other marine fish larvae suggest that laboratory-reared larvae grow much slower than those in the wild because of natural food sources and size-selective predation (Strelcheck et al., 2003). Therefore, although this study failed to find an effect of MeHg on growth in the laboratory, behavioral effects that relate to foraging are expected to reduce growth rates in the field. This would prolong the larval stage duration during a period when mortality is high, thereby reducing survival to the end of the planktonic stage. Top pelagic predators like the blue marlin are known to accumulate considerable levels of MeHg (Downs et al., 1998). In this study, the low dose Atlantic croaker broodstock were fed a diet of MeHg-contaminated blue marlin meat alone and they transferred measurable amounts of MeHg to their eggs. This suggests that blue marlin transfer significant amounts of MeHg to their own offspring. Although we studied Atlantic croaker, our findings imply that larvae of blue marlin and other fishes that are high trophic level predators may also suffer from exposure to MeHg through maternal transfer which would result in reduced recruitment. Further studies on this possible mechanism of larval mortality for blue marlin larvae and its relation to the reported reduction of blue marlin recruitment to the adult population (Cox et al., 2002) are needed. 5. Conclusion This study demonstrated MeHg transfer from food to eggs of Atlantic croaker. Levels in spawned eggs were comparable to MeHg levels expected in the wild, and this maternally derived MeHg induced adverse effects on behavior that are expected to translate into reduced survival of planktonic larvae, with negative consequences for the population. Acknowledgments We thank Drs. G. Joan Holt, B. Scott Nunez, Peter Thomas and Rafael P´erez-Dom´ınguez for their helpful suggestions and critical review of this manuscript. We are also thankful to Dr. Larissa Ford for preparing the fish diets and analyzing MeHg concentrations in the eggs, and Mrs. Susan Lawson for caring for the broodstock. Mr. Hunter Samberson helped with fish husbandry. This study was supported by a grant from the U.S. Environmental Protection Agency, project number R-827399010. Contribution number OOOO of The University of Texas at Austin Marine Science Institute. References Ballatori, N., 2002. Transport of toxic metals by molecular mimicry. Environ. Health Perspect. 110 (Suppl. 5), 689–694. Benoit, J.M., Gilmour, C.C., Mason, R.P., Riedel, G.S., Riedel, G.F., 1998. Behavior of mercury in the Patuxent River estuary. Biogeochemistry 40, 249–265. Colborn, T., vom Saal, F.S., Soto, A.M., 1993. Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environ Health Perspect 101 (Suppl. 5), 378–384.

Cowan Jr., J.H., Houde, E.D., Rose, K.A., 1996. Size-dependent vulnerability of marine fish larvae to predation: an individual-based numerical experiment. ICES J. Mar. Sci. 53, 23–37. Cox, S.P., Martell, S.J.D., Walters, C.J., Essington, T.E., Kitchell, J.F., Boggs, C., Kaplan, I., 2002. Reconstructing ecosystem dynamics in the central Pacific Ocean, 1952–1998. I. Estimating population biomass and recruitment of tunas and billfishes. Can. J. Fish. Aquat. Sci. 59, 1724–1735. Diamond, S.L., Crowder, L.B., Cowell, L.G., 1999. Catch and bycatch: the qualitative effects of fisheries on population vital rates of Atlantic croaker. Trans. Am. Fish. Soc. 128, 1085–1105. Downs, S.G., Macleod, C.L., Lester, J.N., 1998. Mercury in precipitation and its relation to bioaccumulation in fish: a literature review. Water Air Soil Pollut. 108, 149–187. Dusci, L.J., Hackett, L.P., 1976. Rapid digestion and flameless atomic absorption, spectroscopy of mercury in fish. J. Assoc. Off. Anal. Chem. 59, 1183–1185. Edwards, S.C., MacLeod, C.L., Lester, J.N., 1999. Mercury contamination of the eel (Anguilla anguilla) and roach (Rutilus rutilus) in East Anglia UK. Environ. Monit. Assess. 55, 371–387. Faulk, C.K., Fuiman, L.A., Thomas, P., 1999. Parental exposure to ortho, para-dichlorodiphenyltrichloroethane impairs survival skills of Atlantic croaker (Micropogonias undulatus) larvae. Environ. Toxicol. Chem. 18, 254–262. Fuiman, L.A., Cowan Jr., J.H., 2003. Behavior and recruitment success in fish larvae: repeatability and covariation of survival skills. Ecology 84, 53–67. Fuiman, L.A., Magurran, A.E., 1994. Development of predator defenses in fishes. Rev. Fish Biol. Fisher. 4, 145–183. Fuiman, L.A., Ottey, D.R., 1993. Temperature effects on spontaneous behavior of larval juvenile red drum, Sciaenops ocellatus, and implications for foraging. Fish Bull. U.S. 91, 23–35. Fuiman, L.A., Rose, K.A., Cowan Jr., J.H., Smith, E.P., 2006. Survival skills required for predator evasion by fish larvae and their relationship to laboratory measures of performance. Anim. Behav. 71, 1389–1399. Fyhn, H.J., 1989. 1st feeding of marine fish larvae: are free amino acids the source of energy? Aquaculture 80, 111–120. Hall, B.D., Bodaly, R.A., Fudge, R.J.P., Rudd, J.W.M., Rosenberg, D.M., 1997. Food as the dominant pathway of methylmercury uptake by fish. Water Air Soil Pollut. 100, 13–24. Hammerschmidt, C.R., Sandheinrich, M.B., 2005. Maternal diet during oogenesis is the major source of methylmercury in fish embryos. Environ. Sci. Technol. 39, 3580–3584. Hanson, P.J., Zdanowicz, V.S., 1999. Elemental composition of otoliths from Atlantic croaker along an estuarine pollution gradient. J. Fish. Biol. 54, 565–668. Latif, M.A., Bodaly, R.A., Johnston, T.A., Fudge, R.J.P., 2001. Effects of environmental and maternally derived methylmercury on the embryonic and larval stages of walleye (Stizostedion vitreum). Environ. Pollut. 111, 139–148. Lawrence, A.L., Mcaloon, K.M., Mason, R.P., Mayer, L.M., 1999. Intestinal solubilization of particle-associated organic and inorganic mercury as a measure of bioavailability to benthic invertebrates. Environ. Sci. Technol. 33, 1871–1876. Letcher, B.H., Rice, J.A., Crowder, L.B., Rose, K.A., 1996. Variability in survival of larval fish: disentangling components with a generalized individual-based model. Can. J. Fish. Aquat. Sci. 53, 787–801. Locarnini, S.J.P., Presley, B.J., 1996. Mercury concentrations in benthic organisms from a contaminated estuary. Mar. Environ. Res. 41, 225–239. Mason, R.P., Reinfelder, J.R., Morel, F.M.M., 1995. Bioaccumulation of mercury and methylmercury. Water Air Soil Pollut. 80, 915–921. McCarthy, I.D., Fuiman, L.A., Alvarez, M.C., 2003. Aroclor 1254 affects growth and survival skills of Atlantic croaker Micropogonias undulatus larvae. Mar. Ecol. Prog. Ser. 252, 295–301. Miller, M.J., Nemerson, D.M., Able, K.W., 2003. Seasonal distribution, abundance, and growth of young-of-the-year Atlantic croaker (Micropogonias undulatus) in Delaware Bay and adjacent marshes. Fish Bull. U.S. 101, 100–115. Moore, J.C., 2000. A Review of Mercury in the Environment (its Occurrence in Marine Fish). South Carolina Department of Natural Resources, Marine Resources Division, Office of Environmental Management, p. 15.

