Environmental Pollution 157 (2009) 881–886 Contents lists available at ScienceDirect Environmental Pollution journal homepage: www.elsevier.com/locate/envpol Rapid genetic erosion in pollutant-exposed experimental chironomid populations Carsten Nowak a, *, Christian Vogt b, Markus Pfenninger a, Klaus Schwenk a, Jörg Oehlmann b, Bruno Streit a, Matthias Oetken b a Abteilung Ökologie und Evolution, Institut für Ökologie, Evolution und Diversität, J. W. Goethe-Universität Frankfurt am Main, Siesmayerstrasse 70, 60054 Frankfurt am Main, Germany Abteilung Aquatische Ökotoxikologie, Institut für Ökologie, Evolution und Diversität, J. W. Goethe-Universität Frankfurt am Main, Siesmayerstrasse 70, 60054 Frankfurt am Main, Germany b Chronic TBT exposure reduces allelic variation at ﬁve variable microsatellite loci in experimental populations of Chironomus riparius. a r t i c l e i n f o a b s t r a c t Article history: Received 14 May 2008 Received in revised form 4 November 2008 Accepted 5 November 2008 Few studies have evaluated how effectively environmental contamination may reduce genetic diversity of a population. Here, we chose a laboratory approach in order to test if tributyltin (TBT) exposure at environmentally relevant concentrations leads to reduced genetic variation in the midge Chironomus riparius. Two TBT-exposed and two unexposed experimental populations were reared simultaneously in the laboratory for 12 generations. We recorded several life-history traits in each generation and monitored genetic variation over time using ﬁve variable microsatellite markers. TBT-exposed strains showed increased larval mortality (treatments: 43.8%; controls: 27.8%), slightly reduced reproductive output, and delayed larval development. Reduction of genetic variation was strongest and only signiﬁcant in the TBTexposed strains (treatments: 45.9%, controls: 24.4% of initial heterozygosity) after 12 generations. Our ﬁndings document that chemical pollution may lead to a rapid decrease in genetic diversity, which has important implications for conservation strategies and ecological management in polluted environments. Ó 2008 Elsevier Ltd. All rights reserved. Keywords: Allelic variation Chemical pollution Evolutionary ecotoxicology Chironomidae Microsatellites 1. Introduction Numerous studies have documented negative effects of environmental pollutants on life-history traits of ecotoxicological model species in the laboratory (Walker et al., 2001). Although laboratory test systems provide a powerful tool for toxicity assessment, they mostly consider acute or single-generation effects on life-history responses towards chemical exposure (Vogt et al., 2007a). In contrast to single-generation experiments, natural populations are chronically exposed to pollutants over multiple generations and numerous factors inﬂuence the long-term response to chemical exposure in the ﬁeld. For instance, populations may adapt to polluted environments (Gillis et al., 2002; * Corresponding author. Present address: Department of Limnology and Conservation, Senckenberg Research Institute, Clamecystrasse 12, 63571 Gelnhausen, Germany. Tel.: þ49 6051 61954 3122; fax: þ49 6051 61954 3118. E-mail addresses: [email protected], [email protected] (C. Nowak), [email protected] bio.uni-frankfurt.de (C. Vogt), [email protected] (M. Pfenninger), [email protected] (K. Schwenk), [email protected] (J. Oehlmann), [email protected] (B. Streit), [email protected] (M. Oetken). 0269-7491/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2008.11.005 Vogt et al., 2007a). Furthermore, chemical exposure has frequently been predicted to reduce genetic variation in natural populations (Bickham et al., 2000; Van Straalen and Timmermans, 2002). This effect, named genetic erosion, has been reported in ﬁeld surveys for various animal taxa, like ﬁsh (Murdoch and Hebert, 1994), crustaceans (Krane et al., 1999) and marine gastropods (Kim et al., 2003). However, most of these studies have provided little evidence for erosion of genetic variation in response to pollution stress (Brown et al., 2007; Chen et al., 2003; Nadig et al., 1998). Thus, the question if genetic erosion threatens populations in polluted habitats remains unanswered (Van Straalen and Timmermans, 2002). Patterns of genetic variation within and among natural populations are inﬂuenced by various historical and recent factors, like past and present selection processes, migration and chance events (Avise, 2004). Most ﬁeld studies with regard to the effects of pollution on genetic variation, however, do not consider those natural processes adequately (Bickham et al., 2000; Whitehead et al., 2003), and consequently fail to distinguish natural ‘‘background noise’’ from anthropogenic impacts. In order to exclude factors like gene ﬂow among populations and