A Dissertation. Entitled. Genetic Portraits of Introduced Gobies and Mussels: Population Variation Delineates Invasion Pathways. Joshua E.

Transcription

1 A Dissertation Entitled Genetic Portraits of Introduced Gobies and Mussels: Population Variation Delineates Invasion Pathways By Joshua E. Brown Submitted as partial fulfillment of the requirements for The Doctor of Philosophy Degree in Biology (Ecology) Advisor: Dr. Carol A. Stepien Committee Member: Dr. Jonathan M. Bossenbroek Committee Member: Dr. Lynda D. Corkum Committee Member: Dr. Christine M. Mayer Committee Member: Dr. Jeffrey G. Miner Dr. Patricia R. Komuniecki, Dean College of Graduate Studies The University of Toledo December 2009

2 Copyright 2009 This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author.

3 An Abstract of Genetic Portraits of Introduced Gobies and Mussels: Population Variation Delineates Invasion Pathways Joshua E. Brown Submitted as partial fulfillment of the requirements for The Doctor of Philosophy Degree in Biology (Ecology) The University of Toledo December 2009 The growing numbers of species introductions, with many having significant ecological and economic impacts, constitute one of the greatest challenges facing our native ecosystems today. To make correct ecological comparisons among native and introduced populations and minimize their further spread, we must (1) identify the introduced species/taxon/population correctly and (2) determine its source population(s) and its transmission pathways. This dissertation study analyzes and compares the population genetic and phylogeographic structure of three successful invasions by Ponto- Caspian species into the North American Great Lakes and beyond: the round goby Neogobius melanostomus and the dreissenid mussels Dreissena polymorpha and iii

4 Dreissena rostriformis bugensis. In Chapters 2 and 3 we describe evidence for two subspecies of round goby, both of which have expanded their range. From this background information, likely sources are identified for invasive populations in Eurasia and North America. In Chapter 4, the genetic structure of zebra and quagga mussels in North America is analyzed in comparison with sites in Eurasia. Zebra mussel populations have appreciable genetic diversity, whereas quagga mussel populations from the Colorado River and California show some founder effects. The population genetic composition of both species changed over time at given sites; with some adding alleles from adjacent populations, some losing them, and all retaining closest similarity to their original composition. Zebra and quagga mussels from the western United States assign to possible origins from the St. Lawrence River and Lake Ontario, respectively. These assignments suggest that overland colonization pathways via recreational boats do not necessarily reflect the most proximate connections. The results show that all three species likely experienced multiple introductions into the Great Lakes, which brought a large proportion of the native genetic diversity to North America. This created significant genetic structure within their respective introduced ranges. The round goby and the quagga mussel were introduced from their native range, in contrast to the zebra mussel which whose origins North American origins trace to secondary spread from previously invaded regions in Northern and Central Europe. This dissertation demonstrates the utility of molecular techniques to invasive species management, by identifying the pathways connecting source populations to new colonies. iv

5 Acknowledgements This dissertation is an exploration of knowledge and humility for me, and has only been possible through the combined efforts of many people. The highest gratitude goes to Dr. Carol Stepien, who has sacrificed untold time and energy to help me to become a better scholar, a clearer thinker and a wiser man. Thank you for the encouragement, the support, the resources, and the opportunity to progress and learn. Without you, this would never have happened thank you. I thank my advisory committee as well: Drs. Jon Bossenbroek, Lynda Corkum, Christine Mayer, and Jeffrey Miner. Each of you has played a central role in improving the quality of this work, and my experience as a doctoral student. My gratitude is extended to friends and colleagues in the Great Lakes Genetics Laboratory, the Lake Erie Center, and the Department of Environmental Sciences - you provided insights into the processes and mechanisms underlying the data presented here, and helped me to make it all make sense. I appreciate the support and advice of the National Sea Grant Office staff - thank you for helping me see how we bring science to the people and letting me have a hand in doing so it has been an incredible year. My parents and family have been tireless supporters of my education from a very early age thank you for kindling the fire of learning. And last, but never least, my dear wife Amy I know that this has been almost as much work for you as for me. I appreciate your unflagging support and encouragement. v

6 Table of Contents Abstract... iii Acknowledgements... v Table of Contents... vi List of Tables... x List of Figures... xii Preface... xiv Chapter 1: Introduction... 1 The Study Locations... 2 The Study Organisms... 3 Invasion History... 3 Objectives... 5 Chapter 2: Ancient divisions, recent expansions: Phylogeography and population genetics of the round goby Neogobius melanostomus... 7 Abstract... 7 Introduction... 8 Biogeographic history of the Ponto-Caspian region Questions Materials and Methods Sampling, DNA extraction, and genetic data collection Data analysis vi

7 Results Overall genetic variability and patterns Black/Azov Sea vs. Caspian Sea basins Patterns within the Black/Azov Sea basins Patterns within the Caspian Sea basin Non-native locations The exotic Baltic Sea population site Black Sea non-native locations Caspian Sea non-native locations Discussion Ancient Divisions Recent Expansions Summary & Conclusions Tables & Figures Chapter 3: Invasion genetics of the Eurasian round goby in North America: Tracing sources and spread patterns Abstract Introduction Materials and Methods Sampling, DNA extraction, and amplification Genetic analyses Results Genetic composition vii

8 Invasion source and genetic structure Discussion High genetic diversity and its primary source Invasion Population Structure Genetic comparisons with other exotic species Conclusions and summary Tables & Figures Chapter 4: Population genetic history of the dreissenid mussel invasion: Expansion patterns across North America Abstract Introduction Role of genetics Materials and Methods Sampling, DNA extraction, and amplification Genetic analyses Results Invasive population structure Genetic diversity levels Temporal genetic change Discussion The 100 th Meridian Initiative Conclusions Tables and Figures viii

9 Chapter 5: Conclusions and Future research General Conclusions Future Directions References ix

10 List of Tables Table 2.1 Eurasian round goby specimen information.. 32 Table 2.2 Cytochrome b and microsatellite primers used.34 Table 2.3 Partitioning of genetic variation between round goby population divisions and among sampling sites within them using AMOVA..35 Table 2.4 Pairwise divergences between round goby population samples using θ ST for the Black Sea basin and the Caspian Sea basin...37 Table 2.5 Pairwise contingency test results for round goby samples from the Black/Azov Sea and Caspian Sea basins...39 Table 2.6 Nested clade analysis of round goby cytochrome b haplotypes...41 Table 3.1 North American round goby specimen information Table 3.2 Analysis of Molecular Variance results showing partitioning of genetic variation within and among round goby samples.79 Table 3.3 Pairwise ST comparisons between round goby samples from pooled locations Table 3.S1 Occurrence of mtdna cytochrome b haplotypes in North American round goby samples.87 Table 4.1 Zebra mussel and quagga mussel specimen information.119 Table 4.2 Microsatellite primers used for zebra mussel and quagga mussels.121 x

11 Table 4.3 Pairwise F ST and contingency tests for zebra mussel and quagga mussel locations..123 Table 4.4 Analysis of Molecular Variance showing partitioning of genetic variation Table 4.5 Assignment test for zebra mussels and quagga mussels.128 xi

12 List of Figures Fig. 2.1 Eurasian round goby distribution and sampling locations...45 Fig. 2.2 Neighbor-joining tree of phylogenetic relationships among common Eurasian round goby Neogobius melanostomus haplotypes.46 Fig. 2.3 Genetic distance tree among round goby population sites based on combined mitochondrial cytochrome b and microsatellite data.47 Fig. 2.4 Parsimony network among round goby Neogobius melanostomus cytochrome b haplotypes..48 Fig. 2.5 Bayesian STRUCTURE analysis of round goby populations using combined data from seven microsatellite loci and cytochrome b sequences.49 Fig. Appendix 2.1 Statistical parsimony network with nested clade analysis..50 Fig. 3.1 Current distribution of the round goby in North America and Eurasia showing sampling sites 83 Fig. 3.2 Statistical parsimony network showing relationships among the 88 round goby cytochrome b haplotypes using TCS 84 Fig. 3.3 Bayesian STRUCTURE v2.1 analysis of round goby populations 85 Fig. 3.4 Three-dimensional factorial correspondence analysis of round goby microsatellite data...86 Fig. 3.S1 Neighbor-joining tree of cytochrome b haplotypes and their locations 90 Fig. 3.S2 Neighbor-joining trees of round goby population samples..91 xii

13 Fig. 4.1 Maps of current range and sampling locations for zebra mussels and quagga mussels from North America and Eurasia Fig. 4.2 Three dimensional factorial component analysis for zebra mussels and quagga mussels Fig. 4.3 Genetic distance tree among zebra mussel and quagga mussel population sites based on microsatellite data 135 Fig. 4.4 Bayesian STRUCTURE analysis for zebra mussels and quagga mussels.137 xiii

14 Preface The chapters of this dissertation are organized in chronological order, based on the completion of the individual studies. Chapter 2 has been previously published as: Brown JE, Stepien CA (2008) Ancient divisions, recent expansions: Phylogeography and population genetics of the round goby Apollonia melanostoma. Molecular Ecology, 17, Chapter 3 has been previously published as: Brown JE, Stepien CA (2009) Invasion genetics of the Eurasian round goby in North America: tracing sources and spread patterns. Molecular Ecology, 18, Chapter 4 is currently in review for publication in Biological Invasions: Brown JE, Stepien CA. Population Genetic History of the Dreissenid Mussel Invasion: Expansion patterns Across North America. Each chapter is identical to their published/submitted version, except that the scientific name for the round goby has been changed to Neogobius melanostomus along with those of other Ponto-Caspian gobies to bring them into line with current nomenclature (Neilson & Stepien 2009a). xiv

15 Chapter 1 Introduction Introduced species are a primary environmental concern due to their growing impact on ecosystems worldwide (Mack et al. 2000, Sakai et al. 2001). These organisms are transported via human activities and vectors intentional and accidental and often exert major ecological changes. Once in a new community, introduced species may act as predators, competitors, or pathogens; and at the ecosystem level they may significantly alter processes such as nutrient cycling and habitat structuring (Elton 1958, Williamson 1999, Simberloff 2005). Because of these impacts, we need to understand the pathways that facilitate the establishment and spread of invasive species. One way to explore these pathways is by comparing the population genetics of multiple introduced species from congruent donor and recipient areas. Comparative approaches to population genetics and phylogeography have expanded our understanding of a variety of historical events such as vicariance during glaciation, range expansion and long distance colonization (Riginos 2005, Schonswetter et al. 2006). These methods rely on the identification of congruent genetic signals among species, which are indicative of shared past demographic events (Avise et al. 1987, Avise 2000, Lapointe & Rissler 2005). These techniques are also applicable to more recent population events, such as species introductions. The introductions of the round goby (Neogobius melanostomus), the zebra mussel (Dreissena polymorpha), and the quagga mussel (Dreissena rostriformis 1

16 bugensis) into the Laurentian Great Lakes from the Ponto-Caspian region of Eurasia are well suited to a comparative analysis all three species have overlapping ranges in both Eurasia and North America. The Study Locations The Laurentian Great Lakes are a series of inland, freshwater seas in North America that formed after the last glaciation ended about 13,000 years ago (Larson & Schaetzl 2001). Following European colonization of the continent, the Lakes experienced many serious ecological changes. Of particular relevance are the 182 species introduced into the Lakes, making them among the most invaded freshwater systems in the world (Ricciardi 2006). The Great Lakes are commercially important waterways, and 73% of recent Great Lakes invasions are the result of organisms transported by ship ballast (Holeck et al. 2004, Ricciardi 2006). In recent years, 70% of these introduced species have originated from the Ponto-Caspian region of Eurasia (Ricciardi & MacIsaac 2000). In contrast to the recently formed Great Lakes, the Ponto-Caspian region is geologically old, having its aquatic origins in the Tethys Sea 200 million years ago (Zaitsev & Mamaev 1997). During this time span, the region experienced numerous climatic shifts that are believed to have selected for species with flexible requirements and good colonization ability (Reid & Orlova 2002), as demonstrated by their relative successes in both Europe and North America (Ricciardi & MacIsaac 2000, Bij de Vaate et al. 2002). One of the major events geological events was the separation of the Black and Caspian Seas, which vicariantly split the biota of the region into potentially distinct evolutionary 2

17 units about 5 million years ago, followed by occasional reconnections (Briggs 1974). This geographic separation has led to differences in the biota, with subspecies of N. melanostomus in the Black and Caspian Sea (Pinchuk et al. 2003, Brown & Stepien 2008), and a split between D. rostriformis bugensis and D. r. grimmi in the Black and Caspian Seas, respectively (Stepien et al. 2003). The Study Organisms This study explores the interplay of invasion ecology and population genetics using three introduced Ponto-Caspian species now found in the Laurentian Great Lakes: the round goby and the zebra and quagga mussels, building on past studies by the Great Lakes Genetics Laboratory (Dillon & Stepien 2001, Stepien et al. 1999, 2002, 2003, 2005, Stepien & Tumeo 2006). Along with other Ponto-Caspian species, the round goby, the zebra mussel, and the quagga mussel have colonized North America over the last two decades (Herbert et al. 1989, Jude et al. 1992, May & Marsden 1992). All three species have overlapping distributions in their native and introduced ranges. Previous work on the round goby found evidence that the introduction consisted of a large number of propagules (Dillon & Stepien 2001, Stepien et al. 2005). Studies of dreissenid mussels using both mitochondrial cytochrome b sequences and nuclear RAPDs markers likewise suggested large numbers of propagules were introduced for both dreissenid species (Stepien et al. 1999, 2002, 2003). Invasion History 3

18 Zebra mussels, quagga mussels and round gobies have all become established in North America in the last 25 years. Zebra mussels were the first of the three species detected in the Great Lakes, successfully establishing themselves by 1988 in Lake St. Clair (Hebert et al. 1989) and Lake Erie (Carlton 2008) after introductions via ships' ballast water. This bivalve is widespread throughout Western Europe and the drainage basins of the Baltic, Black, and Caspian Seas (Rosenberg & Ludyanskiy 1994), and has expanded throughout the Great Lakes and Mississippi drainages, colonizing rivers and inland lakes, with the possibility of further expansion into some uninvaded regions (Drake & Bossenbroek 2004). The quagga mussel was discovered in the Erie Canal and Lake Ontario in 1989 (May & Marsden 1992, Mills et al. 1996), and is now the most common dreissenid in Lakes Erie and Ontario. The quagga mussel has been replacing zebra mussels in the St. Lawrence River (Ricciardi & Whoriskey 2004; Jones & Ricciardi 2005; Whittier et al. 2008) and the lower Great Lakes (Mills et al. 1999; Wilson et al. 2006; Whittier et al. 2008), and is now doing the same in the upper Lakes (Nalepa et al. 2009). Collectively, the mussels have altered the benthic substrate and community composition (Stewart et al. 1998) and provide protective structure from predators (Stewart et al. 1999, Mayer et al. 2001). The round goby, a predator on both mussel species, was found in the St. Clair River, near Sarnia, Ontario in 1990 (Jude et al. 1992). The goby introduction has also been attributed to ballast water (Mills et al. 1993). An abundant dreissenid prey base likely aided the round goby as it spread throughout the Great Lakes (Jude 1997). Round gobies form a food resource for yellow perch, walleye, smallmouth bass and other predatory 4

19 fishes (Jude 1997, Belanger & Corkum 2003). They also compete with and displace some native benthic fishes, such as the deepwater sculpin, and consume the eggs of walleye, smallmouth bass, and other desirable species (Jude 1997), and have achieved very high abundance in areas that they have invaded (Johnson et al. 2005). Comparing multiple species that have been introduced to the same region from the same donor region may help to illuminate patterns that are common to successful invaders, and highlight processes that uniquely influence each species (Lockwood & Gilroy 2004; Rabitsch & Essl 2006). These three species were introduced from Eurasia to the Great Lakes, and share biogeographic and ecological histories (Reid & Orlova 2002). Biogeographic factors that likely shaped their population genetic histories include the isolation of the Black and Caspian Basins from the Mediterranean Basins (~200 mya, Zaitsev & Mamaev 1997); changes in water salinity, level, and temperature; and later separation of the Black and Caspian basins (~5 mya, Briggs 1974). In addition, each major Ponto-Caspian river system has its own history, which may reflect vicariant divergence of populations, producing distinct lineages or connectivity through flooding and anthropogenic linkage (Avise 2000). A final factor that unites all three species is that they have been transported to the Laurentian Great Lakes at roughly the same time and possibly by the same vectors and along the same pathways, and continue to expand into the rest of North America. Objectives This dissertation addressed the following questions for each species: 5

