Comparative Breeding and Behavioral Responses ... - ACS Publications

Oct 3, 2012 - ... and the effects of exposure to the environmental estrogen ethinylestradiol (EE2) on reproduction in a laboratory (Wild Indian Karyot...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Comparative Breeding and Behavioral Responses to Ethinylestradiol Exposure in Wild and Laboratory Maintained Zebrafish (Danio rerio) Populations Marta Söffker, Jamie R. Stevens, and Charles R. Tyler* University of Exeter, College of Life and Environmental Sciences, School of Biosciences, Geoffrey Pope Building, Stocker Road, Exeter EX4 4QD United Kingdom S Supporting Information *

ABSTRACT: Genetic variation has a significant effect on behavior, fitness, and response to toxicants; however, this is rarely considered in ecotoxicological studies. We compared fitnessrelated behavioral traits, breeding activity, and the effects of exposure to the environmental estrogen ethinylestradiol (EE2) on reproduction in a laboratory (Wild Indian Karyotype, WIK) strain and a wild-caught population (Bangladesh, BLD01) of Danio rerio (zebrafish). In WIK fish, males with higher observed heterozygocity were more active reproductively and more successful in securing parentage, but these relationships were not apparent in the BLD01 fish. The frequency of reproductive behaviors increased in WIK zebrafish for exposure to 0.4 ng/L EE2, which was not apparent in the BLD01 zebrafish. The different strains showed the same threshold for hepatic vitellogenin gene (vtg) induction (2.2 ng EE2/L), but results suggested an elevated response level in the BLD01. There were no effects on total egg production up to 2.2 ng EE2/L in either population; however, there was reduced egg fertilization rate at 2.2 ng EE2/L in the BLD01 fish. These results show consistency in the general responses to EE2 between these two genetically divergent strains of zebrafish, but also illustrate differences in their breeding biology and response sensitivities. These findings highlight the need for due consideration of the source (and genetics) of populations used in ecological risk assessment for accurate comparisons among studies.



HO = 0.7110). Furthermore, the natural genetic variation occurring between wild zebrafish populations was not present between laboratory strains of zebrafish.11 There is evidence for an impact on fish of reduced genetic diversity on growth, survival, reproductive success, and immune response.12,13 Behavioral responses between laboratory (inbred) and wild (outbred) strains of zebrafish can also differ, including in boldness, shoaling, startling response, and anxiety.14−17 In salmonid fish, genetic diversity has also been shown to affect social dominance status and behavior, where dominance is linked with higher individual heterozygocity.18 In a recent study, an outbred population of zebrafish (with higher genetic variation) was found to respond differently to clotrimazole (used in antifungal medication) compared with a more inbred population, the latter of which were more susceptible to its masculinizing effects.19 Much of the work in (eco)toxicological research, however, has not taken into account possible variations in response between strains, e.g. between inbred laboratory strains and outbred wild populations.10 Populations with higher genetic diversity tend to show greater tolerance and phenotypic variability in response to toxicants (reviewed in 20) and given this, it is crucial to

INTRODUCTION The zebrafish is a well established model organism for use in research on development,1 biomedicine,2 stress,3 behavioral ecology,4 and ecotoxicology.5,6 Maintenance of genetic diversity in laboratory maintained animals is generally accepted as being important in chemical risk assessment for the protection of wild populations, but there has been a lack of any standardization in the strains of animals used. There have been very few direct comparisons of the responses to chemicals between strains of zebrafish but in laboratory strains of rats response sensitivity to dioxins has been shown to vary by as much as 2000-fold.7 Some laboratory zebrafish strains are referred to as “wild type” and their genetic diversity is assumed to be heterogeneous and outbred,8,9 but this assumption is rarely tested. There are currently 26 different zebrafish strains registered as “wild type” with zfin.org (zebrafish database), of which three lines only (Darjeeling, India, Nadia) are established from known sources of wild-caught zebrafish. The most recent of these lines of zebrafish (Nadia) was established in 1999. Zebrafish lines most frequently used in ecotoxicology studies reported in the peerreviewed scientific literature include AB, Tübingen Long (TL), Tübingen (TU), and Wild Indian Karyotype (WIK), and only TU and its derivative WIK are descendent from known sources of wild-caught fish. Analysis of the genetic variation in some of these strains has identified both reduced allelic richness (range AR = 1.97−5.48) and reduced observed heterozygocity (HO = 0.4−0.6) in comparison to wild-caught zebrafish (AR = 14.13, © 2012 American Chemical Society