to control ﬂuctuations in population size, we chose an experimental approach here. The experiment was designed to investigate if pollutants at environmentally relevant 882 C. Nowak et al. / Environmental Pollution 157 (2009) 881–886 concentrations reduce genetic variation in populations of an ecotoxicological model organism, the midge Chironomus riparius Meigen (Diptera: Chironomidae). C. riparius is widely distributed in small streams, ditches, ponds and puddles throughout the holarctic. In organically polluted and muddy habitats, the species may reach high densities and it plays a key role in freshwater ecosystems for nutrient cycling and energy ﬂux (Armitage et al., 1995). As detritus feeders living in the upper sediment layers, chironomids come into contact with sediment-associated toxicants like heavy metals, PCBs or organotin compounds (Vogt et al., 2007b). In this study, C. riparius larvae were exposed to sediments contaminated with the highly effective biocide tributyltin (TBT). We used a TBT concentration (measured time-weighted mean concentration ¼ 8.93 mg as Sn kg1 sediment dry weight ¼ dw), which is environmentally relevant (Wade et al., 2004) and has previously been shown to affect life-history traits of C. riparius (Vogt et al., 2007b). The exposure was maintained for 12 generations and genetic variation as well as life-history traits were monitored over time. In particular, we focused on two main questions: (i) Does TBT exposure at an environmentally relevant concentration cause a measurable reduction of genetic variation at neutral loci within only a few generations? (ii) Can we observe changes in life-history responses to TBT exposure over time? 2. Materials and methods 2.1. Experimental procedure and TBT exposure For the experiments we used a laboratory strain of C. riparius originating from a cross-breeding of 11 laboratory strains in 2004 (GENþ; described more in detail in Nowak et al., 2007a). Four experimental populations (¼groups) were separated from the GENþ source strain and reared in the laboratory for twelve consecutive generations. Two groups were kept under control conditions (solvent control ¼ SC), while larvae of the two remaining groups were chronically exposed to sediment-bound TBT as described below. To start the experiment, 42 egg masses were taken from the source strain, and hatched larvae were carefully mixed in a Petri dish. Three hundred and ﬁfty larvae were randomly chosen as start generation for each group. Larvae were transferred in seven 2 l glass vessels per treatment containing 1 l reconstituted water and 100 g artiﬁcial sediment (pure quartz sand; for details see Vogt et al., 2007b). Fifty larvae were dispersed into each replicate. For the TBT treatments, sediment was spiked with nominally 160 mg as Sn kg1 sediment dw (390 mg as TBT kg1 sediment dw) using pure ethanol as solvent. After larval development, emerged imagines were transferred into 10 l reproduction containers (30 20 20 cm) containing a square water-ﬁlled box (11.5 11.5 5.5 cm, 400 ml) for egg laying. Egg masses were collected from the reproduction container and placed into 24-microwell plates for hatching. After hatching, 350 L1 larvae were taken from the two days with the highest egg mass production and mixed in a vessel. Subsequently, larvae were randomly distributed over the replicates to start the next generation. The following life-history parameters were measured in each generation: larval mortality, mean emergence time (EmT50), produced egg masses per female and hatchability of egg masses. Furthermore, daily population growth rate (PGR; calculated according to Vogt et al., 2007b) was used as an integrating endpoint in the experiments. For the genetic analyses, 200 larvae of each experimental group were taken from the same clutches as those used for the multi-generation study and reared in two separate 10 l aquaria. Rearing conditions were exactly the same in these aquaria compared to the test vessels, including the sediment contamination in the TBT exposed treatments. All resulting L4 larvae were frozen in liquid nitrogen and stored at 80 C for further analyses. The experiment was conducted at 20 C (0.5 C) under an 8:16 h dark/light cycle following the OECD guideline 218 for Chironomus sediment tests (OECD, 2004). For further details regarding the experimental procedure see (Vogt et al., 2007a). The experiments were stopped after the 12th generation. 2.2. Acute tests In order to detect alterations in TBT susceptibility of C. riparius larvae, acute tests were performed with L1 larvae of all four groups in generations 4, 8 and 12. Larvae were exposed to nominal TBT concentrations of 2, 6.3, 20, 63 and 200 mg as Sn l1 via water, including a solvent control (10 ml EtOH l1, equals 0.17 mM). After 24 h, survival was determined using a stereo microscope. 