DEVELOPMENTAL STABILITY IN AMPHIBIANS AS A BIOLOGICAL INDICATOR OF CHEMICAL CONTAMINATION AND OTHER ENVIRONMENTAL STRESSORS Project Managers: Howard H. Whiteman (PI) Center of Excellence in Ecosystem Studies and Department of Biological Sciences Murray State University Murray, KY 42071 (270) 762-6753 [email protected] Bommanna G. Loganathan (Co-PI) Chemical Services Laboratory and Department of Chemistry Murray State University Murray, KY 42071 (270) 762-3044 [email protected] Estimated Total Cost Year 1 Year 2 Grand Total Federal: $94,708 $73,641 $168,349 State: $49,250 $38,984 $88,234 Local: $49,771 $38,588 $88,359 Total: $193,729 $151,213 $344,942 1 Sorting Code: 2001-NCER/EPSCoR DEVELOPMENTAL STABILITY IN AMPHIBIANS AS A BIOLOGICAL INDICATOR OF CHEMICAL CONTAMINATION AND OTHER ENVIRONMENTAL STRESSORS Howard H. Whiteman (PI) Center of Excellence in Ecosystem Studies and Department of Biological Sciences Murray State University Bommanna G. Loganathan (Co-PI) Chemical Services Laboratory and Department of Chemistry Murray State University PROJECT SUMMARY The global decline of amphibians is considered a disturbing indicator of environmental degradation because it may forebode of cascading ecological effects, as well as raising health concerns about human populations (Wyman 1990, Wake 1998). Amphibians are ideal biological indicators, because their semi-permeable epidermis and complex life cycle expose them to multiple stressors in both aquatic and terrestrial environments (Wyman 1990). Because of this, amphibians should be among the first vertebrates affected by anthropogenic stressors in either of these environments (Stebbins and Cohen 1995). Furthermore, some of the same stressors affecting amphibians are known to have negative effects on other species, including humans (e.g., PCBs, UV light, etc.; Wake 1998, Carey 2000). Previous research in our laboratories has shown that developmental stability (as measured by fluctuating asymmetry) is correlated to human-induced stressors in two amphibian species (Whiteman unpubl. data). We propose to continue our studies of developmental stability as a biological indicator of stress in amphibian populations, in an effort to improve the ability of biologists to evaluate the health of these populations, the surrounding ecosystem, and the possible environmental risks to humans. Although we propose to evaluate a wide range of possible stressors, a major thrust of this project is to correlate amphibian developmental stability with contaminant levels accumulated in their tissues. Surprisingly, although many studies have used developmental stability as an indicator of such contamination (Moller and Swaddle 1997), none of these studies have concentrated on amphibians. We hypothesize that amphibian developmental stability will decrease with increased levels of anthropogenic (contaminants, land use practices) and natural (population size and density) stressors. To test this hypothesis, we will sample several species of amphibians, including both aquatic and terrestrial life stages, and taxa from various trophic levels which vary in their potential exposure to stressors. In this way we will be able to discover how well developmental stability operates as a bio-indicator across a wide range of conditions. We will measure developmental stability and various natural and anthropogenic stressors, and use the results to assess the utility of developmental stability as a biological indicator in amphibians. Development of these methods will allow scientists to determine the population health of a critical group of organisms, to utilize amphibian development as an inexpensive yet accurate proxy for measuring the effects of stressors such as industrial and agricultural pollutants, and will provide methods to separate the effects of anthropogenic stressors from natural ones. This 2 research will thus be valuable to the estimation and management of ecosystem and human health. 3 PROJECT RATIONALE Rationale for Project Selection This Kentucky EPA/EPSCoR proposal presents a plan of results-oriented, multidisciplinary research aimed at deciphering one of the most timely and urgent environmental problems, the use of biological indicators to understand actual and potential risks to population, ecosystem and human health. Using state-of-the-art, EPA approved methods and protocols, the PIs, graduate students, and undergraduate assistants will obtain baseline data on how anthropogenic and natural stressors affect amphibian development across a range of life stages, species, and habitat types. This research will specifically test the ability of developmental stability to act as a biological indicator of such stressors, and to separate those of anthropogenic versus natural origin. Institutional Goals Murray State University is dedicated to developing the future careers of science and engineering students and faculty. EPA/EPSCoR funds will complement recent awards made by the National Science Foundation (Collaborative Research at Undergraduate Institutions program) and the Howard Hughes Medical Institute which are aimed at developing student and faculty research and teaching with an emphasis on promoting undergraduates that are currently underrepresented in science and technology. To this end, the PIs will advertise broadly for graduate and undergraduate students needed to complete this project, and actively recruit women, minorities, and persons with disabilities. Currently, the PIs have eight undergraduates and four graduate students in their laboratories (eleven women, one minority), and have been actively involved in programs designed to encourage minority development within the sciences (e.g., NSF REU and C-RUI, ESA SEEDS). 