20 1) What is the population genetic structure of each species in its native and introduced ranges? 2) Were the North American introductions founded by single or multiple source populations? 3) Are the introduced populations of each species genetically homogenous/heterogenous across their range? 4) Do the genetic diversity levels of the introduced populations of each species reflect founder events? 5) Do populations outside the Great lakes have reduced/similar/greater genetic diversity compared to those within the Great Lakes? 6

21 Chapter 2 Ancient divisions, recent expansions: Phylogeography and population genetics of the round goby Neogobius melanostomus Previously published as: Brown JE, Stepien CA (2008) Ancient divisions, recent expansions: Phylogeography and population genetics of the round goby Apollonia melanostoma. Molecular Ecology, 17, Abstract During the past two decades, the round goby Neogobius melanostomus expanded its range via shipping transport and canals, extending north and west from the Ponto- Caspian region of Eurasia and to the North American Great Lakes. Exotic populations of the round goby have been very successful in the Baltic Sea and the Great Lakes regions, exerting significant ecological changes. Our study evaluates the population genetic and biogeographic structure of the round goby across its native and nonindigenous ranges, in light of geological history and its expansion pathways. We analyze seven new nuclear microsatellite loci and mitochondrial DNA cytochrome b gene sequences from 432 individuals in 22 locations. Population structure is tested using F ST -analogs, phylogenetic trees, clustering diagrams, Bayesian assignment tests, and nested clade 7

22 analyses. Results show that native populations in the Black Sea versus the Caspian Sea basins diverge by 1.4% and ~350,000 years, corresponding to closure of their prior connections and supporting the taxonomic separation of the Black Sea N. m. melanostomus from the Caspian Sea N. m. affinis. Their within-basin populations diverge by ~0.4% and 100,000 years. Nonindigenous populations in the Baltic Sea and Danube and Dnieper Rivers trace to separate northern Black Sea origins, whereas the upper Volga River system houses mixed populations of N. m. melanostoma and N. m. affinis. Native populations average twice the genetic diversity of most exotic sites; however, sites in the Volga River system have high diversity due to mixing of the two taxa. Our results highlight how vicariance and anthropogenic disturbances have shaped a rapidly expanding species genetic heritage. Introduction Aquatic species introductions have increased in the past two decades due to canal construction, shipping, and ballast water transport. Newly introduced species face many challenges that may thwart their establishment - including novel habitat, food, competitors, predators, and pathogens - as well as potential reproductive and recruitment difficulty. Knowledge of a nonindigenous species phylogeographic history may aid prediction of its relative success in novel habitats. For example, it has been hypothesized that the post-pleistocene expansion of taxa into new aquatic habitats was ecologically similar to their present-day movements via anthropogenic transport pathways and global climate change (Hewitt 2004). Until recently the Ponto-Caspian region (Black, Azov, & Caspian Seas) of Eurasia housed an endemic fauna, such as the neogobiin flock of ~20 8

23 fish species (Stepien and Tumeo 2006), Dreissena mussels (Stepien et al. 2002, 2003, Therriault et al. 2005), and a variety of crustaceans (Cristescu et al 2003, Audzijonyte et al. 2006). Due to human activities many of these Ponto-Caspian endemics, including neogobiin fishes, have spread beyond the region (Bij de Vaate et al. 2002, Stepien et al. 2005, Ricciardi 2006). The round goby Neogobius melanostomus is the most widespread nonindigenous neogobiin fish (Neogobiinae: Gobiidae: Teleostei). It was originally distributed in the Black, Azov, and Caspian Seas and the lower reaches of their tributaries. The round goby recently has spread up the great rivers of the Ponto-Caspian region and Eastern Europe (see Fig. 2.1), as well as to the Baltic Sea and the North American Great Lakes (both ~1990; Jude et al. 1992, Skora & Stolarski 1993). Ballast water is a primary means of their anthropogenic transport, through the uptake of pelagic juveniles that nocturnally rise to feed in the water column (Hensler & Jude 2007). This appears to be the mechanism that led to the Baltic Sea and the North American colonizations, and likely that of the upper Volga River. Expansion populations likely migrated more slowly along canals. The round goby inhabits a gradient of marine to freshwater habitats, reaching its largest sizes in the seas (~25 cm total length; Svetovidov 1964, C.A. Stepien, pers. obs.). Most individuals appear to move little in terms of geography throughout their lives, except for some seasonal migration offshore during the winter (Kostyuchenko 1969, Ray & Corkum 2001). Males guard nests under rocks or other objects and are highly territorial 9

24 (MacInnis & Corkum 2000). They eat molluscs (including dreissenids in freshwaters and mytilids in marine waters), other invertebrates, and fishes (Pinchuk et al. 2003). Our study analyzes population genetic divergence of the round goby across its native and nonindigenous Eurasian ranges, employing a dual approach of mtdna sequencing and seven new nuclear microsatellite loci. This investigation includes re-analysis of samples used by prior round goby genetic studies of mitochondrial DNA genes by Dougherty et al. (1996), Dillon & Stepien (2001), Stepien et al. (2005), and Stepien & Tumeo (2006), which totaled 59 samples from six Eurasian sites. We here provide the first analysis of microsatellite loci, and extend the previous work to include the entire cytochrome (cyt) b gene and add 16 additional Eurasian sites, for a total of 22 sites and 432 individuals. Our approach allows both phylogeographic history of lineages to be assessed (through maternally inherited mtdna) and fine- and broad-scale population genetic relationships to be discerned (comparing biparentally-inherited microsatellites with mtdna patterns). We test for historic vicariant separation patterns of the round goby among native population sites, as well as donor-recipient population relationships for exotic locations in the Black/Azov Sea and Caspian Sea basins. Pathways for round goby expansion and possible jump-transport are examined among the great rivers of the Ponto-Caspian basin, including the Danube, Dniester, Dnieper, Southern Bug, Don, and Volga River watersheds, as well as the Baltic Sea. Biogeographic history of the Ponto-Caspian region 10

25 The Ponto-Caspian region houses an endemic fauna whose history has been dominated by the intermittent union and division of the major basins in response to tectonic events and climatic shifts (Dumont 1998). It is likely that this pattern of change selected for species with flexible niche requirements and good colonization ability, as demonstrated by their historic and recent colonization successes, including post-glacial and recent Eurasian population expansions and establishment in North America (~1990; see Ricciardi & MacIsaac 2000, Bij de Vaate et al. 2002). In addition to its brackish seas, the Ponto-Caspian region drains several major river systems, with the Danube, Dniester, and Dnieper Rivers emptying into the Black Sea, the Don River draining into the Sea of Azov, and the Volga River flowing to the Caspian Sea. The Black and Caspian Sea basins separated ~5 mya, with periodic reconnections to each other through the Sea of Azov and the Manych Depression and to the Mediterranean Sea due to tectonic activity and climate change (Briggs 1974, Reid & Orlova 2002). Pleistocene glacial cycles significantly altered the climate and hydrogeography of the Ponto-Caspian Region (Hewitt 2004, Mangerud et al. 2004, Bahr et al 2005). Water levels in the Black, Azov, and Caspian Seas decreased during dry glacial periods, restricting aquatic fauna to portions of their former distributions (Zaitsev & Mamaev 1997, Dumont 1998, Major et al. 2006) that served as refugia for once widely distributed taxa, including fishes (Griffiths 2006, Reyjol et al. 2007) such as the spined loach Cobitis taenia (Culling et al. 2006), bullhead Cottus gobio (Englbrecht et al. 2000, Hanfling et al. 2002), brown trout Salmo trutta (Bernatchez 2001), chub Leuciscus cephalus (Durand 11

26 et al. 1999), dace Leuciscus leuciscus (Costedoat et al. 2006), and Eurasian perch Perca fluviatilis (Nesbø et al. 1999). Interglacial increases in water levels opened recolonization pathways into western and northern Eurasia via tributaries (Hewitt 2004) and restored connections between the Black/Azov and Caspian Seas at several times during the Pleistocene Epoch (Dumont 1998). Major faunal exchanges accompanied these reconnections three times during the Pleistocene ( mya, mya, and mya; Reid & Orlova 2002). Changing connections also resulted in salinity fluctuations from fresh to brackish to marine conditions, depending on connections to the Mediterranean Sea and freshwater inputs from rivers (Waters et al. 1994). This cycle of vicariance and reconnection resulted in many genetically divergent Black and Caspian Sea lineages, including copepod and amphipod taxa (Cristescu et al. 2003), mysid shrimp (Audzijonyte et al. 2006), dreissenid mussels (Stepien et al. 2003; Gelembiuk et al. 2006), and sturgeon (Choudhury & Dick 1998). A similar pattern may exist in round gobies, as two currently unrecognized subspecies of the round goby were once proposed: N. m. melanostoma in the Black Sea and N. m. affinis in the Caspian Sea (Navozov 1912, Pinchuk et al. 2003). As a result of their distinct though related geological histories, genetic signatures of Black and Caspian Sea fauna may reveal divergent or congruent evolutionary patterns. Moreover, the natural separation of the Black and Caspian Sea drainages has been breached anthropogenically by several canals, of which the largest is the 1952 Volga-Don 12

27 Canal (Zaitsev & Mamaev 1997). These canals created artificial pathways that allow taxa to move between systems, leading to gene flow. Questions Our objective is to quantify the genetic diversity and divergence patterns across the native and introduced Eurasian distribution of the round goby (Fig. 1) to test the following: 1. Are populations in the Black and Caspian Sea basins genetically distinct? 2. Do populations in the Black and Caspian Seas differ in overall genetic diversity levels, and do areas of their expansion show founder effects? 3. Do overall biogeographic divergence patterns correspond to historic glacial refugia and show congruence with other taxa? 4. Do the gene pools of populations in freshwater and marine systems differ? 5. Did populations in the Volga-Don Canal and the northern Volga River system originate from the Sea of Azov or the Caspian Sea? 6. What was the likely founding population source for the nonindigenous population in the Gulf of Gdansk, Baltic Sea? Materials and Methods Sampling, DNA extraction, and genetic data collection We sampled 22 locations from the round goby s native and introduced Eurasian ranges (Table 2.1, Fig. 2.1), including 5-30 individuals per site and totaling 432 individuals. To 13

28 maximize phylogeographic information, we sampled as many locations as possible (see Pons & Petit 1995; Culling et al. 2006), focusing on ports, areas of range expansion, and major watersheds. Unfortunately, we were unable to sample every introduced location, especially along the Volga River in Russia. Sampling locations were defined as nonnative or native based on the reports in the literature and the range presented by Pinchuk et al. (2003) and noted in Fig. 2.1 and Table 2.1. As both the Black and Caspian Seas are brackish, but their tributaries are freshwater, we classified locations with salinity greater than 0 ppt (Table 2.1) as marine for purposes of comparison. These locations were along the sea coast, as opposed to being upstream in a river. Round goby samples were placed directly in 95% ethanol in labeled vials and stored at room temperature. The same individuals were analyzed for mitochondrial and microsatellite loci, facilitating comparison. We compared variation in the round goby with its sister species, the monkey goby N. fluviatilis; and with other neogobiin relatives, including the freshwater tubenose goby Proterorhinus semilunaris, the bighead goby Ponticola kessleri, and the toad goby Mesogobius batrachocephalus (see Stepien & Tumeo 2006). Genomic DNA was extracted from 25 mg fin tissue using Qiagen DNeasy kits (Qiagen, Inc.; Valencia, CA), eluted in 200 µl of water, stored at 4 o C until used for PCR amplification, and archived at -80 o C. The entire mitochondrial cytochrome b (cyt b) gene (1138 bp) and part of the following trna-thr (66 bp) were amplified using the primers L14724, L15066, and H5 (Table 2.2A) in 25 L reactions containing 1 unit Taq polymerase, 200 M dntps, 50 mm KCl, 1.5 mm MgCl 2, 10 mm Tris-HCl, 0.5 M of primers L14724 and H5 (Table 2.2A), and 30 ng of template. Amplifications were 14

29 performed on a MJR DYAD thermalcycler (Bio-Rad Laboratories, Hercules, CA), with initial denaturation at 94 o C for 120 sec, followed by 35 cycles of denaturation (94 o C for 45 sec), annealing (52 o C for 30 sec), and extension (72 o C for 60 sec), plus a final 72 o C extension for 180 sec. Sequencing using the same primers was outsourced to the Cornell University Life Sciences Core Laboratories Center ( Ithaca, NY). Sequences then were aligned with CLUSTAL X v1.8 (Thompson et al. 1994) and adjusted using BIOEDIT v7.0 (Hall 1999, 2004) in our laboratory. We analyzed variation at seven nuclear microsatellite loci developed for the round goby by Kevin Feldheim (Field Museum of Natural History) with us (Feldheim et al. in preparation, Table 2.2B). Amplification used 10 L reactions of 0.6 units Taq, 50 M nucleotides, 50 mm KCl, 1.5 mm MgCl 2, 10 mm Tris-HCl, 0.5 M of each primer (Table 2.2B), and 30 ng of template, with a sterile mineral oil overlay to maintain the reaction volume. A thermal cycle of 2 min at 94 o C for initial denaturation was followed by 35 cycles of denaturation (94 o C, 30 sec), annealing (1:00 min) at a primer-specific temperature (Table 2.2B), and extension (72 o C, 30 sec); followed by a 5 min final extension at 72 o C. Amplification products were diluted 1:50 in water, of which 1µl was added to 13 µl of formamide and ABI (Applied Biosystems Inc., Fullerton, CA) Gene Scan 500 size standard and analyzed on a 96-well plate with an ABI 3130 Genetic Analyzer and GENEMAPPER v3.7 in our laboratory. Output profiles were checked to confirm allelic size variants and representative alleles were sequenced to verify that length polymorphisms were due to variation in copy number of single repeat motifs. 15

30 ARLEQUIN (v3.01; Excoffier et al. 2005) was used to assign individuals to the correct mt haplotype, which were deposited in NIH GenBank as accession numbers EU EU ( We identified codon positions and transitional and transversional substitutions. Data analysis Allelic frequencies, number of private alleles, conformance to Hardy-Weinberg (HW) equilibrium expectations (for microsatellite data), and linkage disequilibrium (for microsatellites) were evaluated in GENEPOP v3.4 (Raymond & Rousset 1995, 2004). Levels of significance for HW and linkage disequilibrium tests were adjusted using nonsequential Bonferroni correction (Sokal & Rohlf 1995). HW deviations were tested for heterozygosity deficiency or excess, and for the presence of null alleles with MICROCHECKER v2.23 ( van Oosterhout et al. 2004, 2006). Genetic composition among samples were analyzed to identify true populations (i.e., those distinguished by significantly divergent gene pools) using a pairwise F-statistic analog (θ ST ; Weir & Cockerham 1984) and contingency tests (Raymond & Rousset 1995, Goudet et al. 1996) for both mt and microsatellite data. Relationships between recently diverged samples, such as those tested here, have been shown to be better resolved in models using contingency tests (see Balloux & Lugon-Moulin 2002), which are independent from Hardy-Weinberg equilibrium assumptions, non-parametric, and little 16

31 affected by small sample size (Raymond & Rousset 1995, Goudet et al. 1996). Our additional use of an F-statistic analog facilitated direct comparisons with other studies (see Stepien et al. 2007). For the cyt b data, we also calculated mean uncorrected pairwise p-distances in MEGA v3.1 (Kumar et al. 2004) among population groups. We then used the genetic distance values to evaluate possible divergence times among populations, based on a 2% divergence per million years calibration for cyt b divergence in the goby Evorthodus across the Isthmus of Panama (Rocha et al. 2005). In order to further analyze the relationships among population sites, pairwise genetic distances were calculated using Cavalli-Sforza chord distances (Cavalli-Sforza & Edwards 1967) for the combined cyt b and microsatellite data in PHYLIP (Felsenstein 1989). Neighbor joining (NJ) trees (Saitou & Nei 1987) were constructed from the chord distances using PHYLIP. We tested correspondence between chord genetic distance (for the combined cyt b and sat loci) and geographic distance, measured as nearest connected waterway paths, using the Mantel (1967) approach (10,000 rearrangements) with ISOLDE in GENEPOP. The fit of the regression line was calculated using Microsoft Excel 2003 (Microsoft, Redmond, WA). Analysis of MOlecular VAriance (AMOVA; Excoffier et al. 1992) in ARLEQUIN tested for hierarchical population structure by partitioning the total θ covariance among geographic groups (Excoffier et al. 1992), including the Caspian Sea vs. Black Sea basins, major river systems, and marine versus freshwater locations. 17