2.3. Chemical sediment analyses TBT analyses of the sediments were performed by GALAB Laboratories, Geesthacht, Germany, according to DIN guideline 19744 (2003). Approximately 2 g freeze-dried sediment were taken from test vessels of the ﬁrst experimental generation. The extracted organotin compounds were derivatized with 4% NaBEt4 in H2O, concentrated into cyclohexane by liquid–liquid extraction, and subsequently analyzed with gas chromatography tandem atomic emission detector GC-AED (GC 6890 Series, Agilent Technologies; AED: HP G2350A, Hewlett Packard, Waldbronn, Germany). The PACS-2 reference material (National Research Council Canada, Ottawa, ON) was analyzed to check extraction efﬁciency, showing recoveries of 95– 100%. The detection limit was 1 mg kg1 sediment dw. TBT and its metabolites were analyzed for controls and for TBT treatments after solvent evaporation 1 and 72 h, and 14 days of test duration. Based on the TBT loss after solvent evaporation and the measured TBT concentrations after 14 days the time-weighted average concentration was calculated according to OECD guideline 211 (OECD, 1998). For the acute tests, no TBT analyses were performed, and hence the calculated LC50 values are based on nominal concentrations. 2.4. Genetic analyses Genetic diversity was estimated using ﬁve variable microsatellite markers described in Nowak et al. (2006). Frozen L4 larvae were homogenized in 1.5 ml EppendorfÒ tubes containing 700 ml standard CTAB buffer and 4 ml 20 mM proteinase K. Tissue was digested in a water bath at 63 C for 1 h followed by standard chloroform/isoamylalcohol 24:1 treatment. DNA was precipitated in 1 ml isopropanol 100% for 1 h at -20 C and washed twice with 300 ml ethanol 70%. DNA was diluted in 600 ml water and stored at 4 C for further treatment. Polymerase chain reaction was performed on a Tetrad thermocycler (InvitrogenÒ) with cycling conditions as follows: 1 min 92 C, 1 min 55 C, 1 min 72 C, repeated 36 times. Reactions were performed in a total volume of 10 ml containing 0.1 ml of 25 mM dNTPs, 0.5 ml 50 mM MgCl2, 1 ml 10 reaction buffer, 0.2 ml of each primer (10 pmol) and 0.1 ml 5 U Taq DNA polymerase (Invitrogen). PCR products were loaded on a 1.4% agarose gel at 120 V and quantiﬁed after staining with ethidium bromide under UV light. For fragment analysis, 1 ml DNA solution containing w1 ng of microsatellite fragment was mixed with 5 ml dextran blue solution and 1 ml of internal size marker lambda DNA. The solution was loaded on a 9% polyacrylamide gel using an ALF sequencer (PharmaciaÒ). Gels were run for 450 min at 1000 V. Fragment lengths were scored manually using the ALFWIN 3.1 software. Genetic diversity was checked for 24 individuals per group and generation (¼96 individuals/generation). Every second generation was included for genetic analysis following the starting generation (¼generations 0, 2, 4, 6, 8, 10 and 12). 2.5. Statistical analysis Prior to statistical analysis, all endpoints were tested for normal distribution with the Kolmogorov–Smirnov normality test. Nested two-way ANOVA with repeated measurements was used in order to reveal the inﬂuence of TBT on variation in larval mortality and mean emergence time (EmT50) and to test for signiﬁcant differences in life-history response between groups and generations. We chose an ANOVA design in which experimental groups (SC I, SC II, TBT I, TBT II) were nested into treatments (SC and TBT). The analyses were performed using STATISTICA 7.1 (StatSoft Inc., USA). We tested for the presence of signiﬁcant linear trends in all lifehistory traits (mortality, EmT50, produced fertile egg ropes per female, population growth rate) and population genetic parameters (see below) with the Mann–Kendall test, and calculated Sen’s slope estimate with the Excel-application MAKESENS 1.0, provided by the Finnish Meteorological Institute. Population genetic parameters (expected and observed heterozygosity [HE and HO], number of alleles per locus [NA], number of polymorphic loci [NP] and test for deviations from Hardy–Weinberg equilibrium) based on the microsatellite data were calculated using GenAlEx 6 software (Peakall and Smouse, 2006). 