4 PROJECT DESCRIPTION This Kentucky EPA/EPSCoR proposal presents a plan of results-oriented, multidisciplinary research aimed at deciphering one of the most timely and urgent environmental problems, the use of biological indicators to understand actual and potential risks to population, ecosystem and human health. Using state-of-the-art, EPA approved methods and protocols, the PIs, graduate students, and undergraduate assistants will obtain baseline data on how anthropogenic and natural stressors affect amphibian development across a range of life stages, species, and habitat types. This research will specifically test the ability of developmental stability to act as a biological indicator of such stressors, and to separate those of anthropogenic versus natural origin. Objectives One of the most important, yet most difficult, tasks associated with understanding and managing the health of ecosystems is the identification of populations subject to stress before such stress has a detrimental effect (Clarke 1995). This is particularly true of amphibians; the global decline of amphibian populations is considered a disturbing indicator of environmental degradation because it may forebode of cascading ecological effects, as well as raising health concerns about human populations (Wake 1998). Amphibians are ideal biological indicators, because their semi-permeable epidermis and complex life cycle expose them to multiple stressors in both aquatic and terrestrial environments (Wyman 1990). Because of this, amphibians should be among the first vertebrates affected by anthropogenic stressors in either environment (Stebbins and Cohen 1995). Furthermore, some of the same stressors affecting amphibians are known to have negative effects on other species, including humans (e.g., PCBs, UV light, etc.; Wake 1998, Carey 2000). Recently, an increase in incidence of malformed frogs has been observed throughout North America. These observations have typically involved gross changes in morphology such as extra or missing limbs and eyes. Deformed frogs are currently hypothesized to be indicators of developmental problems associated with anthropogenic stress, and evidence testing this idea is currently being evaluated (e.g., Diana and Beasley 1998, Helgen et al. 1998, Rowe et al. 1998ab, Johnson et al. 1999, Sessions et al. 1999). Although deformed frogs serve as a warning for the management of nearby amphibian populations as well as human health concerns, they may appear too late to save local populations, and possibly too late to reduce anthropogenic stress to the surrounding ecosystem as well. Biologists thus need an early-warning biological indicator that could identify environmentally-stressed animals before the stressor causes population, ecosystem, and/or regional harm. Such an indicator should be able to measure stress-induced effects before drastic changes in morphology or immunocompetence take place which would subsequently decrease the organism's survival and reproductive abilities. One such indicator is developmental stability (Clarke 1995). Developmental stability is one component of an organism’s ability to withstand environmental and genetic disturbances to produce a genetically predetermined phenotype (Waddington 1942, Lerner 1954; for a review, see Moller and Swaddle 1997). Specifically, developmental stability reduces phenotypic variation resulting from developmental accidents. Under normal conditions, development follows a genetically determined pathway, and minor perturbations are controlled by developmental stability mechanisms. Under stressful conditions 5 (e.g., increased pollutants), the performance of the stability mechanism may be reduced such that development cannot be restored to the original pathway, resulting in the production of abnormal phenotypes (Waddington 1942, Clarke 1995). Developmental stability can thus provide an indirect measure of an organism's fitness, and numerous studies have found significant correlations between measures of developmental stability and fitness (e.g., Quattro and Vrijenhoek 1989, Moller 1992a, b, McKenzie and O'Farrell 1993). Further, stress-induced changes in developmental stability are typically observed before any detectable change occurs in