32 NJ trees from the mt haplotype sequence data were produced using MEGA. Their phylogeographic patterns were further analyzed using Nested Clade Analysis (NCA; Templeton et al. 1995) in GEODIS v2.5 (Posada et al. 2000) to test the null hypothesis of no association among haplotypes and geographic location (Templeton 1998) using a statistical parsimony haplotype network generated with TCS (Clement et al. 2000). Ambiguous loops in the network were resolved according to Pfenninger & Posada (2002). Association between the haplotype network and geography was tested with 10,000 permutations. The GEODIS inference key (Templeton 1998; updated by Posada 2005) was used to evaluate the likely cause(s) of associations; such as isolation by distance, range and population expansion/contraction, long distance dispersal, fragmentation, and demographic connectivity and shifts. Both F-statistics and contingency tests use the sample location as the unit of comparison, whereas the Bayesian model-based methods of Rannala & Mountain (1997) in GENECLASS2 v2.0 (Piry et al. 2004) and STRUCTURE v2.1 (Pritchard et al. 2000, Pritchard & Wen 2004) use the individual as the unit, assigning it to the most likely group (population) regardless of geographic origin. This makes these methods particularly useful for assessing likely source populations (Abdelkrim et al. 2007, D Amato et al. 2007, Schrey et al. 2007). GENECLASS2 tests were run with simulated population sizes of 10,000 individuals for each sampling site and a 0.01 rejection level (Cornuet et al. 1999). An exclusion test was performed to check for false positives. STRUCTURE assigned individuals to groups ranging from K = 1 to K = total N sites, with the relative frequency of predicted group memberships totaling Ten 18

33 independent runs for each K were used, with burn-ins of 100,000 replicates and run lengths of 500,000 replicates. In order to further test the divisions within the microsatellite data, they were subjected to a three-dimensional-factorial correspondence analysis (3D-FCA; Benzecri 1973) using GENETIX v4.05 (Belkhir et al. 2004). That approach has no a priori assumptions about populations and shows both within and among population variation. Results Overall genetic variability and patterns We sampled 432 individuals and recovered 81 cyt b haplotypes (Table 2.1; GenBank accession numbers EU EU331236), with 17 (21%) characterizing multiple individuals and 64 (79%) singletons. There were 220 microsatellite ( sat) alleles recovered from seven loci, with 40 (18%) found exclusively in single population locations and 20 of those (9% overall) characterizing single individuals. All loci were in Hardy-Weinberg equilibrium following Bonferroni correction and were independent, and no null alleles were detected by the MICROCHECKER analysis. Most (76%) cyt b haplotype substitutions occurred at the third codon position, and there were 21 transversions and 83 transitions (1:4) and no indels (Table 2.1). Numbers of haplotypes per location ranged from 0 to 11 (Lisar, Iran of the Caspian Sea), and averaged 5.3 ± 0.7. Haplotype (gene) diversity ranged from 0 to (the latter in two sites in the Sea of Azov), averaging 0.49 ± Number of sat alleles per locus ranged from 15 (Ame10) to 51 (Ame129; Table 2.2B), with a mean of 32. Observed sat 19

34 heterozygosity per locus ranged from (Ame1) to (Ame23), and averaged Total heterozygosity per sampling site ranged from (the Dnieper River site at Kakhovka, Ukraine) to (Sevastopol, Ukraine on the Black Sea) and averaged 0.46 ± 0.01 (Table 2.1). ST values per sat locus ranged from (Ame23) to (Ame194), and averaged (Table 2.2B). Pairwise ST divergences among sampling sites were about two times greater for cyt b data (ranging from to 1.000; mean 0.277) than for the sat loci (range to 0.305; mean 0.126; Table 2.4). Black/Azov Sea vs. Caspian Sea basins The round goby diverged from its congener the monkey goby Neogobius fluviatilis by an average uncorrected pairwise p-distance of 0.12 and an ~3 my divergence using a molecular clock calibration of ~2%/my for cyt b evolution in gobies. Round goby haplotypes were divided into two reciprocally monophyletic clades in the Black and Caspian Sea basins (Fig. 2.2), separated by eight fixed nucleotide differences and an uncorrected p-distance of 0.014, equivalent to ~350,000 y divergence. This division additionally was supported by ST = (cyt b) and ST = ( sat) (AMOVA; Table 2.3A), assignment test distinctions (Table 2.6), and 3D-FCA clustering results. Nested clade analysis (NCA) inferred that this division resulted from allopatric fragmentation (clades 4-1 and 4-2, Table 2.6, Appendix 2.1). In contrast, the majority of lower-level clades were defined by restricted gene flow with isolation by distance or by contiguous range expansion (Table 2.6). 20

35 No round goby haplotypes were shared between the Black and Caspian Sea basins, which additionally housed 83 (38%) and 19 (9%) sat private alleles, respectively. Both basins housed similar numbers of haplotypes and private alleles, and had similar levels of cyt b and sat gene diversity (Table 2.3B). Within each system, marine locations had more private alleles, including the Kerch Strait (7 cyt b, 2 sat), Sea of Azov (7 cyt b, 1 sat), and the southern basin of the Caspian Sea (in Iran; 7 cyt b, 8 sat). Mantel test results based on combined data from cyt b and sat loci revealed significant genetic isolation with geographic distance within each basin (p < for each; Black Sea y = x , R 2 = 0.200; Caspian Sea y = x ; R 2 = 0.232; figures not shown). Assignment tests of combined sat and cyt b data using STRUCTURE Bayesian analysis identified 15 round goby population groups (posterior probability = 0.999; Table 2.7, Fig. 2.5). Most sampling locations assigned to one or two primary groups. Individuals from nonindigenous locations had the greatest assignment levels to single population groups, including the Danube River samples at Slovakia and Serbia, the upper Dnieper River at Kiev, and the Volga-Don Canal, along with native Sea of Azov samples. GENECLASS2 results were congruent with those from STRUCTURE. The 3D-FCA revealed four clusters that were separated by three axes (explaining a total of 43% of the variation; figure not shown). Axis 1 explained 17.95% of the data, and separated the Dnieper River, Danube River, and Odessa populations from all others. Axis 2 explained 9.43% of the data and split the Dnieper River samples from a cluster containing the Danube River and Odessa samples. Axis 3 explained 8.79% of the data, 21

36 and separated the Baltic Sea and marine Black Sea samples from those in the Sea of Azov, Volga River, and Caspian Sea. Patterns within the Black/Azov Sea basins Pairwise ST estimates between sampling locations in the Black/Azov Sea basin ranged from (cyt b) and ( sat) (Tables 2.4A), indicating additional population genetic structure. A three-way division separating the Black Sea freshwater, Black Sea marine, and Sea of Azov population groups was supported by AMOVA (Table 2.3A) and pairwise comparisons (Tables 2.4A, 2.5A). These groups diverged by average p-distances of 0.004, corresponding to ~100,000 years (Table 2.3A). Freshwater locations in the Black Sea had fewer haplotypes and private haplotypes, and lower gene diversities than did the two marine groups (Table 2.3B). The 3D-FCA and the chord distance neighbor-joining tree (Fig. 2.3) grouped the Sea of Azov samples as evolutionarily closer to those from the Volga River and the Caspian Sea. Five population divisions within the Black/Azov Sea basin were designated by STRUCTURE assignments (Fig. 2.5). Samples from the western Black Sea, the Crimean Peninsula, the Dnieper River, and the Sea of Azov formed distinctive groups. Nonindigenous sites in the Danube River assigned with some samples from the Black Sea at Odessa. Within the Dnieper River, the nonindigenous round goby population from Kiev appeared distinct, although related to those further downstream. 22

37 Patterns within the Caspian Sea basin Within the Caspian Sea basin, samples from Iran had the most cyt b haplotypes, more private haplotypes, and higher haplotype diversity (Table 2.3B). Pairwise ST comparisons of Caspian Sea basin samples ranged from (cyt b) and ( sat). Division into Volga River and Caspian Sea population groups was supported by AMOVA (Table 2.3A), pairwise comparisons (Tables 2.4B, 2.5B), the Cavalli-Sforza chord distance tree (Fig. 2.3), and STRUCTURE assignment tests (Fig. 2.5). Further division of the Caspian Sea basin round goby samples into three population groups in Volga River, Azerbaijani Caspian Sea, and the Iranian Caspian Sea was indicated by STRUCTURE assignment tests (Fig. 2.5) and AMOVA (Table 2.3A), corresponding to p = (Table 2.3A) and ~100,000 years. Non-native locations Mean gene diversity was greater in native round goby populations than in nonindigenous locations (Table 2.1, 2.3B; cyt b: ± native vs ± nonindigenous; sat: ± native vs ± nonindigenous). However, some exotic population sites were unusually diverse for cyt b, including the upper Dnieper River (0.539) and the Moskva River (of the Volga system; 0.857) (see Table 2.1). Genetic diversity among native populations was highest in the marine sites at the Kerch Strait and the Sea of Azov (cyt b: 0.978; sat: and cyt b: 0.978; sat: 0.586). The exotic Baltic Sea population site 23

38 The nonindigenous Baltic Sea population site had relatively low gene diversity (cyt b: 1 haplotype and h = 0.000, sat: h = 0.379; Table 1), and clustered with samples from the Black Sea basin (Figs. 2.3 & 2.4). Pairwise tests based on cyt b showed that the Baltic Sea sample significantly differed from those in the Crimean Peninsula (Black Sea) and the Sea of Azov (Table 2.4A, 2.5A). However, the Baltic Sea sample was statistically indistinguishable from other locations in the remainder of the Black Sea and its rivers using cyt b. Pairwise ST (Table 2.4A) and contingency tests (Table 2.5A) using sat data showed significant divergence of the Baltic Sea sample from all other locations. Assignment tests also discerned the Baltic Sea samples as different from all others, which precluded identification of its source (Fig. 2.5). Black Sea non-native locations The two non-native round goby samples from the Danube River in Slovakia and Serbia had low mitochondrial gene diversity (0.051 and 0.044, respectively). However, their sat gene diversities (0.377 and 0.459, respectively) were comparable to mean values for native locations. Nonindigenous sites in the Danube River assigned with each other and with samples from the Black Sea at Odessa in the STRUCTURE (Fig. 2.6) analysis and in the 3D-FCA. The non-native sample in the Dnieper River at Kiev had very high gene diversity (cyt b: 0.539, sat: 0.369) and a private cyt b allele. Both STRUCTURE (Fig. 2.6) and the 3D-FCA grouped it with native locations further downstream, while also highlighting its distinctiveness. 24

39 Caspian Sea non-native locations Of the three exotic sites in the Caspian Sea basin, the Volga-Don Canal sample had very low gene diversity and the introduced location from the Moskva River had very high gene diversity (cyt b: 0.857, sat: 0.425) - higher than almost all of the native locations sampled. The sample from the central Volga River near Svetli Yar also housed high genetic variability (Table 2.1). Introduced round gobies in the Volga-Don Canal, the Volga River, and the Moskva River appeared related to native population locations nearer the mouth of the Volga River. However, NCA additionally revealed a lineage of cyt b haplotypes that was shared between samples from the Moskva River and the Volga River at Svetli Yar, and appeared related to those from the Black/Azov Sea population group (clade 3-6, Appendix 2.1). There thus are two disparate Volga River lineages - one descendent from the Caspian Sea and one similar to the Black/Azov Sea group. NCA was unable to discern whether they originated from contiguous range expansion, long distance colonization, or past fragmentation (Table 2.6). They likely reflect anthropogenic activities, such as long distance transport and the breaching of watershed boundaries with canals. Discussion Ancient Divisions Phylogeography of the Eurasian round goby revealed two major lineages within its native range, corresponding to the two primary Ponto-Caspian basins - the Black/Azov Sea and the Caspian Sea - which likely diverged following a faunal exchange between the 25

40 Euxinian (Black) and the Khazar (Caspian) Sea basins ~350,000 ya (Zubakov 1988). A subspecies divergence between N. m. melanostomus in the Black Sea and N. m. affinis in the Caspian Sea was once recognized (Navozov 1912, Berg 1949), with the Caspian Sea subspecies having lower scale counts (Iljin 1938) and a lower craniological index (see Pinchuk et al for a discussion on these traits). Our results show that the two lineages are markedly divergent based on both mitochondrial and nuclear DNA data, distinguished by eight fixed cyt b substitutions, and reciprocally monophyletic (Fig. 2.2). Our study thus supports their taxonomic distinctiveness (Pinchuk et al. 2003), and they should be recognized as subspecies and further evaluated for possible elevation to species. Marked genetic divergences between round goby taxa from the Black/Azov and Caspian Sea basins are congruent with those of other fishes and invertebrates. They diverged in separate glacial refugia in the Black and Caspian Sea basins and then followed separate glacial meltwater pathways along the Danube, Dnieper, and Volga Rivers. Examples of invertebrates include planktonic cladocerans (~1 mya) and the amphipod Pontogammarus crassus (~1-1.6 mya; Cristescu et al. 2003). Fishes include populations of ruffe Gymnocephalus cernuus across Eurasia that showed pronounced mtdna genetic divergence between lineages in the Black and Caspian Sea basins dated as ~500,000 ± 180,000 y (Stepien et al. 1998), comparable to the separation estimated here and attributed to the most extensive cold period during the glaciations (see Rohling et al. 1998). Similarly, Culling et al. (2006) discerned Ponto-Caspian clades of the spined loach Cobitis taenia, with separation between the Black Sea rivers and the Volga River 26

41 system dated to ~500,000 ya. Black Sea and Caspian/Aral Sea clades of the brown trout Salmo trutta were estimated to have diverged more recently (~150, ,000 ya) based on nested clade analysis (Bernatchez 2001). Other studies of divergence between the Black and Caspian Sea basins discerned older patterns, dating to their formation ~5 mya (Dumont 1998). Work in our laboratory on the monkey goby Neogobius fluviatilis (the sister species of the round goby; Stepien & Tumeo 2006) shows a much older division between Black/Azov and Caspian Sea locations, dating to ~3.75 mya (Neilson and Stepien, in progress). The chub Leuciscus cephalus is divided in two Pliocene-dated Ponto-Caspian clades; one in the Danube River system and the other in the Caspian Sea basin (Durand et al. 1999). Shallower divergence was found for zebra mussel populations dating to ~100, ,000 y according to Stepien et al. (2003, 2005) and Gelembiuk et al. (2006), corresponding to the closure of the most recent connection between the Black and Caspian Sea basins (Reid & Orlova 2002). Patterns within the Black/Azov Sea basin - Within the Black/Azov Sea region, there are significant divergences among round goby population groups that correspond to geographic features, notably basins and tributaries. Round goby populations in the Black Sea rivers, the Black Sea proper, and the Sea of Azov significantly differ, dating to ~100,000 ya. Their resolution is shallow, and may have resulted from expansions into river systems. The Crimean Peninsula in the Black Sea houses round goby lineages from both the Black Sea and the Sea of Azov, reflecting its intermediate geographic location. 27

42 Patterns within the Caspian Sea basin - We identified three round goby population groups in the Caspian Sea basin - one in the Volga River and two within the Caspian Sea proper, dating to 100,000 y. Despite expansions within the Volga River watershed, no Caspian derived lineages were found in the locations we sampled in the Black and Azov Seas. The two Caspian Sea groups may have differentiated in the central and southern deep basins, which likely served as refuges during low water periods. Round goby populations within the Caspian Sea and Black/Azov Sea basins thus differentiated about the same time. About 100,000 ya the Riss Glaciation lowered Caspian water levels 50 m below sea level and the Black Sea became salinized via contact with the Mediterranean Sea (Reid & Orlova 2002). These events may have isolated the respective taxa into groups within these basins. Additional research is needed to further explore these fine scale patterns. Recent Expansions We sampled four sets of non-native populations in this study: the Baltic Sea, the Slovak and Serbian portions of the Danube River, the upper Dnieper River near Kiev, and portions of the Volga River system. Nonindigenous round goby populations in central Europe appear derived from the Black Sea primary clade, whereas those in the Volga River comprise a genetic admixture from native Volga River locations and the Black/Azov Sea region. 28