3. Results 3.1. TBT concentrations in the sediment No TBT or other organotin compounds were detected in the control sediment. After solvent evaporation a TBT concentration of 21.7 mg as Sn kg1 sediment dw was determined in the TBT-spiked sediment, and after 1 h of test duration (containing animals, water and food) a TBT concentration of 9.8 mg as Sn kg1 sediment dw was measured. A similar TBT concentration was also measured after 72 h with 8.9 mg as Sn kg1 sediment dw. Prior to larval emergence C. Nowak et al. / Environmental Pollution 157 (2009) 881–886 (after 14 days of test duration) the TBT concentration decreased to 2.8 mg as Sn kg1 sediment dw. Two TBT metabolites, monobutyltin (MBT) and dibutyltin (DBT), were detected in low microgram per kilogram concentrations after 72 h (MBT ¼ 4.3 mg as Sn kg1 and DBT ¼ 1.6 mg as Sn kg1 sediment dw) and 14 days (MBT ¼ 8.1 mg as Sn kg1 and DBT ¼ 0.8 mg as Sn kg1 sediment dw). During the whole test the time weighted mean TBT concentration was 8.93 mg Sn kg1 sediment dw. 3.2. Life-history traits In the multi-generation study TBT had a signiﬁcant effect on both larval mortality and mean emergence time (EmT50; two way ANOVA, both p < 0.001; Table 1). Mortalities ranged from 14.2% (TBT II in generation 12) to 100% (TBT I and II in generation 10) in the TBT treatments and from 6.3% (SC I in generation 1) to 80.3% (SC I in generation 6) in the controls (Fig. 1A). Over all generations, mean mortality was higher in the TBT-treated groups than in the control groups (43.8% in the treatments compared to 27.8% in the controls). However, only in four generations both TBT-treated groups showed higher mortalities than the two respective controls (generations 1, 5, 9 and 10). Because of the extinction of both TBTexposed groups in generation 10, the study was maintained using backup groups that had been isolated from both TBT-exposed groups in the previous generation and being maintained under control conditions in generation 10. EmT50 values were higher in both TBT groups compared to the controls in all generations except of generations 4 and 5 (Fig 1B). We found considerable variation in mortality and mean emergence time between groups and generations. In addition, a signiﬁcant interaction between treatments and generations as well as between groups and generations was detected (all p < 0.001; nested two-way ANOVA, Table 1). Number of fertile clutches produced per female varied between 0.40 (TBT I, generation 7) and 1.15 (TBT I, generation 3) in the TBT groups and 0.59 (SC I, generation 7) and 1.4 (SC I, generation 11) in the controls (Fig. 1C). Over all generations, TBT-exposed females produced less egg masses than unexposed females (TBT ¼ 0.83 compared to SC ¼ 0.99; Fig. 1C). TBT-treated groups showed reduced population growth rates (PGR) compared to the respective control groups in all but two generations (3 and 4; Fig. 1D). Mean PGR summarized for all generations was 1.22 for TBT (without generation 10) and 1.26 in the controls. No experimental group showed a signiﬁcant time trend in larval mortality or fertility (Table 2). EmT50 values showed a clear tendency to increase over time. However, a signiﬁcant time trend was revealed only for SC II (p < 0.05). Population growth rates tended to decrease with increasing generations in all four groups. Table 1 Inﬂuence of the factors treatment, experimental group nested in treatment, generation and interaction between these factors on mortality and mean emergence time (EmT50) of Chironomus riparius. Trait Cause df SS Mortality Treatment Experimental group (treatment) Generation Treatment generation Experimental group generation 1 2 10 10 20 9482.7 9803.1 69185.4 16375.4 18183.5 EmT50 Treatment Experimental group (treatment) Generation Treatment generation Experimental group generation 1 2 10 10 20 536.8 21.1 1318.4 19.1 8.2 F p 71.960 37.196 37.719 8.928 4.957 529.9 10.4 96.1 13.9 6.0 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Shown are degrees of freedom (df), sums of squares (SS), F-values (F) and signiﬁcance levels (p). 883 Signiﬁcant effects were observed for SC II (p < 0.05) and TBT I (p < 0.01). 3.3. Genetic variation All loci exhibited allelic variation in the source strain prior to the start of the experiment (generation 0, data not shown). Allelic richness slightly decreased in the control groups over time. However, a signiﬁcant negative trend was only revealed for the expected heterozygosity in SC II (Table 2). Values for all genetic parameters recorded in this study decreased more strongly over time in the presence of TBT (Fig. 2, Table 2). Number of alleles decreased from 3.0 at the beginning of the experiment to 2.0 after 12 generations in both stressed treatments. Observed heterozygosity (HO) decreased signiﬁcantly in both TBTexposed groups (Table 2), while no signiﬁcant trend was observed in the controls. Over time, expected heterozygosity (HE) decreased signiﬁcantly only in one exposed group. However, rates of genetic impoverishment were similar in both TBT-affected groups and HE values decreased from 0.39 at the beginning to 0.21 (TBT I) and 0.20 (TBT II) in the last generation. In contrast, reduction of heterozygosity was clearly slower in the absence of sediment contamination (HE ¼ 0.31 for SC I and 0.28 for SC II in generation 12). Under TBT exposure, all genetic parameters showed a signiﬁcant (p < 0.05) decrease or at least a trend (p < 0.1) for decreasing levels of genetic variation at the microsatellite loci (Table 2). At the end of the experiments, one control group had become monomorphic at one locus. In the presence of TBT both groups lost variation at three of the ﬁve loci (data not shown). 