Received: Revised: Accepted: Published: 11377

June 18, 2012 September 18, 2012 September 21, 2012 October 3, 2012 dx.doi.org/10.1021/es302416w | Environ. Sci. Technol. 2012, 46, 11377−11383

Environmental Science & Technology

Article

identification of individuals, and left for 4 days in experimental tanks to allow the establishment of hierarchies. Following this, EE2 exposure was initiated and was run for 14 days. There were nine randomly assigned tanks for each of the treatment groups (water control and EE2 nominally at 2 and 5 ng/L) for each of the two zebrafish strains. Behavioral interactions were assessed between individual fish as described below. All fish were terminated at the end of exposure and measurements were taken on body size (weight and length) and other physiological end points, as described below. The experimental setup was chosen to accentuate male−male interactions and maximize egg output of the shoal (two males and one female) and builds on previous work we have conducted with this matrix.29 The experimental setup was also the most tractable system to identify effects of EDCs on competition between males (mating and paternal success). Although zebrafish are shoaling and group-spawning fish, the actual process of spawning usually includes only two or three fish within the shoal (32 and our own personal observations). Furthermore, this shoal size falls within their natural shoal size range.33 In the wild, zebrafish breed annually and thus breeding fish within a population tend to be of a similar size,34,35 as we have adopted here in this study. This experimental system comprised of 54 tanks, and to allow for the required behavior observations it was designed as a fully crossed, blocked approach with three replicates of each exposure per block. The three blocks were started 3 weeks apart to give 9 replicates of each exposure. The factor “block” was included in subsequent statistical analyses where appropriate, as a random variable with 3 levels. Chemical Dosing and Water Chemistry. Fish were maintained and exposed in tap water filtered by reverse osmosis (RO; Osmonics E625 with cellulose membranes; GE Water and Process Technologies, Trevose, PA, USA) then reconstituted with Analar grade mineral salts to concur with U.S. Environmental Protection Agency (EPA) guidelines on standardized synthetic freshwater. The chemical dosing was run using a solvent-free method. Six mg of EE2/L acetone (100%) was added into 2.5-L dosing bottles and left overnight for the acetone to evaporate. Two L of reverse osmosis water was then added and the solutions were shaken vigorously. EE2 was pumped into the tanks continuously at the required rate for the exposures, and dilution water was maintained at a flow rate of 1 L/h. To determine EE2 exposure concentrations, 1-L water samples were collected on days 2, 9, and 14 from all tanks, preserved with 3 mL of 30% hydrogen chloride and 0.25 g of copper nitrate, and analyzed for EE2 via liquid chromatography−mass spectrometry (by Severn Trent Laboratories, Coventry, U.K.). Genetic Diversity. DNA was extracted from the caudal fin of each fish using the ammonium acetate precipitation method.36 Six microsatellite loci were used to assess genetic diversity (Z226, Z374, Z1213, Z1233, Z13614, and Z5058). Reaction mixtures, microsatellite primers, and PCR profiles were as described previously by Coe et al.37 We multiplexed the chosen loci for PCR amplification. Genetic diversity was evaluated by means of observed and expected heterozygocity,38 number of alleles, and allele richness. Allele richness was calculated using FSTAT 1.2.39 Reproductive Output. The number of eggs produced by each colony was determined daily. Eggs/embryos were collected 1.5 h after onset of artificial sunrise, using a system designed to minimize disturbance to the fish.40 The eggs were

consider genetic variation where effects derived from exposures using laboratory-based animals are applied in the protection of wild populations. The synthetic estrogen 17α-ethinylestradiol (EE2), used in oral contraceptives, is an environmental pollutant of considerable concern. Ethinylestradiol is found widely in wastewater treatment works (WwTWs) at concentrations up to 10 ng/L (but more commonly at between 0.5 and 3 ng/L) and in receiving rivers at concentrations up to 1 ng/L.21−24 It is an exquisitely potent estrogen and has been shown to induce disruption of sexual development,25 sex reversal, and even population failure.9,26 Studies on zebrafish (much of it on the Wild Indian Karyotype [WIK] strain) have shown exposure to EE2 can inhibit male gonad development,27 reduce fertilization success,5 and induce changes in aggressive and sexual behaviors.28−30 To date, however, very little work has been done that has considered how responses to EDC exposure in fish are affected by their genetics. In this study we first compared breeding and behaviors in wild-caught zebrafish (derived from Bangladesh, BLD01) with a laboratory maintained strain (WIK). We then assessed their responses to EE2 at environmentally relevant exposure concentrations by measuring effects on reproductive output and behavior in small breeding colonies.