43 Microsatellite data similarities suggest that the Dniester and Dnieper Rivers in the northwest Black Sea were likely donors to the Baltic Sea nonindigenous population site, and should be investigated further. This assignment gains additional support from the study of the round goby invasion by Dillon & Stepien (2001) using the mtdna control region, which eliminated the western Black Sea region of Varna, Bulgaria as a likely donor source for the Baltic Sea population site, but did not have samples available from the Dnieper and Dniester Rivers. Round gobies from the Dniester and Dnieper Rivers thus should be sequenced for the control region. Nonindigenous round goby sites in the Danube River at Bratislava and Prahovo had relatively low genetic diversity in our study. Similarly low genetic diversity was found in an expansion population of the racer goby Babka gymnotrachelus from the Bratislava site (Ohayon & Stepien 2007). In contrast, a racer goby sample from another nonindigenous Danube River location located southward at Tekija, Serbia had much higher diversity. Exotic populations of the racer goby thus appeared to lose genetic variation as they progressed northward along the Danube River (Ohayon & Stepien 2007). The nonindigenous round goby population in the upper Dnieper River contrasts with those from the Baltic Sea and the Danube River, in having higher genetic diversity (cyt b: 0.539, sat: 0.369) and a private cyt b allele. The differences observed between this location and others downstream are likely due to founder effects and colonization admixture resulting from extensive shipping activity as the Dnieper River is a commercially important waterway. 29

44 Round goby samples from the Volga-Don Canal had a single cyt b haplotype that also occurred in the central Volga River and the Volga River delta. This was surprising, since the water filling the canal reservoirs originates from the Don River (V. Boldyrev, State Institute of Lake and River Fisheries, Volgograd, Russia, pers. comm.). However, Audzijonyte et al. (2006) found that mysid shrimp populations in the Canal were of Volga River origin, despite stockings from the Don River. Round goby samples from the Moskva River and the central Volga River contained both N. m. melanostomus (the Black Sea basin subspecies that we are resurrecting here) and N. m. affinis (the Caspian Sea basin subspecies), and consequently had high gene diversity. The Black Sea taxon likely reached the Volga River via human activities, whereas the Caspian Sea taxon likely spread upriver from the Volga River delta. The Volga River and its tributaries need to be further sampled to determine whether the taxa hybridize. Summary & Conclusions Our findings support long-term primary genetic division between round goby taxa in the Black/Azov and Caspian Sea basins leading to their differentiation as Neogobius melanostomus melanostomus and N. m. affinis, respectively. These taxa should be recognized as valid subspecies and further evaluated for possible elevation to species. Both have expanded beyond their historic distributions due to human activities, especially during the past two decades (Bij de Vaate et al. 2002, Stepien et al. 2005, Ricciardi 2006). Expansion populations in central and eastern Europe are composed of N. m. 30

45 melanostoma. Both taxa co-occur in the Volga River system, making this region potentially valuable for future evolutionary studies and worthy of more intensive research. Our results highlight how glacial and anthropogenic disturbances have shaped the genetic heritage of a fish that has recently escaped the Ponto-Caspian and rapidly expanded its range on two continents. 31

46 Tables & Figures Table 2.1 Round goby sampling locations, latitude and longitude, typical salinity (ppt), sample size (N), cyt b haplotypes, and genetic diversity values determined using mt cyt b sequence data and seven nuclear microsatellite loci. Bold haplotypes are shared among locations, * = introduced locations, H O = observed microsatellite heterozygosity, H E = expected heterozygosity, and F IS = inbreeding coefficient. Salinity values for locations A-M are from Yuriy Kvach (pers. comm.) and N-V are from Audzijonyte et al. (2006). Water Body Baltic Sea, Gulf of Gdansk Danube River Black Sea Location & Map label A. Gdynia, Poland* B. Bratislava, Slovakia* C. Prahovo, Serbia* D. Varna, Bulgaria E. Bilgorod, Ukraine F. Odessa Ukraine G. Sevastopol, Ukraine Gene sat Diversity Latitude Longitude Salinity N Haplotype(s) Cyt b H O H E F IS (ppt) , , ,10,11,62,65,67,68,77,78, ,2,3,4,5,6,9, ,10,18,48,60,61,62,63,64, ,18,48,65,66,69,74,75,

47 Southern Bug River Dnieper River H. Mikolaev, Ukraine I. Kiev, Ukraine* J. Kakhovka, Ukraine ,18,49,50,51,58,72, ,58, , Kerch Strait Sea of Azov Moskva River Volga- Don Canal Volga River Caspian Sea K. Kherson, Ukraine L. Kerch, Ukraine M. Mariupol, Ukraine N. Moscow, Russia* O. Iliovka, Russia* P. Svetli Yar, Russia* Q. Damchik, Russia R. Nabran, Azerbaijan S. Shikh, Azerbaijan T. Alet, Azerbaijan U. Lenkoran, Azerbaijan ,7,8, ,41,42,43,44,45,46,47, ,12,13,14,15,16,17,18, ,25,26,27, ,24,25,29, ,21, ,33,70, ,32,35,81,82,83, , V. Lisar, Iran ,32,33,34,35,36,37,38,39,40,

48 Table 2.2 Cytochrome b (A) and microsatellite (B) primers used for amplification of round goby DNA, with sequence, annealing temperature (T A ), and publication source. Table 2B contains the range of repeat numbers (R N ) and sizes (R S ; bp), number of alleles (N A ), average observed heterozygosity (H O ), and average F IS and F ST for each microsatellite locus. A Primer Sequence ( ) T A ( o C) Source L14724 GTGACTTGAAAAACCACCGTT 52 Meyer et al H5 GAATTYTRCGTTTGGGAG 52 Akihito et al L15066 TTGGTCGAGGCCTCTATTACG 52 This study B Primer Sequence ( ; (Feldheim et al. in prep) T A ( o C) R N R S (bp) N A H O F IS F ST Ame1F AGAACAGTCTGGAGGACTCTTTG Ame1R GCGCTTTGTGACCATGTCT 58 Ame10F ATGCGAAGCCGATTTCTG Ame10R CCATATGTCAGGCGATATTCC 52 Ame17F GGCGCAACCTCATTTTAATC Ame17R GTTTAGGCGGGGGTTAAGAG 58 Ame23F AAAGCATCAGCAGCAGTTGT Ame23R TATGTGAGTGTGCGGATGGT 58 Ame129F TGCTCGGTCCTACTTCAAGC Ame129R GCATTCACATTCCTCCCACT 56 Ame133F GCCCACCCCTTCACTCTT Ame133R GGCTATGGCATTTTCTCTCC 56 Ame194F AAACACACAGTCACAAGCACA Ame194R CACAGCTAATGGGGATCCTA 52 34

49 Table 2.3 A. Partitioning of genetic variation between round goby population divisions and among sampling sites within them using AMOVA (Excoffier et al. 1992). Measures include ST and % variation between the divisions and among sampling sites within them (for which all p < ), and between group mean uncorrected p-distances for cyt b. The remainder % of the variation is found within the sampling sites. Within group mean uncorrected p-distance was for all Black, Azov, and Caspian Sea groupings. B. Summary statistics for regional comparisons. A. AMOVA ( ST ) Genetic Variation (Cyt b/ sat) Uncorrected p-distance Cyt b sat Between Cyt b divisions Among sites Black/Azov Seas vs. Caspian Sea basins % / 8% 3% / 12% Black Sea freshwater vs. marine vs. Sea of Azov % / 6% 10% / 9% Caspian Sea vs. Volga River system % / 2% 37% / 10%

50 B. Comparison Mean N Cyt b haplotypes N of Private haplotypes Gene Diversity (Cyt b) Ho ( sat) Black/Azov S. Basin 5.9 ± ± ± ± Caspian S. Basin 4.3 ± ± ± ± Black S. Freshwater 2.6 ± ± ± ± Black S. Marine 9.0 ± ± ± ± Azov S. 9.0 ± ± ± ± Iranian Caspian S ± ± Azerbaijan Caspian S. 3.5 ± ± ± ± Volga R. 3.5 ± ± ± ± Native 6.5 ± ± ± ± Non-Native 2.7 ± ± ± ±

51 Table 2.4 Pairwise divergences between round goby population samples using θ ST for the Black Sea basin (A) and the Caspian Sea basin (B): cyt b below diagonal, microsatellites above diagonal. * = significant (p < 0.05), = remained significant after Bonferroni correction. A Danube River NW Black Sea Dnieper River Baltic S. Slovakia Serbia Varna Bilgorod Odessa S. Bug R. Kiev Khakhovka Kherson Crimea NC Black Sea Sea of Azov Kerch Strait Azov S. A. Baltic Sea B. Slovakia C. Serbia * D. Varna 0.115* E. Bilgorod 0.085* * F. Odessa 0.165* * H. S. Bug R * 0.026* 0.086* 0.067* I. Kiev J. Khakhovka * 0.072* 0.145* K. Kherson * 0.057* * G. Crimea * L. Kerch Strait * * M. Sea of Azov *

52 B Volga River Azerbaijani Caspian Sea Iranian Caspian Sea_ Moskva R. Volga-Don Canal Svetli Yar Damchik Nabran Shikh Lenkoran Alet Lisar N. Moskva River O. Volga-Don Canal * 0.032* P. Svetli Yar 0.349* * Q. Damchik R. Nabran * * * S. Shikh * U. Lenkoran 0.601* T. Alet 0.577* V. Lisar *

53 Table 2.5 Pairwise contingency test results for round goby samples from the Black/Azov Sea (A) and Caspian Sea (B) basins: Cyt b below diagonal, sats above diagonal. N = not significant, * = significant (p < 0.05), = remained significant after Bonferroni correction. A Danube River NW Black Sea Dnieper River Baltic S. Slovakia Serbia Varna Bilgorod Odessa S. Bug R. Kiev Khakhovka Kherson Crimea NC Black Sea Sea of Azov Kerch Strait Azov S. A. Baltic Sea - * N N * N N N B. Slovakia N - * * * * * C. Serbia N N - * * D. Varna * - N * N N N N N E. Bilgorod * * - N * N N N * F. Odessa N N - N * N * H. S. Bug R. N * * * * * - * N N N N I. Kiev * - J. Khakhovka N N N N * * N - N * * K. Kherson N * * * N N N - N N N G. Crimea * N - N * L. Kerch Strait * * N - N M. Sea of Azov * * * N - 39

54 B Volga River Azerbaijani Caspian Sea Iranian Caspian Sea_ Moskva R. Volga-Don Canal Svetli Yar Damchik Nabran Shikh Lenkoran Alet Lisar N. Moskva River - O. Volga-Don Canal - * * P. Svetli Yar * N - * Q. Damchik N N - R. Nabran - * N * S. Shikh N - N N U. Lenkoran * * N N - N T. Alet * * N N N - V. Lisar N N N N - 40

55 Table 2.6 Nested clade analysis of round goby cytochrome b haplotypes showing significant nesting clades and significant subclades, clade dispersion (D C ) and displacement values (D N ), the inference chain, inference results, and map locations (lettered according to Fig. 1, with Caspian Sea locations italicized). For clade dispersion and displacement values, (S) = a significantly small value, (L) = a significantly large value. Figure is in electronic appendix. Nesting Clade Significant Subclades DC DN Inference Chain Result Map Locations Total 4-1 (S) p < (S) p < ,19 N Allopatric Fragmentation A*,B*,C*,D,E,F,G,H,I*,J,K,L,M,N*,P* 4-2 (S) p < (L) p < N*,O*,P*,Q,R,S,T,U,V I-T (S) p = (L) p < (S) p = ,2,3,4 N Restricted Gene Flow A*,B*,C*,D,E,F,G,H,I*,J,K,M 3-2 (L) p = with Isolation by Distance D,E,F,G,H,L,M,N*,P* 3-3 (S) p = (S) p = E,F,G I-T (L) p = (L) p = (L) p = ,2,3,4 N Restricted Gene Flow N*,P*,R,S,U 3-6 (S) p = (S) p = with Isolation by Distance O*,P*,Q,R,S,T,U,V I-T (L) p = (L) p = (L) p = (L) p = ,2,3,4 N Restricted Gene Flow A*,B*,C*,D,E,F,G,H,I*,J,K 2-2 (S) p = (S) p = with Isolation by Distance C*,H,I*,K,M I-T (L) p = (L) p = (S) p < (S) p < ,2,11,12 N Contiguous Range Expansion E,F,H,G,L,M 2-4 (S) p = M 2-5 (S) p = D,E,F 2-6 (L) p < N*,P* I-T (S) p = (S) p < (S) p < ,19,20,2,11,12,14 Y Contiguous Range Expansion, R,S,T,U,V 2-13 (S) p < (L) p < Long Distance Colonization, O*, P*, Q I-T (S) p < or Past Fragmentation (S) p < (S) p < , 2, 3, 4 N Restricted Gene Flow E 1-3 (S) p = with Isolation by Distance E, G 41

56 1-8 (L) p < (L) p < A*, B*, C*, D, E, F, G, H, I*, J, K I-T (L) p < (L) p < (S) p = , 2, 11, 12 N Contiguous Range Expansion E, F, G, H, L, M 1-15 (S) p = F, G, L (S) p = , 2, 11, 12 N Contiguous Range Expansion S, V I-T (S) p = (L) p = (L) p = , 2, 3, 4 N Restricted Gene Flow A*,B*,C*,D,E,F,G,H,I*,J,K 8 (S) p = with Isolation by Distance J,K I-T (L) p = (L) p = (L) p = , 2, 3, 4 N Restricted Gene Flow C, K 58 (S) p = (S) p = with Isolation by Distance H, I* I-T (L) p = (L) p = I-T (S) p = , 2, 11, 12 N Contiguous Range Expansion R, S, T, U, V 42

57 Fig. 2.1 Eurasian round goby distribution (hatched areas), showing sampling locations (lettered). Circles = the Black-Azov Sea basin clade, squares = the Caspian Sea basin clade, with filled symbols indicating non-native locations. Fig. 2.2 Neighbor-joining tree of phylogenetic relationships among common Eurasian round goby Neogobius melanostomus haplotypes (found in multiple individuals), constructed in MEGA v3.1 (Kumar et al. 2004), and rooted to its monkey goby sister species (N. fluviatilis) and three other neogobiin species. Bootstrap % 50% pseudoreplications. Branch lengths are proportional to genetic divergence. Timeline shows major geologic Ponto-Caspian events, and is proportional to tree length, using a molecular clock calibration of 2%/million years for cytochrome b variation in gobies (Rocha et al. 2005). The Black and Caspian Sea clades split during the Chauda-Baku faunal exchange ~ 350,000 ya. Fig. 2.3 Genetic distance tree among round goby population sites constructed in PHYLIP v3.6 (Felsenstein 1989) with Cavalli-Sforza chord distances (Cavalli-Sforza & Edwards 1967) and combined mitochondrial cytochrome b and microsatellite data. Bootstrap % values from 1000 pseudo-replicates. Branch lengths are proportional to genetic divergence. Fig. 2.4 Parsimony network among round goby Neogobius melanostomus cytochrome b haplotypes. Symbol size is proportional to haplotype frequency and is colored to indicate location. The sister species, the monkey goby N. fluviatilis, connects to the network at 43

58 haplotype 37 (*), after <100 mutational steps. The nesting diagram is available in electronic appendix 1. Fig. 2.5 Bayesian STRUCTURE v2.1 (Pritchard et al. 2000, Pritchard & Wen 2004) analysis of round goby populations using combined data from seven microsatellite loci and cytochrome b sequences. K=15 (posterior probability, pp = 0.999). Each individual is represented by a thin vertical line, which is partitioned into K colored segments that represent the individual s estimated membership fractions. Black lines separate individuals from different sampling sites, which are labeled above the figure. Ten STRUCTURE runs at each K produced nearly identical individual membership coefficients, having pairwise similarity coefficients above 0.95, and the figure illustrating a given K is based on the highest probability run at that K. Results show five main population genetic divisions in the Black/Azov Sea basin and three within the Caspian Sea basin. 44

59 Fig

60 Fig

61 Fig

62 Fig

63 Fig

64 Appendix 2.1 Statistical parsimony network with nesting groups from nested clade analysis. Results of this analysis are given in Table 2.6. Significant clades in the analysis are shaded on the figure. 50