3.4. Acute tests LC50 values based on TBT acute tests ranged from 7.33 mg Sn l1 (SEM 1.05, generation 4) to 35.7 mg Sn l1 (SEM 1.01, generation 8) in the pre-exposed groups (TBT I, TBT II). Values in the control treatments (SC I, SC II) ranged from 11.1 mg Sn l1 (SEM 1.05, generation 4) to 25.4 mg Sn l1 (SEM 1.09, generation 12; Table 3). No signiﬁcant differences were observed between pre-exposed and control groups in any test (t-test). 4. Discussion The results presented in this study provide clear evidence that chemical exposure may lead to genetic impoverishment in populations. Although signiﬁcant stressor effects were observed throughout the study, the impact of TBT on larval mortality, mean emergence time and reproduction was moderate throughout this investigation when compared with the controls. However, both number of alleles and levels of observed and expected heterozygosity decreased over time in the TBT-exposed groups. The unexposed groups showed only a slight decrease of genetic variation and no signiﬁcant negative trends were observed for most genetic parameters. Chemical exposure may lead to decreased genetic variation in populations in different ways. First, selection against less stresstolerant genotypes may increase frequencies of alleles (potentially leading to ﬁxation) that provide a selective advantage under stressed conditions (Bickham et al., 2000). This kind of selectionbased reduction of genetic variation, however, will solely affect loci which are under direct selection or physically linked to those. Neutral genetic variation, which can be visualized by microsatellite markers, is not directly affected by selection. However, contaminant exposure may lead to increased mortality and reduced fertility and thus reduce the number of individuals that contribute to the next generation (expressed as effective population size, NE). The extent of allelic variation in a population is positively correlated 884 C. Nowak et al. / Environmental Pollution 157 (2009) 881–886 SC I SC II A TBT I TBT II B 30 mortality [%] 100 EmT50 [d] 75 50 25 20 25 15 0 1 2 3 4 5 6 7 8 1 9 10 11 12 2 3 4 generation 6 7 8 9 10 11 12 generation D population growth rate [d-1] C 1.5 clutches/female 5 1.0 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 generation 1.4 1.3 1.2 1.1 1 2 3 4 5 6 7 8 9 10 11 12 generation Fig. 1. Life-history of two TBT exposed (TBT) and two unexposed (SC) Chironomus riparius populations over 12 generations. Shown are mean mortality (in %, SEM; A), Mean emergence time (EmT50, in days, SEM; B), mean number of fertile egg clutches produced per emerged female (clutches/female; C) and the population growth rate (in days1, SD; D). with its effective population size, because random genetic drift is more effective in small populations (Frankham, 1996). Thus, chemical stress can reduce neutral genetic variation due to nonselective reduction of NE. In this study we obtained no evidence for selection processes, as no signiﬁcant time trend in any life-history trait was observed, and L1 larval tolerance towards TBT did not signiﬁcantly change over time. However, TBT led to increased mortalities and a reduced Table 2 Time series analyses of four genetic parameters and ﬁve life-history traits of Chironomus riparius Trait n Repl. SC p Level Genetic variation Life-history traits þ HO 7 HE 7 NA 7 NP 7 Mortality 12 EmT50 12 Clutches/female 12 Clutch size 12 PGR 12 I II I I II I II I II I II I II I II I II * þ þ * þ * TBT Slope 0.005 0.022 0.009 0.013 0.133 0.100 0.200 0.000 2.285 0.964 0.183 0.273 0.011 0.018 10.18 6.125 0.003 0.007 p Level Slope * * þ * þ * þ þ 0.025 0.028 0.036 0.036 0.200 0.150 0.500 0.333 þ þ ** 1.946 0.048 0.483 0.290 0.024 0.013 7.825 1.375 0.010 0.004 p < 0.1; *p < 0.05; **p < 0.01. HO, observed heterozygosity; HE, expected heterozygosity, NA, mean number of alleles per locus; NP, number of polymorphic loci; EmT50, mean emergence time, n, number of data points (¼generations) considered in the analysis. number of fertile clutches in most generations. Consequently, a lower number of individuals contributed to the next generation. Our observation of decreasing genetic diversity is therefore most likely the result of reduced effective population sizes in the exposed groups, although absolute population sizes were held constant at the beginning of each generation. In a previous study using a similar experimental design, but a lower TBT concentration (measured time-weighted average: 4.46 Sn kg1 sediment dw), presence of the biocide did not signiﬁcantly lower ﬁtness traits in most generations, and