fitness (Clarke et al. 1986, Clarke and McKenzie 1992), providing an early-warning mechanism for population monitoring. One of the most widely used measures of developmental stability is fluctuating asymmetry (FA). FA is nondirectional differences between the left and right sides of paired bilateral characters within a population (Thoday 1955, 1958, Van Valen 1962). The underlying assumption of this measure is that development of both sides of a bilaterally symmetric organism is influenced by identical genes, and thus nondirectional differences between sides must be environmental in origin (Waddington 1942, Moller and Swaddle 1997). Because developmental stability acts to reduce such changes, FA will measure the efficiency of developmental stability and the magnitude of the environmental perturbation (Clarke 1995). Because measures of developmental stability, such as FA, can be used to identify stressed populations before more significant deleterious effects are observed, and because such measures can also be used to estimate changes in fitness before they occur, developmental stability has the potential to be a critical biological indicator. Many studies have used developmental stability as an indicator of stressors, including lead, benzene, mercury, PCBs, DDT, various pesticides, and UV radiation in a diverse array of organisms from algae to grey seals (Moller and Swaddle 1997). Surprisingly, although development has been studied more extensively in amphibians than perhaps any other group (Duellman and Trueb 1986), and many stressors have been shown to have major developmental effects on amphibians (Diana and Beasley 1998, Rouse et al. 1999; Appendix 1) developmental stability is only now being applied to amphibians. Previous studies by the PIs and their students have shown that developmental stability can be used as a biological indicator of anthropogenic stressors. Studies of bullfrog (Rana catesbeiana) tadpoles and tiger salamander (Ambystoma tigrinum) adults revealed that FA increased with increasing levels of agricultural disturbance and water quality degradation (Fig. 1). We have also documented tissue contamination associated with developmental deformities in bullfrog tadpoles. Normal tadpoles exhibited total PCB congener and total pesticide concentrations ranging from detection limit (0.35 ng g-1 ) to 10.36 ng g-1 wet weight and detection limit (0.35 ng g-1) to < 6 ng g-1 wet weight respectively. Grossly deformed tadpoles (tumors, axial deformations), in contrast, exhibited elevated total PCB concentration (>20 ng g-1 wet weight), but did not differ from normal tadpoles in total pesticide concentration (Seaford et al. 2000). Unfortunately, lack of funds has limited our ability to test whether FA correlates with these tissue contamination levels, and to analyze amphibian tissues for other contaminants such as butyltins and PAHs. 6 0.030 EEN (FH = 5L.6, p = (0.0F4) Fig. 1: FA increases with anthropogenic 2,6 disturbance level (low, medium, high) in 0.025 = 10.0, p = 0.01) bullfrog(Rana catesbeiana) tadpoles from 2,6 western Kentucky. Low disturbance 0.020 areas are ponds surrounded completely by forest, medium are farm ponds with 0.015 associated trees and shrubs, and high areas are farm ponds without any trees or other non-agricultural vegetation. Two different 0.010 FA measures are shown: EN (eye to nares) and EHL (eye to hind limb). 0.005 0.000 Low Medium High Disturbance We propose to continue our research into this biological indicator by correlating FA with a number of natural and anthropogenic stressors that might influence variation in developmental stability. These variables include: 1. amphibian tissue concentrations of organic, organometallic, and trace elements commonly associated with industrial and agricultural pollution (Appendix 2), which have been implicated in endocrine disruption and developmental problems, including amphibian abnormalities (Appendix 1); 2. important water chemistry measures, which have been associated with amphibian mortality and deformities (Appendix 1) and have been correlated with significant FA in previous studies (see above); 3. land use practices, i.e., undisturbed forested sites, moderately disturbed agricultural sites and highly disturbed industrial sites; 4. population density, which at high levels can induce stress and FA (Clark and McKenzie 1992); 5. adult population size, which may increase FA via inbreeding depression in small populations (Quattro and Vrijenhoek 1989). Population estimates will allow separation of natural stress levels from those that may be human induced, which a recent NSF workshop identified as a critical goal of amphibian conservation (Wake 1998). We generally predict that FA will increase with increased tissue contamination, decreased water quality, increased land use degradation, and in populations with high densities (particularly in larvae) and/or very small adult population sizes. Approach and Methods The goal of this project is to