MATERIALS AND METHODS Test Fish. We collected the Bangladeshi strain of zebrafish (BLD01) from a household pond in Mozahdi village in Tarakanda, Mymensingh Upazila, Bangladesh in February 2010 and transported them to the University of Exeter, where they were acclimated for 6 weeks in an aquarium. Test fish for the WIK strain were bred from an existing stock obtained originally from the Max-Planck-Institute in Tübingen, Germany. Both strains were approximately 9 months in age. All experiments and animal procedures were carried out in accordance with the U.K. Home Office Animals (Scientific Procedures) Act 1986. Experimental Procedure. Twenty-seven groups of both the WIK and BLD01 zebrafish, each comprising two male and one female fish were placed into 15-L experimental tanks (Figure 1). Fish were size matched by wet body mass to ensure equal opportunity for dominance (as dominance is affected by body size31), tagged with Visible Implant Elastomers for

Figure 1. Experimental design for assessing effects of exposure to EE2 on reproductive behavior and output, using temporal block design. Wild BLD01 and laboratory-bred WIK zebrafish were exposed to two concentrations of EE2 (0.4 and 2.2 ng/L) for 14 days. 11378

dx.doi.org/10.1021/es302416w | Environ. Sci. Technol. 2012, 46, 11377−11383

Environmental Science & Technology

Article

washed, counted, and incubated at 26 (±1) °C in small plastic containers to assess egg fertilization rate, embryo viability, and to increase DNA quantity for parentage analysis. Eggs were observed under the microscope just after fertilization, 1 h after collection (approximately at 32- to 64-cell stage of development), and unfertilized or fungally infected eggs were removed. Viability of the fertilized eggs was assessed at 8 and 26 hpf, and viable, fertilized embryos at 26 hpf were then preserved in 100% ethanol for subsequent parentage analysis. Induction of Vitellogenin (vtg) mRNA. Response to EE2 exposure was assessed by determining induction of hepatic expression of vitellogenin (vtg) mRNA in male fish. In females, vitellogenin is produced in the liver in response to endogenous estrogen but males carry the gene and express it readily when exposed to exogenous estrogens.41,42 Total RNA was extracted from livers using TRI reagent (Sigma Aldrich) according to the manufacturer’s instructions. It was then DNase treated, and vtg mRNA expression was determined by quantitative real-time PCR as we have described previously.43,44 cDNA synthesis, primer design, and PCR profiles were conducted as described previously43,44 and the primer pairs used are detailed in Lange et al.44 Expression of efficiency-corrected hepatic vtg was determined by relative expression normalized against the ribosomal protein L8 (rpl8), measured in each sample, which is not affected by estrogen exposure.45 Behavioral Analysis. Reproductive behaviors in the different strains of zebrafish and effects of EE2 exposure on sexual and aggression behaviors were assessed at 5 time-points during the course of the study. After the acclimation period (4 days after placing the fish into their respective tanks) behaviors were assessed every 3 days. At each time-point observations were made on fish in the different treatment tanks in random order for a period of 4 min each, scoring interactions among all individuals. Aggressive behavior was scored as chasing, snapping, biting, and circling/fighting (agonistic behavior, see 29 and Table S1 in the Supporting Information). Sexual behavior was defined as the sum of all pursuits, displays, and attempted and successful spawnings (see Table S2), with pursuits scored as a multiple of 3 s and the remaining behaviors based on frequencies alone. Spatial position of fish in the tank during sexual displays was also recorded and categorized as being in the upper, middle, or lower third of the tank water column. Parentage Analysis. For parentage analysis 15 randomly selected fertilized eggs (this number was derived following a power analysis to determine adequate sampling size, see 46) were sampled from each control tank over the duration of the study in direct proportion to the overall number of eggs laid each day. DNA was extracted from fins and eggs using the ammonium acetate precipitation method (as above), and the same six microsatellite loci used for the genetic diversity measurements (described above) were used for parentage assignment. Eggs that could not be assigned were excluded from the analysis; in the WIK strain 11.8% of eggs could not be assigned and in the BLD01 strain 2.4% of eggs could not be assigned. Statistical Analysis. All data were analyzed with the statistical software R v. 2.13.0 or SPSS v. 16.0. Behavioral patterns were tested using generalized linear models adjusted to overdispersion and accounting for time and repeat (fixed factors), or linear mixed effects models accounting for repeated observations with tank nested in observation block as random factor, where appropriate. Relationships among heterozygocity,