M.d.C. Alvarez et al. / Aquatic Toxicology 80 (2006) 329–337 Murphy, C.A., 2006. Modeling the Effects of Endocrine Disrupting Chemicals on Atlantic Croaker: understanding Biomarkers and Predicting Population Responses. Doctoral Dissertation. Louisiana State University, Coastal Fisheries Institute and Department of Oceanography and Coastal Sciences, Baton Rouge, Louisiana. Murphy, C.A., Rose, K.A., Alvarez, M.C., Fuiman, L.A. Modeling larval fish behavior: scaling the sublethal effects of methylmercury to population relevant endpoints (in prep.). Myers, G.J., Davidson, P.W., Cox, C., Shamlaye, C., Cernichiari, E., Clarkson, T.W., 2000. Twenty-seven years studying the human neurotoxicity of methylmercury exposure. Environ. Res. 83, 275–285. Nixon, S.W., Jones, C.M., 1997. Age and growth of larval and juvenile Atlantic croaker Micropogonias undulatus, from the Middle Atlantic Bight and estuarine waters of Virginia. Fish Bull. U.S. 95, 773–784. Randall, D.J., Burggren, W., French, K., 1997. Eckert Animal Physiology: Mechanisms and Adaptations, 4th ed. W.H. Freeman and Co., New York, NY, USA. Rice, D.C., 1995. Neurotoxicity of lead, methylmercury, and PCBs in relation to the Great Lakes. Human health and environmental pollution in the Great Lakes. Environ. Health Perspect. 103, 71–87. Sager, D.R., 2004. Mercury in tissues of selected estuarine fishes from minimally impacted bays of coastal Texas. B Environ. Contam. Tox. 72, 149–156.

337

Simmons-Willis, T.A., Koh, A.S., Clarkson, T.W., Ballatori, N., 2002. Transport of a neurotoxicant by molecular mimicry: the methylmercury-l-cysteine complex is a substrate for human L-type large neutral amino acid transporter (LAT) 1 and LAT2. Biochem. J. 367, 239–246. Spry, D.J., Weiner, J., 1991. Metal bioavailability and toxicity to fish at low alkalinity lakes: a critical review. Environ. Pollut. 71, 243– 304. Strelcheck, A.J., Fitzhugh, G.R., Coleman, F.C., Koenig, C.C., 2003. Otolith-fish size relationship in juvenile gag (Mycteroperca microlepsis) of the eastern Gulf of Mexico: a comparison of growth rates between laboratory and field populations. Fish. Res. 60, 225–265. Watras, C.J., Bloom, N.S., 1992. Mercury and methylmercury in individual zooplankton: implications for bioaccumulation. Limnol. Ocean 37, 1313– 1318. Zar, J.H., 1999. Biostatistical Analysis, 4th ed. Prentice Hall, Upper Saddle River, New Jersey, p. 662. Zhou, T., Rademacher, D.J., Steinpreis, R.E., Weis, J.S., 1999. Neurotransmitter levels in two populations of larval Fundulus heteroclitus after methylmercury exposure. Comp. Biochem. Phys. C 124, 287–294. Zohar, Y., 1989. Fish reproduction: its physiology and artificial manipulation. In: Shilo, M., Sarig, S. (Eds.), Fish Culture in Warm Water Systems: Problems and Trends. CRC Press, Boca Raton, FL, USA, pp. 65–119.