utilize developmental stability as an biological indicator in several amphibian species and life stages that vary in habitat, life history, behavior, and resource utilization (Table 1). All of these factors should influence the exposure and susceptibility of these organisms to toxins through biomagnification. We will explore these general questions across: A. areas with variable exposure to toxins, including relatively pristine sites, moderately disturbed (agricultural) sites, and heavily disturbed (industrialized) sites. B. a variety of species and life stages that we predict will have differential exposure to stressors, including: i. larval vs. adult forms; ii. species that vary in their use of aquatic vs terrestrial habitats; and iii. species that vary in trophic level, and thus bioaccumulation and biomagnification levels. 7 Clearly, studying FA across habitats with different potential stress levels (A) will directly evaluate the use of FA as a surrogate estimator of contamination and ecosystem health. Testing a wide variety of species and life stages (B) will allow us to determine which kinds of species and stages are the best ecological indicators, and thus which should be utilized in future investigations. Table 1: Life stages, trophic levels, and habitats of proposed study species. Species Larvae Adult1 Habitat Trophic Level Habitat Breeding Habitat Bullfrog (Rana catesbeiana) aquatic herbivore2 semi-terr. aquatic Leopard frog (Rana utricularia) aquatic herbivore2 semi-terr aquatic Spotted Sal. (Ambystoma maculatum) aquatic carnivore3 terrestrial aquatic Slimy Sal. (Plethodon glutinosis) terrestrial4 carnivore3 terrestrial terrestrial 1All adults are carnivorous. 2 primary consumer 3secondary consumer 4larvae metamorphose within the egg; thus we will sample juveniles. Each of these different comparisons set up tests of different potential exposure levels to assess FA as an biological indicator. For example, in most of the proposed study species there are aquatic larval stages and terrestrial adult stages (Table 1). From a temporal standpoint, adults should have accumulated more contaminants and thus have increased FA compared to larvae. However, aquatic larval stages may experience increased exposure due to their aquatic nature when compared to terrestrial (often fossorial) adults. Further, adults may reflect selection against heavily exposed individuals, i.e., larvae may reveal indications of asymmetry better than adults because heavily exposed and potentially asymmetric larvae never survive to adulthood (see also Burger and Snodgrass 2000). Our comparisons of salamander species with aquatic versus terrestrial reproduction should help test this hypothesis. Similarly, different amphibian species spend various amounts of time in aquatic vs. terrestrial habitats (e.g., adult habitat vs. breeding habitat in Table 1). This differential exposure to the aquatic environment may also influence the degree of exposure and bioaccumulation of contaminants, and thus the predicted level of FA. Bioaccumulation and FA should also differ among species and life stages that vary in their trophic level, with herbivorous tadpoles having the lowest levels, carnivorous salamander larvae intermediate ones, and carnivorous adults the highest. Targeted Contaminants Although we will be assessing a wide variety of water and soil chemistry parameters, as well as tissue contaminantion (Appendix 2), we will target a number of variables which have particular relevance to amphibian development and life history (Appendix 1), and which are found in the western Kentucky area (Table 2). These include polychlorinated biphenyl (PCB) congeners, chlorinated hydrocarbon pesticides, polynuclear aromatic hydrocarbons (PAHs), butyltins, trace inorganic elements (including trace metals such as mercury, cadmium, and aluminum), hydrogen ions (i.e., pH) and associated aluminum, and nitrogen (ammonium, nitrate). Table 2: Selected Industrial and Agricultural Pollutants in Western Kentucky Sample Medium References Industrial Pollutants Polychlorinated biphenyls pine needles Loganathan et al. 1998a sediment Loganathan et al. 1998b freshwater mussels Loganathan et al. 1998b amphibians Seaford et al. 2000 8 Butyltins sediment Loganathan et al. 2000 (Mono, di-, tri-) freshwater mussels Agricultural Pollutants DDT and metabolites pine needles Loganathan et al. 1998a Chlordane compounds sediment Loganathan et al. 1998b Hexachlorocyclohexane freshwater mussels Loganathan et al. 1998b amphibians Seaford et al. 2000 Hydrogen ions (pH) pond water Whiteman et al. unpubl. Nitrogen pond water Whiteman et al. unpubl. Phosphate pond water Whiteman et al. unpubl. Previous work by the PIs in western Kentucky has provided evidence of these pollutants and, in some cases, their bioaccumulation. For example, westernmost Kentucky is endowed with a variety of industries and state-of-the-art agricultural operations. A report by the Kentucky Environmental Protection Cabinet (KEPCR 