had thus no clear effects on NE (Vogt et al., 2007a). Consequently, no reduction in genetic variation due to pollution was observed in this study, although some indications for selection processes were found. The fact that there was no gene ﬂow between the experimental strains contributed to the fast reduction of genetic variation. Gene ﬂow among populations restores genetic variation very effectively (Slatkin, 1987). In our experiment we simulated completely isolated populations, and thus the obtained results are not easily transferable to natural conditions. However, loss of genetic variability is rather a problem of small and isolated than of large populations. Rare and endangered species usually persist in small and isolated relict populations (Amos and Balmford, 2001). As we showed, chemical pollution might decrease genetic variation in small and isolated populations, which are the main targets of conservation efforts. Hence, our ﬁndings provide a functional link between ecotoxicology and conservation genetics. Besides the investigation of genetic variation under TBT stress, high variation in life-history traits was observed not only between treatments, but also among groups and generations throughout the multi-generational study. For instance, no stressor effect on larval mortality was apparent in generations 3 and 8. In contrast, TBT severely reduced larval survival in other generations (e.g. 5, 9 and 10). These ﬁndings document the limited signiﬁcance of single generation surveys in standard ecotoxicological studies. Although C. Nowak et al. / Environmental Pollution 157 (2009) 881–886 SC I A SC II TBT I B 3.0 2.5 TBT II 3.0 NA NA 2.5 2.0 2.0 1.5 1.5 0 2 4 6 8 10 0 12 2 4 6 8 10 12 generation generation C D 0.4 0.4 0.3 0.3 HE HE 885 0.2 0.2 0.1 0.1 0 2 4 6 8 10 12 0 2 generation 4 6 8 10 12 generation Fig. 2. Genetic variation (A, B ¼ mean number of alleles per locus [NA]; C, D ¼ expected heterozygosity [HE]) of two TBT exposed (TBT) and two unexposed (SC) Chironomus riparius populations over 12 generations. Lines show combined slope estimates (see Table 2 for values and signiﬁcance levels). conditions were kept constant throughout the study and equal concentrations of TBT were applied in all generations, a comparison of different generations (e.g. eight and nine) leads to completely different conclusions concerning the effects of the chosen TBT concentration on Chironomus larval survival. There are several ways to explain the large variation in lifehistory traits between experimental groups and generations, like heterogeneous contamination of the test sediment or slight variability of test parameters. As our data document, however, variation in the genetic composition between generations is likely to contribute to the observed life-history variation. For instance, population growth rates of all four groups investigated tended to decrease over time (signiﬁcant for SC II and TBT I) and this reduction was overall slightly stronger in the exposed groups. This decrease can most likely be explained by reduced ﬁtness due to loss of genetic variation in the respective groups. Twelve generations might not be sufﬁcient, though, to detect a signiﬁcant trend here. It is likely that genetic effects contributed to extinction of TBTexposed groups after generation 10. Larval development was already delayed in the two former generations 8 and 9. A further delay in development might have led to a complete extinction because of rapidly declining water quality, which can be observed in most test vessels after approximately 30 days of test duration (data not shown). High mortalities have also been found in the previous multi-generational experiment by Vogt et al. (2007a), Table 3 Acute tests with Chironomus riparius L1 larvae from generations 4, 8, and 12 of the multi-generation study. Treatment Experimental group Gen. 4 Gen. 8 Gen. 12 SC I II 11.1 1.05 17.2 1.11 21.9 1.04 21.5 1.04 25.4 1.09 17.1 1.02 TBT I II 7.33 1.05 20.4 1.08 35.7 1.01 23.4 1.07 22.4 1.07 20.7 4.92 Shown are calculated LC50 values (in mg TBT-Sn l1 SEM). probably indicating the presence of survival-limiting factors that can hardly be controlled in the laboratory. However, we decided to continue experiments after the breakdown using backup groups of the respective populations taken from the previous generation. These groups were kept for one generation under unstressed conditions, which should not severely affect experimental outcomes. However, even when considering only the ﬁrst 10 generations, signiﬁcant reduction of genetic variation was still obvious for most parameters (data not shown). The reduction of genetic variation, however, does not provide sufﬁcient explanation for the extinction event in generation 10, because the backup strains showed considerable survival rates in the last two generations of the experiment. These backup strains derive from the same egg masses as the strains used in the lifehistory experiments. It would thus be reasonable to expect similar reduced survival in the backup strains when exposed to TBT in generations 11 and 12. Continuing the experiments for two additional generations after the breakdown did prove that other explanations, such as seasonal effects or variations in exposure conditions, likely contributed to the observed high variation in life history among generations. 