65 Chapter 3 Invasion genetics of the Eurasian round goby in North America: Tracing sources and spread patterns Previously published as: Brown JE, Stepien CA (2009) Invasion genetics of the Eurasian round goby in North America: tracing sources and spread patterns. Molecular Ecology, 18, Abstract The Eurasian round goby Neogobius melanostomus invaded the North American Great Lakes in 1990 through ballast water, spread rapidly, and now is widely distributed and moving through adjacent tributaries. We analyze its genetic diversity and divergence patterns among 25 North American (N = 744) and 22 Eurasian (N = 414) locations using mtdna cytochrome b gene sequences and 7 nuclear microsatellite loci in order to: (1) identify the invasion s founding source(s), (2) test for founder effects, (3) evaluate whether the invasive range is genetically heterogeneous, and (4) determine whether fringe and central areas differ in genetic diversity. Tests include F ST analogs, neighborjoining trees, haplotype networks, Bayesian assignment, Monmonier barrier analysis, and 3-dimensional factorial correspondence analysis. We recover 13 cytochrome b haplotypes and 232 microsatellite alleles in North America and compare these to variation we 51

66 previously described across Eurasia. Results show: (1) the southern Dnieper River population was the primary Eurasian donor source for the round goby s invasion of North America, likely supplemented by some alleles from the Dniester and Southern Bug Rivers, (2) the overall invasion has high genetic diversity and experienced no founder effect, (3) there is significant genetic structuring across North America, and (4) some expansion areas show reduced numbers of alleles, whereas others appear to reflect secondary colonization. Sampling sites in Lake Huron s Saginaw Bay and Lake Ontario significantly differ from all others, having unique alleles that apparently originated from separate introductions. Substantial genetic variation, multiple founding sources, large number of propagules, and population structure thus likely aided the goby s ecological success. Introduction Introductions of exotic species often provide unplanned experiments to test evolutionary theory and patterns (Huey et al. 2005) as their new habitats offer novel environmental and ecological pressures (Jeschke & Strayer 2005; Kajita et al. 2006), along with possible release from co-evolved predators, pathogens, and parasites (e.g., Torchin et al. 2003; Joshi & Vrieling 2005; Kvach & Stepien 2008). Ecological theory predicts that a new introduction is likely to be founded by only a few individuals that contain a fraction of the source population s genetic variation, which may limit adaptive potential and success (Frankham 2005; Poulin et al. 2005). However, some exotic introductions experience little to no reduction in genetic diversity, due to large numbers of founding propagules and multiple founding sources, which likely translate to greater adaptability (Durka et al. 52

67 2005; Stepien et al. 2005; Roman & Darling 2007). Population level genetic diversity appears important for long-term stability (Stepien 1995), especially if adaptive genetic variation is maintained along with neutral variation (Zayed & Whitfield 2008). Correctly identifying donor-recipient populations and their transport pathways are vital to evaluating a species relative ecological success in its introduced versus native ranges (Amsellem et al. 2000; Kang et al. 2007). Here we analyze the invasion genetics of the round goby Neogobius melanostomus (Teleostei: Gobiidae; see Stepien & Tumeo 2006, Neilson & Stepien 2009a for nomenclature) in the North American Great Lakes (Jude et al. 1992; Corkum et al. 2004), which rank among the world s most invaded freshwater systems with ~182 introduced species (Mills et al. 1993; Ricciardi 2006). The round goby was discovered in the St. Clair River in 1990, and then rapidly spread throughout the Great Lakes and beyond (Fig. 3.1A; Jude et al. 1992; Charlebois et al. 1997; Irons et al. 2006). It is now one of the most abundant nearshore fishes in the lower Great Lakes, reaching 90 individuals/m 2 (Ray & Corkum 2001; Johnson et al. 2005). The round goby occurs in all five Great Lakes, as well as many of their connecting rivers. It has expanded eastward through the St. Lawrence River and the Erie Canal, and southward into the Mississippi watershed via the Illinois River and the Chicago Sanitary and Shipping Canal (United States Geological Survey Nuisance Aquatic Species Fact Sheet, ) Like the round goby, ~73% of Great Lakes invaders entered via ballast water exchange from oceanic vessels (Holeck et al. 2004; Ricciardi 2006) and ~70% of recent invaders 53

68 ( ) trace to the Eurasian Ponto-Caspian region (Ricciardi & MacIsaac 2000). Three shipping routes extend from the Ponto-Caspian region, across the Atlantic, and into the Great Lakes, including: (1) from the Black Sea through the Mediterranean Sea, (2) from the Danube River into the North Sea, or (3) from the Dnieper River into the Baltic Sea (Ricciardi & MacIsaac 2000; Grigorovich et al. 2002). These pathways are evaluated in our study for the round goby introduction(s). The round goby s invasion success likely was facilitated by the abundance of its native Ponto-Caspian dreissenid mussel prey in the Great Lakes, which preceded the goby s introduction - also via ballast water (Vanderploeg et al. 2002). Such ecological facilitation is hypothesized to lead to more rapid establishment, successful spread, and long-term persistence (Simberloff & Van Holle 1999; Ricciardi 2001; Sutherst et al. 2007). Moreover, Kvach & Stepien (2008) found that Great Lakes populations of the round goby remain relatively parasite-free in comparison with native populations, supporting the enemy release hypothesis (Williamson 1996; Keane & Crawley 2002). That hypothesis predicts that the ecological success of invaders is enhanced by their leaving their native parasites (and likely predators and pathogens) behind. Most fish parasites within the Great Lakes have not yet adapted to colonizing the round goby, and their absence likely has continued to enhance the goby s success from the early course of its invasion through the present day (Kvach & Stepien 2008). In the Great Lakes, the round goby is consumed by native predatory fishes, including yellow perch, walleye, and smallmouth bass (Jude 1997; Belanger & Corkum 2003; Truemper et al. 2006); and in 54

69 turn competes with and displaces native benthic fishes, and eats fish eggs (Chotkowski & Marsden 1999; Nichols et al. 2003; Steinhart et al. 2004). Few Great Lakes introduced species have been genetically assessed to date. Both the spiny water flea Bythotrephes longimanus (Berg et al. 2002) and the ruffe Gymnocephalus cernuus (Stepien et al. 1998, 2004, 2005) show evidence of single source colonization. The water flea originated from Lake Ladoga in Russia (Berg et al. 2002) and the ruffe from the Elbe River system in Germany (Stepien et al. 2004, 2005). All genotypes of the ruffe from the Elbe River region are present in the upper Great Lakes, in similar frequencies, and were likely introduced via a Baltic Sea transport pathway into the upper Great Lakes through grain shipping coinciding with increased U.S. trade with Germany near the time of its reunification (Stepien et al. 2004, 2005). In contrast, sporadic introductions of the Chinese mitten crab Eriocheir sinensis into the Great Lakes trace to multiple European ports rather than native Asian populations (Tepolt et al. 2007); however, this species is not established as it does not reproduce in freshwater (Herborg et al. 2007). Marsden et al. (1995) and Stepien et al. (1999, 2002, 2005) separately described relatively high levels of genetic variation in the dreissenid zebra (Dreissena polymorpha) and quagga mussel (D. r. bugensis) invasions of the Great Lakes, suggesting little overall loss of genetic diversity in comparison with putative source populations. The Great Lakes introduction of the freshwater tubenose goby Proterorhinus semilunaris (formerly known as P. marmoratus; Stepien & Tumeo 2006; Neilson & Stepien 2009b) another Ponto-Caspian goby is characterized by levels of genetic diversity that are similar to native river source populations (Stepien et al. 2005; 55

70 Stepien & Tumeo 2006; Neilson & Stepien 2009b). Moreover, prior genetic study of the round goby invasion by our laboratory discerned considerable genetic variation and population structure in comparison with native populations (Dillon & Stepien 2001; Stepien et al. 2005; Stepien & Tumeo 2006), and this is the first study to comprehensively test its source(s) and analyze variability across its present-day range. We test the following hypotheses underlying the genetic history of the round goby invasion in the Great Lakes: 1) The introduction was founded by a single versus multiple source population(s), 2) Genetic diversity levels do/do not reflect a founder effect, 3) The invasive population is genetically homogenous/heterogeneous across its range, and 4) The fringe population areas have reduced/similar genetic variability compared to its original establishment center. These hypotheses are tested using both mitochondrial (mt) DNA cytochrome b gene sequences and 7 nuclear microsatellite ( sat) DNA loci, which allow both phylogeographic history of lineages to be assessed (through maternally inherited mtdna) and fine-scale population genetic relationships to be discerned (via comparisons of biparentally inherited microsatellites with the mtdna patterns). We then relate the round goby s invasion genetics to its ecological success, in light of invasion theory and patterns of other exotic species. 56

71 Materials and Methods Sampling, DNA extraction, and amplification We analyze 25 North American (N = 744) locations (Table 3.1, Fig. 3.1A, sites AA-YY). These samples are compared with previously published results from 22 (N = 414) introduced and native Eurasian locations in the Baltic, Black, and Caspian Seas watersheds (Fig. 3.1B, sites A-V; see Brown & Stepien 2008). Samples tested here encompass locations across the North American Great Lakes, with fine-scale concentration in Lake Erie. We also include an expansion population site in the Des Plaines River outside the Great Lakes watershed (Fig. 3.1A). We focus on locations of initial round goby sightings and include areas of major shipping traffic and ballast water exchange. Samples were placed directly in 95% ethanol in labeled vials and stored at room temperature. Genomic DNA was extracted from fin tissue using Qiagen DNeasy kits (Qiagen, Inc.; Valencia, CA), eluted in 100 μl of water, stored at 4 o C until used for amplification, and archived at -80 o C. Amplification and sequencing primers for the mt cytochrome (cyt) b gene (1204 bp) were L14724 (Meyer et al. 1990), H15149 (Kocher et al. 1989), H5 (Akihito et al. 2000), and L15066 (Brown & Stepien 2008). The mtdna PCR amplification mixture contained 25 ul of 1 unit Taq polymerase, 200 μm dntps, 50 mm KCl, 1.5 mm MgCl 2, 10 mm Tris-HCl, 0.5 μm of each primer (L14724, H5), and ~30 ng of template. Amplification was conducted on a MJR DYAD thermalcycler (Bio-Rad 57

72 Laboratories, Hercules, CA) with an initial 120 sec denaturation at 94 o C, followed by 35 cycles of 45 sec at 94 o C, annealing at 52 o C for 30 sec, and extension at 72 o C for 60 sec; with a final 180 sec extension at 72 o C. Sequencing was outsourced to Cornell University Life Sciences Core Laboratories Center ( Ithaca, NY), which uses Applied Biosystems Automated 3730 DNA Analyzers (ABI; Foster City, CA). Sequence data were processed in our laboratory, aligned using BIOEDIT v7.0 (Hall 1999, 2004) and CLUSTAL X v1.8 (Thompson et al. 1994). Seven nuclear sat loci were developed for the round goby by Dr. Kevin Feldheim of the Field Museum of Natural History (Brown & Stepien 2008). The PCR amplification mixture contained 0.6 units Taq, 50 μm nucleotides, 50 mm KCl, 1.5 mm MgCl 2, 10 mm Tris-HCl, 0.5 μm of each primer, and ~30 ng of template in 10 μl, with an oil overlay to ensure a constant reaction volume. Amplification encompassed a 2 min initial denaturation for 94 o C; followed by 35 cycles of 30 sec at 94 o C, 1 min annealing at a primer-specific temperature, and 30 sec at 72 o C; with a final 5 min extension at 72 o C. Amplification products were diluted 1:50, of which 1 µl was added to 13 µl of formamide/ ABI Gene Scan 500 size standard and loaded onto a 96-well plate for analysis on an ABI 3130 Genetic Analyzer using ABI GENEMAPPER v4.0. Output profiles were manually checked to confirm allelic size variants. Genetic analyses 58

73 We used ARLEQUIN (v3.01; Excoffier et al. 2005) to identify individual round goby mtdna haplotypes, in comparison with our results from their Eurasian range (Brown & Stepien 2008) and other Ponto-Caspian gobies (Stepien & Tumeo 2006; Neilson & Stepien 2009b). Allelic frequencies, number of private alleles, and (for sat data) conformance to Hardy-Weinberg (HW) equilibrium expectations, and linkage disequilibrium were evaluated using GENEPOP v3.4 (Raymond & Rousset 1995, 2004). Levels of significance for HW and linkage disequilibrium tests were adjusted using sequential Bonferroni correction (Rice 1989), and compared with results from the False Discovery Rate method (Benjamini & Hochberg 1995). HW deviations were tested for heterozygote deficiency or excess, and for the presence of null alleles with MICROCHECKER v2.23 ( van Oosterhout et al. 2004, 2006). Genetic data (cyt b and sat) were analyzed to identify true populations, i.e., samples having significantly divergent gene pools, using the pairwise F-statistic analog θ ST (Weir & Cockerham 1984) and 2 contingency tests (Raymond & Rousset 1995; Goudet et al. 1996). Fine-scale population relationships, such as those tested here, have been shown to be better resolved with contingency tests (see Balloux & Lugon-Moulin 2002); which are independent from Hardy-Weinberg equilibrium assumptions, non-parametric, and less affected by small sample size (Raymond & Rousset 1995; Goudet et al. 1996). Our additional use of θ ST facilitated direct comparisons with other population genetic studies (see Stepien et al. 2007; Allendorf & Luikart 2007). 59

74 Both F-statistics and contingency tests employ the population sample as the unit of comparison, whereas the Bayesian model-based approach in STRUCTURE v2.1 (Pritchard et al. 2000; Pritchard & Wen 2004) use the individual as the unit, assessing whether it belongs to one or more population group regardless of geographic origin. These Bayesian methods were used to assess likely donor sources for the invasion (Abdelkrim et al. 2007; D Amato et al. 2007; Schrey et al. 2007), and were based on combined cyt b and sat data. STRUCTURE assigned individuals to population groups ranging from K = 1 to K = 47 (total sites), with the relative frequency of their predicted group memberships totaling Ten independent runs for each K were evaluated, with burn-ins of 100,000 replicates and run lengths of 500,000. Groupings obtained from the STRUCTURE results then were tested independently with Analysis of MOlecular VAriance (AMOVA; Excoffier et al. 1992) in ARLEQUIN, with significance determined using the nonparametric permutation approach described by Excoffier et al. (1992). To further analyze the relationships among population sites, neighbor-joining (NJ) trees (Saitou & Nei 1987) were constructed from Cavalli-Sforza chord distances (Cavalli- Sforza & Edwards 1967) for the cyt b and sat data in PHYLIP v3.6 (Felsenstein 1989, 2008). Phylogeographic relationships among the cyt b haplotypes were examined with a statistical parsimony haplotype network generated at a 95% confidence level with TCS (Clement et al. 2000) and with a neighbor-joining tree constructed in PHYLIP. Ambiguous loops in the haplotype network were resolved according to Pfenninger & Posada (2002). We tested the hypothesis of genetic isolation with geographic distance using cyt b, sat, and combined data, θ ST genetic distances, nearest connected waterway 60

75 distances (km), and 10,000 rearrangements in Mantel (1967) tests in GENEPOP s ISOLDE. The fit of data to the regression line was evaluated with Microsoft Excel 2003 (Microsoft, Redmond, WA). Three-dimensional factorial correspondence analysis (3D-FCA; Benzecri 1973) in GENETIX v4.05 (Belkhir et al. 2004) further explored population relationships using the sat data, making no a priori assumptions about population groupings. Finally, the relative magnitude of genetic divisions across the round goby s North American distribution was investigated using an analytical computational geometry approach by Manni et al. (2004a, b; BARRIER 2.2), which identified geographically continuous and discontinuous assemblages of sampling sites independent from a priori knowledge of geographical population structure (e.g. lakes or river drainages). Pairwise estimates of θ ST from the sat data were mapped onto a matrix of their geographical coordinates (latitude and longitude). The spatial organization of subpopulations was modeled by Voronoi tessellation and a Monmonier (1973) maximum difference algorithm identified which neighboring populations were most genetically differentiated (Manni et al. 2004a, b). Initial estimates of genetic discontinuities were made with a multilocus θ ST matrix, followed by a second analysis that incorporated single-locus θ ST values. This procedure ranked each identified barrier in relative magnitude, according to respective support from individual locus θ ST values. Results 61