social status, behaviors, and parentage were tested either through generalized linear models (poisson errors) or logistic binomial regressions (binomial errors).



RESULTS Breeding in WIK vs BLD01 Zebrafish. Microsatellite analysis indicated that the WIK strain was more inbred than the BLD01 strain (F = 0.36 vs F = 0.23), having a lower level of heterozygocity and lower allelic diversity (Table 1). Overall,

Table 1. Microsatellite Diversity Indices for Wild (BLD01) and Laboratory-Bred (WIK) Strains of Zebrafisha locus z1213 BLD01 N 32 AN 11 AR 18.50 HE 0.77 HO 0.45 WIK N 27 AN 9 AR 12.96 HE 0.81 HO 0.37

z1233

z13614

z266

z5058

z374

average value

32 9 9.50 0.82 0.36

32 22 20.24 0.55 0.91

32 6 11.44 0.83 0.33

32 9 9.54 0.64 0.73

32 13 18.53 0.73 0.55

32 11.67 14.63 0.72 0.56

27 7 7.92 0.76 0.48

27 8 10.84 0.69 0.63

27 9 9.70 0.76 0.48

27 7 9.00 0.89 0.22

27 12 11.40 0.63 0.74

27 8.67 10.56 0.76 0.49

a

N = number of individuals; AN = number of alleles; AR = allelic richness; HE = expected heterozygocity; HO = observed heterozygocity.

there appeared to be a lower total egg production (cumulative egg numbers over 14 days) in the WIK compared with BLD01 zebrafish, but this was not statistically significant (F3,38 = 1.106, p = 0.09). In WIK fish, the most aggressive, dominant fish (hierarchy dominance) also displayed the highest level of sexual behavior (binomial logistic regression, p = 0.001), but this was not the case in BLD01 groups (p = 0.08). In WIK control groups, males with higher observed heterozygocity were more reproductively active (F1,9 = 7.111, p = 0.03), and were, in general, more successful in securing parentage (F1,9 = 4.85, p = 0.055). In BLD01 control groups, for males, there was no such relationship between heterozygocity and dominance/sexual behavior (p = 0.39), or with parentage success (p = 0.78). The parentally dominant male sired 91.9% of the offspring in WIK colonies, and 78.1% of the offspring in BLD01 colonies, and there was no difference statistically between the strains (Welch two-sample t test, p = 0.14). Effects of EE2 on Reproductive Output. The measured EE2 water exposure concentrations were 0.4 ± 0.14 ng/L (lowdosed tanks) and 2.2 ± 1.14 ng/L (high-dosed tanks). Hepatic expression of vtg mRNA in males in both BLD01 and WIK zebrafish was low, as expected. In WIK fish males hepatic vtg expression was significantly higher in dominants compared with in subordinates (two-sample t test, p = 0.03), but no such relationship was found in the Bangladesh strain. There was no induced hepatic expression of vtg in males in either the BLD01 or WIK zebrafish exposed to 0.4 ng EE2/L. In contrast, exposure to 2.2 ng EE2/L induced hepatic vtg expression in males of both strains (F2,35 = 5920, p = 0.0008 for BLD01, F2,22 = 3999, p = 0.0005 for WIK, Figure 2) compared with both the 11379

dx.doi.org/10.1021/es302416w | Environ. Sci. Technol. 2012, 46, 11377−11383

Environmental Science & Technology

Article

Figure 2. Hepatic vtg mRNA expression in two strains of male zebrafish following exposure to EE2. The central line within the box is the median; the box represents the upper and lower quartiles, and the whiskers represent the 95% confidence intervals. Open circles are outliers beyond the 95% confidence intervals. Letters indicate significant differences in vtg expression.