5. Conclusions We showed that toxic substances have the potential to reduce genetic variation before severe population declines due to high mortality rates or low fertility occurs. While life-history responses to contamination can remain approximately constant over time, resulting in stable population sizes, genetic variation may constantly decrease further over time. The consequences of genetic impoverishment on ﬁtness and tolerance towards chemical exposure and other stress factors may only be visible over longer time periods. Therefore, loss of genetic variation has to be considered as an additional threat to populations subjected to chronic environmental pollution (Theodorakis, 2001; Van Straalen and Timmermans, 2002). It is also important to stress the relevance of genetic variation in 886 C. Nowak et al. / Environmental Pollution 157 (2009) 881–886 populations for the survival in contaminated environments. Nowak et al. (2007b), for instance, have shown that C. riparius groups with low degrees of allelic variation display a signiﬁcantly lowered cadmium tolerance compared to more diverse groups. Genetic impoverishment might therefore have severe long-term consequences for populations in human-affected environments when stressors change over time. Even if selection may allow for local adaptation, its effectiveness relies on the presence of a sufﬁcient degree of genetic variation (Frankham et al., 2002). The loss of intraspeciﬁc diversity might therefore lead to an increased susceptibility to environmental stress, accompanied by a lowered ability to adapt to stressful environmental conditions. In order to ensure longterm viability of populations in polluted environments, it is therefore crucial to maintain sufﬁciently high population sizes and to ensure gene ﬂow between populations, keeping genetic variation high and preventing genetic erosion. In addition, rare and isolated populations should be kept from pollution more rigidly than large populations, because they might lose variation more rapidly, have lowered potential for genetic adaptation and might display enhanced stress susceptibility, as documented for C. riparius in laboratory tests (Nowak et al., 2007b, 2008; Vogt et al., 2007c). Investigating the inﬂuence of pollution on patterns of genetic variation represents an important future challenge for ecotoxicologists and conservation geneticists (Belﬁore, 2001). Only comprehensive ﬁeld studies, which consider the high complexity of factors impacting the genetic structure of natural populations, will show the relevance of genetic erosion in polluted environments. Linking ecological and population genetic processes will eventually allow us to predict the long-term fate of natural populations and to assess the consequences of human made changes to natural communities. Acknowledgments We appreciate the kind assistance of several undergraduate students working on this project in the Departments ‘‘Aquatic Ecotoxicology’’ and ‘‘Ecology & Evolution’’. This work was ﬁnancially supported by the ‘‘Programm Lebensgrundlage Umwelt und ihre Sicherung (BWPLUS)’’ of the federal state Baden-Württemberg (contract number BWR 22018). References Amos, W., Balmford, A., 2001. When does conservation genetics matter? Heredity 23, 257–265. Armitage, P., Cranston, P.S., Pinder, C.V., 1995. The Chironomidae: Biology and Ecology of Non-Biting Midges. Chapman & Hall, London. Avise, J., 2004. Molecular Markers, Natural History, and Evolution. Sinauer Associates, Sunderland. Belﬁore, N.M., 2001. Effects of contaminants on genetic patterns in aquatic organisms: a review. Mutation Research-Reviews in Mutation Research 489, 97–122. Bickham, J.W., Sandhu, S., Hebert, P.D.N., Chikhi, L., Athwal, R., 2000. Effects of chemical contaminants on genetic diversity in natural populations: implications for biomonitoring and ecotoxicology. Mutation Research-Reviews in Mutation Research 463, 33–51. Brown, J.W., Van Coeverden de Groot, P.J., Birt, T.P., Seutin, G., Boag, P.T., Friesen, V.L., 2007. Appraisal of the consequences of the DDT-induced bottleneck on the level and geographic distribution of neutral genetic variation in Canadian peregrine falcons, Falco peregrinus. Molecular Ecology 16, 327–343. Chen, X.Y., Li, N., Shen, L., Li, Y.Y., 2003. Genetic