76 Genetic composition We identify 13 round goby mt cyt b North American haplotypes; whose GenBank accession numbers are EU331156, EU331160, EU , EU331204, EU331207, and EU ( Supplemental Material Table 3.S1). All haplotypes found in North America belong to the subspecies Neogobius melanostomus melanostomus from the Black/Azov Sea drainage, and none are the N. m. affinis subspecies from the Caspian Sea drainage (Fig. 3.2; Brown & Stepien 2008, Supplemental Material Fig. 3.S1). Six of the 13 North American round goby haplotypes (1, 5, 7, 8, 49, & 57) also were described by us from Eurasian populations (Brown & Stepien 2008), and the remaining 7 (52-56 & 87-88) are rare and have not yet been identified in Eurasia. Six haplotypes (1, 7, 8, 57, 87, & 88) occur in multiple North American locations (Tables 3.1 & 3.3). Of the remaining 7 haplotypes, 6 are singletons and haplotype 54 characterizes 3 individuals from the Bay of Quinte in Lake Ontario. Haplotype 1 is the most abundant in both North America (73.7%; Tables 3.1 & 3.3) and Eurasia (46.8%; Brown & Stepien 2008). The three next most common Great Lakes haplotypes (7, 8, 57) also occur in the Dnieper River, near the port of Kherson, Ukraine in about the same proportions as in Lake St. Clair (Fig. 3.1, Supplemental Material Table S1). These 4 haplotypes (1, 7, 8, & 57) constitute 97.5% of all North American round goby individuals sampled, and 100% of those found in Kherson (Supplemental Material Table 3.S1). Two of the North American singletons (5 & 49) respectively also occur at low frequency in the Eurasian sites of the Dniester River estuary and the Southern Bug River. 62

77 Round goby sampling locations in North America average 2.5 (± 0.3) cyt b haplotypes (range: 1-6), compared with 6.8 (± 1.1) from Eurasian Black Sea watershed locations (range: 1-10; Brown & Stepien 2008). Mean haplotype diversity of North American samples is (± 0.042), compared with (± 0.092) in the native Black Sea drainage (Table 3.1). However, haplotype diversity of the St. Clair River population the original point of introduction averages and contains 6 haplotypes, which is higher than the mean of other Great Lakes sites and similar to that found in native Dnieper River locations ( ). Higher cyt b diversity occurs in round goby samples from Lakes Superior, Michigan, Huron, St. Clair, Erie and Ontario. However, some fringe locations in these lakes have low genetic diversity, suggesting secondary founder effects during their spread (Table 3.1). We discern 305 alleles among 7 sat loci assayed for the round goby across North America and Eurasia (see Brown & Stepien 2008). Of those, 232 (76%) characterize North American samples (mean: 33.1 alleles/locus), representing a high proportion of the native genetic diversity. Individual North American sampling locations have total alleles, similar to the number found in native populations (Table 3.1). Locus Ame10 has the fewest alleles (16), and locus Ame129 contains the most (48). Round goby sites in North America average 55.9 (± 5.2) alleles, with the fewest found in the Des Plaines River (23; the site outside the Great Lakes watershed) and the most occurring in the St. Clair River (111; the original introduction site). Loci do not deviate from Hardy- Weinberg equilibrium expectations and are unlinked. Microsatellite allelic diversity from North American samples averages (± 0.012), slightly more than the level found in 63

78 the native Black Sea drainage (0.460 ± 0.023); thus the introduction experienced no founder effect. In comparison, sat gene diversity is uniformly high - showing no founder effect or bottlenecks. Invasion source and genetic structure Our results show that the North American invasion principally originated from a source in the southern Dnieper River system, which drains into the Black Sea at the major shipping port of Kherson, Ukraine. The neighbor-joining trees of sampling sites (Supplemental Material Fig. 3.S2A) links the North American invasion and Dnieper River samples as a single population group with high bootstrap support, excluding all other native Eurasian sites. Bayesian assignment tests using STRUCTURE (best K = 2, posterior probability = 0.97; Fig. 3.3), AMOVA tests (Table 3.2), and three-dimensional factorial correspondence analysis (3D-FCA; Fig. 3.4) also group the North American samples with those from the Dnieper River, excluding other Eurasian locations. The 3D- FCA highlights the distinctiveness of invasive populations in the Baltic Sea and Danube River from North American populations, showing that they were not sources for Great Lakes. MtDNA cyt b genetic composition and haplotype frequency are equivalent between the St. Clair River introduction site and the Dnieper River port of Kherson, Ukraine (Table 3.3, Fig. 3.1). However, the St. Clair River location also has 2 rare round goby haplotypes that have not been found in the Dnieper River. This may be due to sampling or reflect contribution from another Eurasian founding source. 64

79 Round goby samples across the Great Lakes demonstrate significant genetic heterogeneity (i.e., marked divergence among sites). Almost all sites diverge markedly in all pairwise tests using the sat data and Lake Huron s Saginaw Bay and Lake Ontario s Bay of Quinte populations also diverge using the cyt b data (Table 3.3; see Avise 2004). These results thus show high genetic structure across the round goby s invasive range, with little gene flow among locations following their establishment. BARRIER analysis of the sat data reveals 5 primary genetic divisions among round goby population groups (Fig. 3.1A). Barriers I (= the strongest barrier) and V distinguish populations in Lake Huron s Saginaw River (KK) and Saginaw Bay (LL) from all other locations. Barriers II and III characterize populations in Georgian Bay of Lake Huron (HH) and southern Lake Michigan (EE-FF), respectively, and barrier IV splits the population sites sampled in Lake Ontario. The patterns of the cyt b, sat, and combined data sets do not correspond to a genetic isolation by geographic distance model (y = x ; R 2 = ; p = 0.065). Low genetic diversity in peripheral areas of the colonization in the cyt b data likely reflects founder effects (Supplemental Material Fig. 3.S2B), which also are shown by significant differences in the sat data. Central locations near the invasion s origin (e.g., MM, QQ) have the greatest genetic diversity and most closely resemble the genetic composition of the source population from the southern Dnieper River (Table 3.3). Beyond the central locations, there are 2 general cyt b haplotype patterns. The first shows increasing prevalence of haplotype 1 (Supplemental Material Fig. 3.S2B), which is monomorphic in some areas of recent population expansion (e.g., Des Plaines River, site 65

80 FF and Lake Huron s Georgian Bay, HH). The second pattern is an increase in the frequency of haplotype 8 relative to the St. Clair River initial introduction site (MM, 14%; Supplemental Material Fig. 3.S2B), reaching 63% in Lake Huron s Saginaw River (KK) and 60% in Saginaw Bay (LL). The sample from the Bay of Quinte in eastern Lake Ontario (VV; Fig. 1A) does not follow this pattern, being dominated by haplotype 7 (68%; Supplemental Material Table 3.S1) and containing other rare haplotypes (<10%; Supplemental Material Table 3.S1). Haplotype 7 also occurs at the original introduction site in the St. Clair River (2%, MM), as well as in the St. Lawrence River (4%, YY), the Danube River (1%, B-C), and the Dnieper River at Kherson founding source (8%, K). The Bay of Quinte population appears markedly distinct, grouping outside of all other North American samples on the cyt b genetic distance tree (Supplemental Material Fig. 3.S2B). The Bay of Quinte population also contains private haplotypes and unique sat alleles at all loci examined. It appears nestled with other North American sites in the sat tree (Supplemental Material Fig. 3.S2A) and other analyses (Fig. 3.4); and thus appears closer to the other samples according to the nuclear DNA data. Discussion High genetic diversity and its primary source Our data indicate that the North American round goby invasion principally originated from the mouth of the southern Dnieper River at Kherson, Ukraine, which is a major Black Sea port where ballast water is exchanged (Ricciardi & MacIsaac 2000). Round gobies likely were taken up in ballast water at night, when the juveniles rise into the 66

81 water column to feed (Hensler & Jude 2007). Kherson is linked to an invasion corridor that leads through the Mediterranean Sea to the Atlantic Ocean and then to the Great Lakes (Ricciardi & MacIsaac 2000). We find no support for round goby introductions via other hypothesized routes to North America, either through the Danube and Rhine Rivers to the North Sea or via the Dnieper and Vistula Rivers to the Baltic Sea (Ricciardi & MacIsaac 2000, Vanderploeg et al. 2002). Some exotic introductions succeed despite few propagules, founder effects, and consequent low genetic diversity (Planes & Lecaillon 1998; Tsutsui et al. 2000; Grapputo et al. 2006); however, the Great Lakes introduction of the round goby has very high diversity and most likely was founded by a very large number of propagules. Almost all of the alleles in the primary source population of the southern Dnieper River at Kherson, Ukraine (Brown & Stepien 2008) also occur in the North American samples (Fig. 3.1A, Supplemental Material Table 3.S1), indicating a very large introduction size. We discern no evidence for a founder effect, with the overall number of alleles from the Great Lakes introduction site at the St. Clair River being greater than the number recovered from the Dnieper River founding source. Some additional rare alleles occur in some North American samples that have not been identified in Europe to date. These alleles likely are found in very low frequencies in a Eurasian founding source population, which may be the southern Dnieper River or another (unidentified) location. These allelic frequencies may have changed stochastically post-introduction. 67

82 Other Eurasian round goby source populations may have augmented the North American invasion, suggested by the high genetic diversity and significant divergences we find across their introduced range. Similar cases of high diversity and structure of invasions have been documented by some other studies also based on neutral loci (e.g., Ellstrand & Schierenbeck 2000; Lavergne & Molofsky 2007; Kolbe et al. 2008). If this high neutral genetic diversity also corresponds to adaptive genetic variation in coding loci, then the round goby introduction likely could have experienced enhanced adaptability (Zayed & Whitfield 2008). The large number of colonists and high genetic diversity then presumably aided the round goby s ecological fitness in spreading throughout the Great Lakes and beyond. Invasion Population Structure Despite evidence supporting a primary founding source population area for the round goby invasion, considerable genetic structure occurs across its exotic range. Mitochondrial DNA cyt b and nuclear sat data show congruent genetic structure, with the latter showing greater resolution due to the faster evolutionary rate and larger effective population size of the nuclear genome (see Avise 2004). The genetic composition of the round goby at its original discovery site in the St. Clair River is statistically indistinguishable from the southern Dnieper River population. Other areas of high genetic diversity in the Great Lakes have similar genetic composition to those from the St. Clair and Dnieper Rivers. 68

83 However, Great Lakes sampling locations located farther from the invasion s center show evidence of secondary founder effects during expansion (Austerlitz et al. 1997; Hallatschek & Nelson 2008). Lower genetic diversity at the expansion s edge is a common phenomenon, with alleles then accumulating over time (Hewitt 1996; Ibrahim et al. 1996). According to this leading edge hypothesis, short-term bottlenecks occur as the species spreads, which then disappear with subsequent gene flow (Hewitt 1996; Bensch & Hasselquist 1999; Merilä et al. 1996). This pattern is observed in the mtdna genome in our data, likely due to its smaller effective population size (Avise 2004). Our results do not support a genetic isolation by geographic distance hypothesis for the round goby invasion, suggesting a mixture of short- and long-distance dispersal. These likely resulted from a combination of natural expansion and human-mediated jump dispersal, the latter via bait bucket transport by anglers and commercial shipping (Clapp et al. 2000; Carman et al. 2006). Most sampling sites are genetically distinguishable, supporting high genetic structure across the round goby s invasion. The most divergent populations occur in the Saginaw River/Bay system in Lake Huron and the Bay of Quinte in Lake Ontario, which are each distinguished from others by great genetic differences (F ST = ) and appeared later, respectively in 1997 and These genetic divergences likely resulted from independent founder events, followed by little gene flow (Table 3.1). The samples have high genetic diversities, are dominated by alleles that are either absent or rare in other locations, and share some alleles with the St. Clair River location but in markedly different frequencies. Two hypotheses that may explain these differences are: 1) they 69

84 may each have been colonized by a biased sample of founders from other Great Lakes sites, or 2) they may have been independently founded or augmented by separate introduction events originating from overseas. The first hypothesis likely explains the divergence of the round goby sample from the Saginaw Bay/River system, since this system has only a single rare sat allele. In contrast, the Bay of Quinte population has diversity levels comparable to the St. Clair River population, yet includes 15 additional private alleles that are distributed among all loci (cyt b and sat). Thus, the Bay of Quinte population likely resulted from a novel round goby introduction, setting it apart from the rest of the round goby invasion. A direct introduction to the Bay of Quinte from Eurasia is unlikely, since it is not a trans- Atlantic port and all of its intra-great Lakes shipping originates from the international port of Toronto, ON (T. Johnson, Ontario Ministry of Natural Resources, pers. com.). It is possible that one or more ships that travel between Toronto and the Bay of Quinte may have introduced unique alleles arriving from overseas. Alternatively, a ship may have exchanged its ballast upon entering eastern Lake Ontario, providing unique founders for that part of the lake, and its members then may have dispersed into the Bay of Quinte. We are sampling more round goby population sites from Lake Ontario to discern among these scenarios. Genetic comparisons with other exotic species High genetic diversity of Great Lakes populations of the exotic round goby likely is due to the establishment of numerous diverse founding propagules from the Black Sea. A 70

85 large-scale introduction, high genetic diversity, and no founder effect are similar to results characterizing other Ponto-Caspian invaders, including the introduced zebra Dreissena polymorpha and quagga D. r. bugensis mussels (Marsden et al. 1995; Stepien et al. 2002, 2005), ruffe Gymnocephalus cernuus (Stepien et al. 1998, 2005), and freshwater tubenose goby Proterorhinus semilunaris (Stepien et al. 2005; Stepien & Tumeo 2006; Neilson & Stepien 2009b). Some evidence supports the leading edge hypothesis (Hewitt 1996), as peripheral expansion areas in the Great Lakes have lower cyt b genetic diversity than is found at the center of its invasive range. However, this hypothesis needs further testing to confirm its occurrence in Great Lakes round goby samples. A leading edge pattern also has been observed for the sea lamprey Petromyzon marinus expansion across the Great Lakes, with locations in Lakes Huron, Michigan, and Superior having lower genetic diversity than do central areas in Lakes Erie and Ontario (Bryan et al. 2005). This phenomenon is similar to that observed in historic range expansions of the meadow grasshopper Chorthippus parallelus in Eurasia (Hewitt 1996), the greenside darter Etheostoma blennioides in North America (Haponski & Stepien 2008), and the alpine rock-cress Arabis alpina in northern Europe (Ehrich et al. 2007); suggesting that reduced diversity may be a general pattern of some expanding populations. In the round goby, this reduced genetic variation is likely temporary (Bensch & Hasselquist 1999), and is not seen in our sat data. 71

86 Although we identify the port of Kherson as the primary source for the North American round goby s introduction, the presence of other alleles raises the possibility of secondary augmentation from other parts of the Black Sea drainage. Multiple introduction sources have been shown to enhance the introduction success of other introduced species. For example, the expansion of the amphipod Gammarus tigrinus into the Great Lakes and Europe traces to 3 allopatric native locations along the northwestern Atlantic Ocean (Kelly et al. 2006). Likewise, the European green crab Carcinus maenas was introduced to the North American Atlantic coast from multiple sources (Roman 2006). The original introduction had low diversity, originated mostly from southern Europe, and ranged from New Jersey, USA to southern Nova Scotia, Canada for almost 200 years. In the 1980s and 1990s, the green crab rapidly expanded northward due to new introduction of northern European lineages from ballast water (Roman 2006). Likewise, the recent successful expansion of the brown anole Anolis sagrei in the southeastern United States was fueled by the admixture of distinct native lineages from Cuba (Kolbe et al. 2004, 2008). In contrast, successful Great Lakes introductions of the spiny water flea Bythotrephes longimanus (Berg et al. 2002) and the ruffe Gymnocephalus cernuus (Stepien et al. 1998, 2004) originated from single locations in northern Europe, albeit with retention of high local genetic diversity. We are analyzing additional Eurasian round goby populations and are testing temporal patterns across North America, in order to further understand the genetic factors fueling its invasional success. Conclusions and summary 72

87 Comparing the population genetic characters of the North American round goby invasion with Eurasian samples leads to the following conclusions, which aid our understanding of its rapid spread and ecological success: 1) The primary source population for the North American introduction originated from the mouth of the Dnieper River at the port of Kherson, Ukraine on the Black Sea. Other founding events and propagules from additional Eurasian sources may have supplemented its genetic diversity. 2) The invasion is characterized by high genetic diversity and underwent no founder effect, attributed to large numbers of introduced propagules. 3) Round goby populations are genetically divergent across the Great Lakes, differing in allelic composition. This appears due to spread patterns that are a mixture of short- and long-distance dispersal, as well as possible genetic additions from other Eurasian founding sources. 4) Some spread locations display founder effects, fitting a leading edge model; whereas other edge areas have high genetic diversity and unique alleles. The latter may have been founded independently or significantly augmented from additional Eurasian sources. In summary, the round goby s North American introduction provides an important model for exploring a range of demographic processes that occur during an exotic species 73