1997) documented high levels (>9.68 million lbs toxic release inventory (TRI)) of toxic pollutants released by the industries of this region. Recent biomonitoring studies have revealed that pine needles (an excellent bioindicator of atmospheric organic pollutants), sediments, and mussel tissues from this region contain significant levels of inorganic ions, as well as detectable concentrations of persistent, bioaccumulative and toxic (PBTs) compounds. These include organochlorine (chlorinated pesticides and polychlorinated biphenyls) and organometallic (butyltin) compounds and their metabolites (Loganathan et al. 1998a, 1998b, 1999, 2000). The presence of organochlorine and organometallic parent compounds such as 4,4’-DDT and tributyltin (TBT) respectively in sediments and mussel tissues indicates recent input into the environment (Loganathan et al. 1998a, 1998b). Furthermore, recent studies on extractable organohalogens (EOX) in sediment and mussel tissues have revealed the presence of additional unknown organochlorine residues in this region (Loganathan et al. 2001). Suspected sources for these toxic contaminants include: non-point sources, industrial or agricultural origin, septic tank effluents, atmospheric transport and shipping activities. Toxic pollutants such as polychlorinated biphenyls (PCBs), chlorinated pesticides, polynuclear aromatic hydrocarbons (PAHs) and butyltin compounds are of particular concern because they are known or suspected of causing cancer and/or other health effects such as endocrine disruption in a wide variety of animals in freshwater and marine ecosystems (Loganathan et al. 1994, Jobling et al. 1998, Giesy and Snyder 1997, Khim et al. 1999, Loganathan et al. 2000). In particular, amphibian development and metamorphosis is hormonally regulated and may be especially vulnerable to the effects of endocrine disrupting chemicals (Patlack 1996, Diana and Beasley 1998). General Amphibian Sampling and Measurements of Fluctuating Asymmetry (FA) Ponds and surrounding habitats will be sampled for the various species and life stages listed in Table 1. At least 40 ponds with varying potential stress levels (pristine, agricultural, and industrial sites) and which contain breeding populations of target species will be chosen for study. Ponds will be sampled for adults during breeding events (Table 1) and for larvae several times a year. Forty-six potential study sites have already been strategically located throughout western Kentucky (Fig. 2) to take advantage of land use practices (forested, industrial, and agricultural) as well as prevailing winds (i.e., downwind of major industrial centers). Many of these sites were used in previous studies, and other sites that are immediately adjacent to industrialized areas have been located. Slimy salamanders, which breed terrestrially, will be collected in nearby forested areas using standard herpetological capture techniques. 9 Fig. 2: Locations of 46 sampling sites in the Purchase region of western Kentucky; each dot represents a single pond. Note the strategic locations of sites in forested areas (Land Between the Lakes), agricultural areas (west and north of Murray), and those downwind of industrial areas (east of Paducah, Calvert City, and Murray). Amphibians will be captured using dip nets, minnow traps, and seines. At least 25 larvae and 10-12 adults (males, because they are typically much more abundant than females) will be sampled from each population. Previous work has shown that these sample sizes will provide excellent power (> 0.8) for statistical analysis (Whiteman et al. unpubl. data). After capture, individuals will be transported in buckets back to Murray State for digital imaging. Each individual will be anesthetized using tricaine methylchloride (MS-222), sexed (adults only) and measured for snout-vent length (SVL; to nearest mm) with a ruler and mass (g) with a balance. Each individual will then be photographed with a Pixera Professional digital camera which is connected to a PC (already available). We utilize digital imagery because it reduces measurement error considerably when compared to other techniques (Fig. 3). In addition, digital imagery allows us to easily take multiple measurements at a later time, rather than immediately while the animal is in the laboratory. After imaging is complete, most animals will be submerged in aged water to revive them from sedation and released back to their pond/area of capture. A subsample of captured animals will be preserved for analysis of bioaccumulated contaminants (see below). Measurements of FA will concentrate on morphological structures directly related to amphibian fitness, particularly measures associated with the head, mouth, and torso which have already been shown to exhibit significant FA in other species (Table 3; Fig. 1, 3; Whiteman et al., unpubl. data). Each individual will be measured three separate times in order to statistically analyze measurement error (Palmer 1994) 10