structure along a gaseous organic pollution gradient: a case study with Poa annua L. Environmental Pollution 124, 449–455. Frankham, R., 1996. Relationship of genetic variation to population size in wildlife. Conservation Biology 10, 1500–1508. Frankham, R., Ballou, J.D., Briscoe, D.A., 2002. Introduction to Conservation Genetics. Cambridge University Press, Cambridge. Gillis, P., Diener, L.C., Reynoldson, T.B., Dixon, D.G., 2002. Cadmium-induced production of a metallothionein like protein in Tubifex tubifex (Oligochaeta) and Chironomus riparius (Diptera): correlation with reproduction and growth. Environmental Toxicology and Chemistry 21, 1836–1844. Kim, S.J., Rodriguez-Lanetty, M., Suh, J.H., Song, J.I., 2003. Emergent effects of heavy metal pollution at a population level: Littorina brevicula as a study case. Marine Pollution Bulletin 46, 74–80. Krane, D.E., Sternberg, D.C., Burton, G.A., 1999. Randomly ampliﬁed polymorphic DNA proﬁle-based measures of genetic diversity in crayﬁsh correlated with environmental impacts. Environmental Toxicology and Chemistry 18, 504–508. Murdoch, M.H., Hebert, P.D.M., 1994. Mitochondrial L-DNA diversity of brown bullhead from contaminated and relatively pristine sites in the great-lakes. Environmental Toxicology and Chemistry 13, 1281–1289. Nadig, S.G., Lee, K.L., Adams, S.M., 1998. Evaluating alterations of genetic diversity in sunﬁsh populations exposed to contaminants using RAPD assay. Aquatic Toxicology 43, 163–178. Nowak, C., Hankeln, T., Schmidt, E.R., Schwenk, K., 2006. Development and localization of microsatellite markers for the sibling species Chironomus riparius and Chironomus piger (Diptera: Chironomidae). Molecular Ecology Notes 6, 915–917. Nowak, C., Vogt, C., Barateiro, J., Schwenk, K., 2007a. Genetic impoverishment in laboratory cultures of the test organism Chironomus riparius. Environmental Toxicology and Chemistry 26, 118–122. Nowak, C., Jost, D., Vogt, C., Oetken, M., Schwenk, K., Oehlmann, J., 2007b. Effects of inbreeding and reduced genetic variation on tolerance to cadmium stress in the midge Chironomus riparius. Aquatic Toxicology 85, 278–284. Nowak, C., Czeikowitz, A., Vogt, C., Oetken, M., Streit, B., Schwenk, K., 2008. Variation in tolerance to cadmium exposure among genetically characterized laboratory populations of the midge Chironomus riparius (Diptera: Chironomidae). Chemosphere 71, 1950–1956. OECD, 1998. Guideline for testing of chemicals no. 211. Daphnia magna reproduction test, adopted September 1998. Organization for Economic Development and Cooperation, Paris. OECD, 2004. Sediment-water chironomid toxicity test using spiked sediment. OECD guidelines for the testing of chemicals. (Original guideline 218, adopted 13th April 2004). Organization for Economic Development and Cooperation, Paris. Peakall, R., Smouse, P.E., 2006. GenAlEx 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes 6, 288– 295. Slatkin, M., 1987. Gene ﬂow and the geographic structure of natural populations. Science 236, 787–792. Theodorakis, C.W., 2001. Integration of genotoxic and population genetic endpoints in biomonitoring and risk assessment. Ecotoxicology 10, 245–256. Van Straalen, N.M., Timmermans, M., 2002. Genetic variation in toxicant-stressed populations: An evaluation of the ‘‘genetic erosion’’ hypothesis. Human and Ecological Risk Assessment 8, 983–1002. Vogt, C., Nowak, C., Barateiro Diogo, J., Oetken, M., Schwenk, K., Oehlmann, J., 2007a. Multi-generation studies with Chironomus riparius - Effects of low tributyltin concentrations on life-history parameters and genetic diversity. Chemosphere 67, 2192–2200. Vogt, C., Belz, D., Galluba, S., Nowak, C., Oetken, M., Oehlmann, J., 2007b. Effects of cadmium and tributyltin on development and reproduction of the non-biting midge Chironomus riparius (Diptera) - baseline experiments for future multigeneration studies. Journal of Environmental Science and Health-Part A 42, 1–9. Vogt, C., Pupp, A., Nowak, C., Jagodzinski, L.S., Baumann, J., Jost, D., Oetken, M., Oehlmann, J., 2007c. Interaction between genetic diversity and temperature stress on life-cycle parameters and genetic variability of Chironomus riparius populations. Climate Research 33, 207–214. Wade, T.L., Sweet, S.T., Quinn, J.G., Cairns, R.W., King, J.W., 2004. Tributyltin in environmental samples from the former Derecktor Shipyard, Coddington Cove, Newport RI. Environmental Pollution 129, 315–320. Walker, C.H., Hopkin, S.P., Sibly, R.M., Peakall, D.B., 2001. Principles of Ecotoxicology. Taylor & Francis, London. Whitehead, A., Anderson, S.L., Kuivila, K.M., Roach, J.L., May, B., 2003. Genetic variation among interconnected populations of Catostomus occidentalis: implications for distinguishing impacts of contaminants from biogeographical structuring. Molecular Ecology 12, 2817–2833.