88 invasion and expansion. The invasion is characterized by a suite of genetic factors that favored its success; including a large number of colonists, high genetic diversity, and the possibility of multiple donor locations. The combination of these population genetic factors, along with ecological facilitation (Simberloff & Van Holle 1999; Ricciardi 2001) provided by its native dreissenid mussel prey (Vanderploeg et al. 2002) and enemy release from its native and new parasites (Kvach & Stepien 2008), likely augmented the round goby s invasional success. 74

89 Tables & Figures Table 3.1 Locations of round goby samples in North America (Fig. 3.1), with latitude, longitude, date of first sighting, sample size (N), cytochrome b haplotypes and haplotype diversity, microsatellite number of alleles (N A ) heterozygosity measures (H O = observed, H E = expected), F IS, and proportion of private alleles. Bold haplotypes are shared among Great Lakes locations. Sighting dates are from the United States Geological Survey Nuisance Aquatic Species fact sheet ( and from the joint Ontario Federation of Anglers and Hunters and the Ontario Ministry of Natural Resources Invasive Species Program (John Zoltak, Ontario Federation of Anglers and Hunters, pers. comm.). Water body Map label and location Latitude Longitude First sightin g Cytochrome b N Haplotypes Haplotype diversity L. Superior AA. Duluth, WI , 52, Microsatellite N A H O H E F IS % private alleles L. Michigan Total (BB-FF) % BB. Petoskey, MI , % CC. Grand Haven, MI % DD. Muskegon, MI , 8, % EE. Chicago, IL % FF. Des Plaines R., IL % % 75

90 3 1 1 L. Huron Total (GG-LL) % GG. Alpena, MI , % HH. Severn, ON % II. Linden, MI % JJ. Russelville, MI % KK. Saginaw, MI , 8, 57, % LL. Saginaw Bay, MI , % St. Clair R. MM. Port Huron, MI , 7, 8, 57, 87, % L. St. Clair NN. Mt. Clemens, MI , 8, 57, % L. Erie Total (OO-TT) % OO. Oregon, OH , 8, % PP. Gibraltar Island, OH , 8, 57, % QQ. Avon, OH , 5, 8, % RR. Chagrin, OH , % SS. Erie, PA % TT. Erie County, NY , % L. Ontario Total (UU-VV) % UU. Hamilton, ON , % 76

91 VV. Picton, ON , 7, 8, 54, 55, % Erie Canal WW. Lockport, NY , % St. Lawrence R. Total (XX-YY) % XX. Montreal, QC , % YY. Quebec City, QC , 8, % North America Total (AA-YY) % Dnieper R. Total (I-K) % I. Kiev, Ukraine , 58, % J. Khakhovka, Ukraine Native 15 1, % K. Kherson, Ukraine Native 25 1, 7, 8, % Baltic Sea A. Gdynia, Poland % Danube R. Total (B, C) , 7, % Black Sea Total (D, E, F, G, H) % E. Dniester R Native 20 1, 2, 3, 4, 5, 6, 9, % H. S. Bug R Native 27 1, 18, 49, 50, 51, % 58, 72, Native Black Sea (D, F, G) Native 61 1, 10, 11, 18, 48, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 74, 75, 76, 77, 78, % Sea of Azov Total (L, M) Volga R. Total (N-Q) , 80 Native 20 11, 12, 13, 14, 15, 16, 17, 18, 19, 41, 42, 43, 44, 45, 46, 47, 48 Native 50 20, 21, 22, 24, 25, 26, 27, 28, 29, % 0% 77

92 Caspian Sea Total (R-V) Native 66 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 70, 71, 81, 82, 83, 84, 85, % 78

93 Table 3.2 Analysis of Molecular Variance (AMOVA; Excoffier 1992) results showing partitioning of genetic variation within and among round goby samples using cytochrome (cyt) b and microsatellites ( sat). -statistics (F-statistic analogs) significant at p < A. Comparisons between 2 primary population groups: all North American locations (AA-YY) plus the Dnieper River (I-K) versus introduced and native Eurasian locations (A-H, L-V). B. Comparisons between 2 primary population groups: all North American locations (AA-YY) plus the Dnieper River (I-K) locations versus native Eurasian locations (excluding nonindigenous ones). Division of Variation Partitioning Tested A. 2 primary population groups: NA + Dnieper R. sites vs. all other Eurasian locations Among sampling sites within the 2 groups Within sampling sites Measure of Variation Partitioning cyt b sat Proportion of Variation CT Proportion of Variation SC Proportion of Variation ST B. 2 primary population groups: NA + Dnieper R. sites vs. native Eurasian locations Among sampling sites within the 2 groups Within sampling sites Proportion of Variation CT Proportion of Variation SC Proportion of Variation ST

94 Table 3.3 Pairwise ST comparisons between round goby samples from pooled locations (in ARLEQUIN v3.01; Excoffier et al. 2005) for cytochrome b below the diagonal and microsatellites above. * significant (p<0.05), remains significant after sequential Bonferroni correction (Rice 1989). Results from contingency tests (Raymond & Rousset 1995; Goudet et al. 1996; not shown) are congruent. Sampling Locations L. Superior (AA) ~ L. Michigan (BB-FF) 0.01 ~ L. Huron proper (GG-HH) ~ L. Huron/Saginaw (II-LL) ~ L. St Clair (MM-NN) 0.04* 0.02* ~ L. Erie (OO-TT) 0.08* * 0.08* 0.00 ~ L. Ontario/Hamilton (UU) 0.10* 0.03* 0.10* 0.16* ~ L. Ontario/Picton (VV) ~ Erie Canal (WW) * 0.11* 0.53 ~ St. Lawrence R. (XX-YY) * * ~ Dnieper R./Kiev (I) ~ Dnieper R./Khakhovka (J) * * ~ Dnieper R./Kherson (K) 0.04* * * 0.00 ~ Baltic Sea (A) ~ Danube R. (B-C) * ~ Dniester R. Estuary (E) * 0.17* * * 0.26 ~ S. Bug R. (H) * 0.03* * * * * ~ Black Sea (D, F, K) * * * 0.07 ~ 80

95 Fig. 3.1 Current distribution of the round goby (shaded), in (A) North America and (B) Eurasia with circles showing our sampling sites (lettered according to Table 3.1; with double letters and denoting North American sites from this study and single letters designating Eurasian sites from Brown and Stepien (2008)). Pie charts show the relative frequencies of cytochrome b haplotypes (center) and alleles from 3 representative microsatellite loci (rings; Ame 1, 33, & 117). (A) In North America, the Great Lakes watershed boundary is shown by solid line, along with 5 primary barriers to gene flow among population sites (ranked I-V; I = greatest difference) determined from BARRIER analysis (Manni et al. 2004a, b). (B) Circles = the Black-Azov Sea basin clade, squares = the Caspian Sea basin clade, with filled symbols indicating non-native locations in Eurasia. Note the similarity between the St. Clair River site of the original introduction (site MM and the port of Kherson in the lower Dnieper River, site K; 2 : p=0.762). Fig. 3.2 Statistical parsimony network showing relationships among the 88 round goby cytochrome b haplotypes using TCS v1.2 (Clement et al. 2000). Double letters (AA-XX) = North American sites (Fig. 3.1; this study), single letters (A-V) = Eurasian sites from Brown & Stepien (2008). Fig. 3.3 Bayesian STRUCTURE v2.1 (Pritchard et al. 2000; Pritchard & Wen 2004) analysis of round goby populations using combined data from 7 microsatellite loci and cytochrome b sequences. K = 2 (posterior probability = 0.97) was best-supported in separate runs, which ranged from K = 1 to K = 47 (the total number of sampling locations). Each individual is represented by a thin vertical line, which is partitioned into 81

96 K colored segments that represent the individual s estimated membership fractions. Black lines separate different sampling sites, which are labeled above the figure. Ten STRUCTURE runs at each K produced nearly identical individual membership coefficients, having pairwise similarity coefficients above 0.95, and the figure shows the highest probability run. Results show that the North American and the southern Dnieper River (site K) round goby sampling locations group together, excluding the other Eurasian locations. Fig. 3.4 Three-dimensional factorial correspondence analysis (Benzecri 1973, GENETIX v4.05; Belkhir et al. 2004) of round goby microsatellite data showing clustering among North American (AA-XX) and Dnieper River (I-K) sites. Individuals from the Baltic (A) and Black Seas (D, E, G, H) group together, with component 3 separating them from sites in the Danube River (B, C) and the northwest Black Sea site at Odessa (F). 82

97 Fig

98 Fig

99 Fig

100 Fig

101 Supplemental Material Table 3.S1 Occurrence of mtdna cytochrome b haplotypes in North American round goby samples (lettered according to Fig. 3.1, with GenBank ( accession numbers and frequency. Eurasian frequencies of haplotypes present in North American round goby samples included for comparison. * = identified in Eurasia by Brown & Stepien (2008). The Sea of Azov, Volga River, and Caspian Sea locations share no haplotypes with North America. Accession Number Total N.A % L. Superior 24 (AA) 92.3% L. Michigan 95 (BB-FF) 86.4% L. Huron 48 proper 94.1% (GG-HH) L. Huron/ Saginaw (II-LL) L. St. Clair (MM-NN) L. Erie (OO-TT) L. Ontario/ Hamilton (UU) L. Ontario/ Picton (VV) Erie Canal (WW) St. Lawrence R. (XX-YY) Dnieper R./ Kiev (I) Haplotype 1* 5* 7* 8* 49* * Total N Haplotype EU EU EU EU EU EU EU EU EU EU EU EU EU % % % % 2 6.5% 24 96% % 5 33% 1 0.1% % % 1 0.1% 1 0.1% 1 0.1% 3 0.4% 1 0.1% 1 0.1% % 2 0.3% 7 1% % % 3.9% 100% % 2.7% 3.6% 100% % 100% % % % % 0 1 1% % % 1 1% 2 1.9% % % % 9% 0.4% 100% % 100% % 9.7% 9.7% 3.2% 3.2% 100% % 100% % 3.4% 1.7% 1.7% 100% %

102 Dnieper R./ Khakhovka (J) Dnieper R./ Kherson (K) Baltic Sea (A) 14 93% 19 76% % % 0 2 8% 3 12% % % % % Danube R. (B-C) 82 98% 0 1 1% % 3 Dniester R. Estuary (E) 9 45% 1 5% % 8 S. Bug R. (H) 20 71% % % 9 Native Black Sea (D, F, K) 22 36% % 22 88

103 Fig. 3.S1 Neighbor-joining tree of 88 cytochrome b haplotypes and their locations. NA = North America, BS = Black Sea, AS = Azov Sea, VR = Volga River, CS = Caspian Sea. Bootstrap values (1000 replicates) are given as percentages on nodes. Fig. 3.S2 Neighbor-joining trees of round goby population samples based on Cavalli- Sforza & Edwards (1967) chord distances, in PHYLIP v3.6 (Felsenstein 1989, 2008). Bootstrap values (1000 replicates) are given on nodes. * = Eurasian source locations (Brown & Stepien 2008). A. Distances based on microsatellite data. B. Distances based on cytochrome b data, showing secondary (expansion) founder effects. 89

104 Fig. 3.S1 90

105 Fig. 3.S2A 91

106 Fig. 3.S2B 92

107 Chapter 4 Population genetic history of the dreissenid mussel invasion: Expansion patterns across North America Submitted for publication in Biological Invasions as: Brown JE, Stepien CA. Population Genetic History of the Dreissenid Mussel Invasion: Expansion patterns Across North America. Abstract This study tests population genetic patterns across the Eurasian dreissenid mussel invasion of North America encompassing the zebra mussel Dreissena polymorpha (1986 detection) and the quagga mussel D. rostriformis bugensis (detected in 1990 and now has largely displaced the former in the Great Lakes). We evaluate their sourcespread relationships and invasion genetics using 8-9 nuclear microsatellite loci for 583 zebra (21 sites) and 269 quagga (12 sites) mussels from Eurasian and North American range locations, with the latter including the Great Lakes, Mississippi River basin, Atlantic coastal waterways, Colorado River system, and California reservoirs. Additionally, the mtdna cytochrome b gene was sequenced to verify species identity. Our results indicate that North American zebra mussels originated from multiple nonnative northern European populations, whereas North American quagga mussels trace to native estuaries in the Southern Bug and Dnieper Rivers. Invasive populations of both species show considerable genetic diversity and structure (zebra F ST = , quagga F ST = ), without founder effects. Most newer zebra mussel 93

108 populations have appreciable genetic diversity, whereas quagga mussel populations from the Colorado River and California show some founder effects. The population genetic composition of both species changed over time at given sites; with some adding alleles from adjacent populations, some losing them, and all retaining closest similarity to their original composition. Zebra mussels from Kansas and California appear genetically similar and assign to a possible origin from the St. Lawrence River, whereas quagga mussels from Nevada and California assign to a possible origin from Lake Ontario. These assignments suggest that overland colonization pathways via recreational boats do not necessarily reflect the most proximate connections. In conclusion, our microsatellite results comprise a valuable baseline for resolving present and future dreissenid mussel invasion pathways. Introduction Genetic studies using high-resolution microsatellite DNA loci offer means to elucidate the pathways and vectors of invasive species, and assess their temporal population dynamics. This is the first comprehensive population genetic analysis of the Eurasian dreissenid mussel (Mollusca: Bivalvia: Dreissenidae) invasion across North America, which has been one of the most ecologically and economically important aquatic introductions. The zebra mussel Dreissena polymorpha appeared in the Great Lakes in 1986, where it was introduced via ballast water (Hebert et al. 1989; Carlton 2008), followed by the quagga mussel D. rostriformis bugensis in ~1989 (May & Marsden 1992; Mills et al. 1996). Prior to its introduction to North America, the zebra mussel had previously expanded its original Ponto-Eurasian range via central European canals in the 94

109 Black Sea and Baltic Sea drainages (Morton 1993; Bij de Vaate et al. 2002), whereas the Great Lakes was one of the quagga mussel s first documented expansions (Pligin 1979; Karatayev et al. 2008). The zebra mussel also spread from the Baltic Sea drainage to Italy (1960s), Ireland (1997), and Spain (2001) by recreational boats (Bij de Vaate et al. 2002; Aldridge et al. 2004; Gosling et al. 2008). Dreissenid mussels widely disperse via their planktonic larvae, whereas juveniles and adults adhere to hard substrates with byssal threads (summarized by Ackerman et al. 1994). These life history traits distinguish dreissenids from most native North American freshwater bivalves, which lack a planktonic larval stage and do not attach to hard substrates (Johnson & Padilla 1996). The spread of dreissenids was facilitated by their life history traits, including unintentional transport of larvae in ships ballast and bilge water, and by attachment of their juveniles and adults to engines, hulls, and anchors (Ricciardi et al. 1995; Mackie & Schloesser 1996; Johnson et al. 2001). Their fouling of power generation and water treatment facilities in the United States and Canada has cost $161-$467 million/year (Connelly et al. 2007). They filter feed on phytoplankton with high efficiency that shifted the basis of Great Lakes aquatic food chains from pelagic to benthic (Berg et al. 1996; Zhu et al. 2006). Spread of dreissenid mats on the lake floors - where they host an associated community of plants and invertebrates in their shells and interstices - has further increased benthification, markedly changing the ecosystem. Dreissenids likely facilitated the invasion of the round goby Neogobius melanostomus (Vanderploeg et al. 2002), their natural predator that became established throughout the Great Lakes with high genetic diversity and no founder effect (Stepien & Tumeo 2006; Brown & Stepien 2009). 95