control and the 0.4 ng EE2/L treatment. In BLD01 males exposed to 2.2 ng EE2/L vtg expression was 2.7 times higher than in WIK males (median relative expression was 113.5 fold and 40.9 fold, respectively), but this apparent difference was not significant due to large response variation in BLD01 males (two-sample t test, p = 0.12). There was no effect of exposure to EE2 on total egg production for either strain of zebrafish for the 10-day study period (F1,22 = 1.76, p = 0.2 for BLD01, F1,13 = 0.26, p = 0.62 for WIK; Figure 3). Exposure to EE2 did, however, appear to affect the pattern of egg production in the WIK colonies over the study period (F3,36 = 2.55, p = 0.03; Figure S1), with a trend for an increased egg output over time in colonies exposed to 2.2 ng EE2/L (p = 0.01) which was not seen in the BLD01 fish (F1, 22 = 1.50, p = 0.23). Exposure to 2.2 ng EE2/L resulted in a reduced egg fertility rate in BLD01 fish (F1,22 = 3783, p = 0.01; Figure 3b), but no such effect occurred for the WIK fish. There were no subsequent differences in the embryo viability rates (at 26 hpf) between the strains and the exposures. Effects of EE2 on Reproductive Behavior. Total aggression (aggressive acts displayed by the colonies as a whole) was significantly higher in WIK fish compared with the BLD01 fish (lme, p = 0.01), and this was independent of exposure (lme, p = 0.4). This was primarily due to the dominant male in the group, which was more aggressive in WIK fish compared with the BLD01 fish (lme, p = 0.009). Aggression was significantly higher in WIK fish exposed to 0.4 ng EE2/L compared with controls and fish exposed to 2.2 ng EE2/L (lme, p = 0.009). The dominant fish in WIK groups were more likely to “patrol” (Table S1) than the dominant fish in BLD01 groups (p = 0.08). Total sexual behavior (by colonies) was not significantly different between WIK and BLD01 fish, or between exposures (lme, p = 0.6). However, dominant WIK fish showed an increase in their display of sexual behavior over time when exposed to both 0.4 and 2.2 ng EE2/L (lme, p = 0.04, Figure 4), which was not the case for the BLD01 fish.

Figure 3. (A) Total number of eggs spawned in wild (BLD01) and laboratory-bred (WIK) zebrafish groups exposed to EE2. (B) Relative difference of unfertilized eggs following exposure to EE2 in wild (BLD01) and laboratory-bred (WIK) zebrafish. Asterisks indicate significant differences in fertilization rate. The total number of eggs did not vary between strains or exposures; however, proportion of unfertilized eggs was significantly higher in BLD01 groups exposed to 2.2 ng EE2/L.

The position of the fish in the water column during spawning and sexual behaviors was affected by EE2 exposure in both strains. Under unexposed conditions spawning took place toward the base of the tank close to the spawning substrate, but under EE2 exposure these behaviors were shifted toward the water surface (binomial logistic regression, p = 0.01) in both study populations.



DISCUSSION In the study presented we identify some comparative differences in breeding and behavior between wild-caught and laboratory-bred zebrafish and show that although the different strains of zebrafish showed consistency in their general responses to EE2 exposure, there were differences in response sensitivities. Breeding and Behavior in WIK and BLD01 Zebrafish Colonies. In this study we identify some of the basic behaviors in the establishment of dominance in zebrafish colonies and suggest that genetic variation and allelic richness may contribute to dominance status and response to toxicant exposure. The behavioral phenotypes associated with aggres11380

dx.doi.org/10.1021/es302416w | Environ. Sci. Technol. 2012, 46, 11377−11383

Environmental Science & Technology

Article

production where there were no observed effects of the estrogen treatments. Responses to EE2 exposure however differed for some of the other end points measured. In WIK fish exposure to a low level of EE2 (0.4 ng/L) induced greater sexual behavior displays in the dominant (and more heterozygous) males progressively over time. Egg output in the WIK group females was also found to increase over time for the higher (2.2 ng EE2/L) exposure. Previous studies have shown that low-dose exposure to EE2 (0.05 ng/L) can increase sperm motility in WIK males5 and heighten fertilization success (at concentrations