110 The zebra mussel spread rapidly, expanding throughout the Great Lakes and into the Mississippi and Hudson River drainages within five years of detection (Nalepa & Schloesser 1993; Allen & Ramcharan 2001; Kelly et al. 2009). The quagga mussel spread more slowly (Kelly et al. 2009), gradually replacing zebra mussels in the St. Lawrence River (Ricciardi & Whoriskey 2004) and the lower Great Lakes (Mills et al. 1999; Jarvis et al. 2000; Wilson et al. 2006), and is now doing the same in the upper Lakes (Nalepa et al. 2009). Researchers and managers attempted to stop westward spread at the 100 th meridian, to prevent their colonization of the heavily dammed and diverted waterways of the western United States (Mangin 2001; Britton & McMahon 2005; Bossenbroek et al. 2007, 2009). However, in 2007 the quagga mussel was detected in Lake Mead, Nevada - a reservoir on the Colorado River - and then discovered downstream (Stokstad 2007). Dreissenids since have spread into many southern California reservoirs via the extensive canal system radiating from the Colorado River ( In 2008, zebra mussels were detected in San Justo Reservoir in central California ( speciesid=5). It is important that we understand the population sources and pathways of these recent expansions so that we can evaluate the efficacy of our control efforts, and attempt to avoid further expansions. Role of genetics Modern genetic techniques can play a crucial role to aid invasive species identification and evaluate their population linkages. Hebert et al. (1989), May & Marsden (1992), and Stepien et al. (1999) used genetics to determine the species identify of dreissenid 96

111 invaders in the Great Lakes. Additional genetic research has examined portions of the invasive ranges for zebra (Elderkin & Klerks 2001; Stepien et al. 2001, 2003, 2005; Gosling et al. 2008) and quagga mussels (Stepien et al. 2001, 2003, 2005), illustrating the need for more rapidly-evolving reproducible genetic markers to understand their origin and spread, compare genetic variation, determine population genetic structure, and predict their expansion pathways. The genetic composition of an introduced population can be substantially altered from its native source(s), including the amount of overall variation and its geographic partitioning. Genetic variation comprises the raw material for adaptive evolution, which can play a critical role governing invasive success (Blossey & Notzold 1995; Bossdorf et al. 2005). Selectively neutral molecular markers can serve as a proxy for estimating quantitative genetic variation in phenotypic traits that underlies adaptive evolution (Merila & Crnokrak 2001; Reed & Frankham 2001; Chun et al. 2009). Population genetic data also are useful for elucidating the number and sources of an introduction (Stepien et al. 2005; Rosenthal et al. 2008; Brown & Stepien 2009), as well as determining whether changes occur during the time course of its establishment and spread (Andreakis et al. 2009; Henry et al. 2009). Such information can aid control efforts, for example, determining the genetic origin of the invasive climbing fern Lygodium microphyllum led to a biological control agent that was matched to its native genotypes (Goolsby et al. 2004). The genetic variation of an introduced population may be reduced, increased, or unchanged relative to native populations; with reduction as the most common outcome (Nei et al. 1975; Ficetola et al. 2008; Henry et al. 2009). Three processes may reduce 97

112 genetic diversity of the introduction and preclude significant variation among local population groups. First, most introductions are founded by only a few individuals that represent only a small part of native variation, and subsequent bottlenecks and Allee effects may further reduce it (DeWalt & Hamrick 2004; Grapputo et al. 2005; Griffen & Drake 2008). Second, introduction and spread likely lead to high gene flow that homogenizes population genetic structure (Viard et al. 2006; Kim et al. 2009). Lastly, a recent introduction may lack sufficient time for local adaptation or drift to generate local population structure in the new range (Maron et al. 2004; Whitney & Gabler 2008). Any new genetic variation would need to arise from mutations and recombinations, before population structure could develop (Perez et al. 2008). Selection also might lead to reduced genetic diversity; since when an introduced genotype is more fit, its numbers will increase relative to others, resulting in a net loss of genetic variability (Kliber & Eckert 2005). If these processes predominate during an introduction, the new population should be genetically depauperate in comparison to native populations (Amsellam et al. 2000; DeWalt & Hamrick 2004; Grapputo et al. 2005). If an introduction is founded from multiple genetically differentiated native populations, its resultant genetic diversity can be higher than that characterizing native populations; increasing heterozygosity and/or generating novel allelic combinations (Ellstrand & Schierenbeck 2000). Several studies have shown patterns indicating multiple introductions and high genetic diversity in invaders (Kolbe et al. 2004; Williams et al. 2005; Stepien et al. 2005; Brown & Stepien 2009). Because increased genetic diversity likely also increases potential for adaptive evolution, an invader with high diversity may evolve quickly in its new range. Another possible evolutionary outcome for an 98

113 introduction is no significant change in genetic diversity between the native and introduced ranges. This is possible if large population size is maintained during an invasion (Brown & Marshall 1981) or regained shortly following introductions (Zenger et al. 2003). Understanding the origin of population genetic structure, or lack thereof, in an invasive species also can provide valuable information on transport pathways from source populations. Correctly identifying transport pathways and source populations is critical to interpreting an invader s relative ecological success in introduced versus native ranges and for targeting management efforts to shut down invasion pathways (Amsellam et al. 2000; Kang et al. 2007). Like the zebra and quagga mussels, 73% of Great Lakes invaders entered via ballast water exchange from oceanic vessels (Holeck et al. 2004; Ricciardi 2006; Kelly et al. 2009) and 70% of those from traced to the Eurasian Ponto-Caspian region (Ricciardi & MacIsaac 2000). Three shipping routes extend from the Ponto-Caspian region, across the Atlantic, and into the Great Lakes, including: (i) from the Black Sea through the Mediterranean Sea, (ii) from the Danube River into the North Sea, or (iii) from the Dnieper River into the Baltic Sea (Ricciardi & MacIsaac 2000; Grigorovich et al. 2002). These pathways are evaluated in our study for the zebra and quagga mussel introductions. Relatively few Great Lakes-introduced species have been genetically assessed to date. Both the spiny water flea Bythotrephes longimanus (Berg et al. 2002) and the ruffe Gymnocephalus cernuus (Stepien et al. 1998, 2004, 2005) showed evidence of singlesource colonization. The water flea originated from Lake Ladoga in Russia (Berg et al. 99

114 2002) and the ruffe from the Elbe River system in Germany (Stepien et al. 2004, 2005). All the Elbe River ruffe genotypes are present in the upper Great Lakes in similar frequencies, and likely were introduced into the upper Great Lakes via increased U.S. shipping trade with Germany near the time of its reunification (Stepien et al. 2004, 2005). In contrast, Chinese mitten crab Eriocheir sinensis introductions into the Great Lakes traced to multiple European ports rather than native Asian populations (Tepolt et al. 2007); however, this species does not reproduce in freshwater and thus failed to establish (Herborg et al. 2007). The Great Lakes introduction of the freshwater tubenose goby Proterorhinus semilunaris has been characterized by levels of genetic diversity that are similar to native river source populations (Stepien et al. 2005; Stepien & Tumeo 2006; Neilson & Stepien 2009b). Invasive Ponto-Caspian round goby Neogobius melanostomus populations in the Great Lakes have considerable genetic variation and population structure in comparison with native populations (Dillon & Stepien 2001; Stepien et al. 2005; Stepien & Tumeo 2006; Brown & Stepien 2008), and the Dnieper River population was identified as its primary source (Brown & Stepien 2009). Marsden et al. (1995) and Stepien et al. (1999, 2002, 2005) separately described relatively high levels of genetic variation in the dreissenid zebra and quagga mussel invasions of the Great Lakes, suggesting little overall loss of genetic diversity in comparison with putative source populations. Allozyme markers by Marsden et al. (1995) appeared to suggest a single source for the zebra mussel invasion, yet a later allozyme study showed significant genetic heterogeneity among populations in the Great Lakes and a complex of inland lakes, suggesting an important role for stochastic processes governing their population genetic structure (Lewis et al. 2000). Subsequent genetic studies using mitochondrial and 100

115 nuclear DNA markers similarly found high genetic diversity and implicated multiple founding sources, pointing to western- central European sources for the zebra mussel (Stepien et al. 2002, 2005). In this new study, we analyze the invasion genetics of the zebra mussel and the quagga mussel in North America and Eurasia using high-resolution nuclear microsatellite markers for the first genetic study across their expanded geographic range. We test the following hypotheses underlying the genetic history of the dreissenid invasions in North America: 1. Each species was founded by a single versus multiple source populations. 2. The invasive populations are genetically homogenous/heterogeneous across their ranges. 3. Genetic diversity levels reflect overall founder effects/show no founder effects. 4. Populations outside the Great Lakes have reduced/similar/greater genetic diversity compared to those within the Great Lakes. 5. Once established, populations remain genetically static/change over time. Materials and Methods Sampling, DNA extraction, and amplification We sampled 24 zebra mussel (N=583) and 13 quagga mussel (N=269) population locations across North America and Eurasia (Fig. 4.1, Table 4.1). Samples were collected by hand or by fishery agency trawls, and were either immediately frozen, or placed in 101

116 95% ethanol and stored at room temperature. Genomic DNA was extracted from mantle tissue using Qiagen DNeasy kits (Qiagen, Inc.; Valencia, CA), eluted in 100 μl of water, stored at 4 C until used for amplification, and then archived at -80 C. Genetic variation was assessed from eight microsatellite (μsat) loci for zebra mussels (identified as Dpo) and six for quagga mussels (Dbu) developed by Feldheim et al. (in review), and three additional previously published loci per species (Table 4.2). PCR amplifications were performed in our laboratory using 10μL reaction volumes of 0.6 units Taq, 50 μm nucleotides, 50 mm KCl, 2.5 mm MgCl 2, 10 mm Tris-HCl, 0.5 μm of each primer, and 30 ng of template; with a mineral oil overlay to maintain reaction volume. A thermal cycle of 2 min at 94 C for initial denaturation was followed by 35 cycles of denaturation at 94 C for 30 sec, annealing for 1 min at C (primer-specific; Table 2), and extension at 72 C for 30 sec; capped by a final extension at 72 C for 5 min. Amplification products were diluted 1:50, of which 1µl was added to 13µl of formamide and ABI (Applied Biosystems, Inc.; Foster City, California) Gene Scan 500 size standard, and loaded onto a 96-well plate for analysis on an ABI 3130xl Genetic Analyzer using GeneMapper v4.0. Output profiles were checked to confirm allelic size variants. A portion of the mtdna cyt b gene was amplified from representative individuals to confirm species identity at sites using polymerase chain reactions (PCR) of 25μL, including 1 unit Taq polymerase, 200 μm dntps, 50 mm KCl, 2 mm MgCl 2, 10 mm Tris-HCl, 0.5 μm of each primer (151F, 270R; Merritt et al. 1998), and 30 ng of 102

117 template. Amplification on a MJR DYAD thermalcycler (Bio-Rad Laboratories, Hercules, CA) consisted of an initial denaturation step at 94 C for 2 min, followed by 35 cycles of denaturation at 94 C for 45 sec, annealing at 52 C for 30 sec, and extension at 72 C for 1 min; concluded by a final 72 C extension for 3 min. Sequencing of the mitochondrial (mt) cytochrome (cyt) b gene (336bp) used the primers 151F and 270R, and was outsourced to the Cornell University Life Sciences Core Laboratories Center ( Ithaca, NY). Sequences then were trimmed by us to remove primer sequences, aligned with CLUSTAL X v2.0 (Larkin et al. 2007), and adjusted using BIOEDIT v7.0 (Hall 1999, 2004). These sequences are deposited in GenBank as accession numbers GQ for D. polymorpha and GQ for D. rostriformis bugensis. Genetic analyses ARLEQUIN (v3.11; Excoffier et al. 2005) was used to assign individuals to the correct mtdna cyt b haplotype; and allelic frequencies, number of private alleles, conformance of sats to Hardy-Weinberg (HW) equilibrium expectations, and linkage disequilibrium were evaluated in GENEPOP v4.0 (Rousset 2008). HW deviations were tested for heterozygosity deficiency or excess, and for the presence of null alleles using MICRO- CHECKER v2.23 ( van Oosterhout et al. 2004, 2006). Levels of significance for all tests were adjusted using Bonferroni correction (Sokal & Rohlf 1995). 103

118 Microsatellite genetic composition of the samples was analyzed to identify true populations (i.e., those distinguished by significantly divergent gene pools) using the F- statistic analog θ ST (Weir & Cockerham 1984) and contingency tests (Raymond & Rousset 1995; Goudet et al. 1996). Relationships among recently diverged samples, such as those tested here, have been shown to be better resolved in models using contingency tests (see Balloux & Lugon-Moulin 2002), which are independent of HW equilibrium assumptions, non-parametric, and unaffected by sample size (Raymond & Rousset 1995; Goudet et al. 1996). Our use of the F-statistic analog facilitated direct comparisons with other studies (see Brown & Stepien 2008, 2009). Probability values of both tests were adjusted using sequential Bonferroni corrections (Rice 1989). In order to further analyze the relationships among population sites, pairwise genetic distances were calculated based on the microsatellite data using Cavalli-Sforza chord distances (Cavalli-Sforza & Edwards 1967) and neighbor-joining trees (Saitou & Nei 1987) were constructed in PHYLIP v.3.68 (Felsenstein 1989). The trees were rooted to their respective close relatives D. stankovici and D. rostriformis grimmi (see Stepien et al. 2003, 2005 for phylogeny) using microsatellite data from Feldheim et al. (in review). Three-dimensional factorial correspondence analysis (3D-FCA; Benzecri 1973) was used to examine population divisions in GENETIX v.4.05 (Belkhir et al. 2004) via a visual representation of the relationships among populations without a priori assumptions about populations and evaluates variation within and among sites. Populations that experience a genetic bottleneck usually exhibit a decrease in the number of alleles at polymorphic microsatellite loci faster than their respective heterozygosity 104

119 declines (Luikart et al. 1998a). In order to detect such a signal, a test for heterozygosity excess was carried out using the program BOTTLENECK (Cornuet & Luikart 1996, Piry et al. 1999). This test compared the observed gene diversity (Nei 1987) to the expected equilibrium gene diversity, which is calculated from the observed number of alleles assuming a constant-size (i.e., equilibrium) population (Luikart et al. 1998b). Wilcoxon tests using a stepwise mutation model were used to establish whether the number of loci showing heterozygosity excess was significantly greater than that expected in populations at equilibrium, with distribution of expected gene diversity at equilibrium estimated from10,000 simulations (Cornuet & Luikart 1996). The Bayesian model-based methods of Rannala and Mountain (1997) in STRUCTURE v (Pritchard et al. 2000; Pritchard & Wen 2004) and GENECLASS2 (Piry et al. 2004) use the individual as the unit, assigning it to the most likely population group(s) regardless of geographic origin. STRUCTURE was used to assign individuals to population groups, ranging from K = 1 (a single population group, i.e., the null hypothesis of panmixia) to K = N (the total N of sampling groups), with the relative frequency of individual membership per group totaling Ten independent runs for each K were used with burn-ins of 100,000 replicates and 1,000,000 replicates. We complemented this analysis with assignment tests using GENECLASS2 v 2.0 (Piry et al. 2004), which assigned each individual a probability of membership among each sampled location using simulated population sizes of 10,000 individuals per sampling site and a 0.01 rejection level (Cornuet et al. 1999). 105

120 Results We recovered 386 zebra mussel alleles from 11 loci and 228 quagga mussel alleles using nine loci. All populations for both species conformed to Hardy-Weinberg equilibrium expectations, all loci were unlinked, and there was no evidence of null alleles. Our results support that both dreissenid introductions to North America originated from multiple Eurasian sources. For the zebra mussel, the 3D-FCA (Fig. 4.2A) and the neighbor-joining tree (Fig. 4.3A) analyses indicate at least 2 distinct introductions, with Lake Erie samples from both years being distinct from other North American samples. The majority of the North American samples group with introduced locations in northern Europe and the Volga River, but the Lake Erie samples do not group with any locations we tested here. The 3DFCA (Fig. 4.2B) and the neighbor-joining tree (Fig. 4.3B) for quagga mussels are largely congruent, highlighting distinct introduction events one into that contributed the majority of the diversity found in North American samples, and a second that influenced diversity in Lake Erie (evident in samples from both years). Quagga mussels that were introduced to the North American Great Lakes appear to trace to the general area of the Southern Bug and Dnieper River estuaries, which are geographically proximate (~60 km). 106

121 Invasive population structure All introduced zebra mussel populations appear significantly differ from each other in pairwise tests (Table 4.3A, 4.4A), showing a high degree of genetic heterogeneity across their ranges, with the exception of a single comparison. The two Kansas locations do not significantly differ from each other, but diverge from all other populations (Table 4.3A). Similarly, the only population sites with low self-assignment in the GENECLASS test were the Kansas samples, which had high assignments to each other (Table 4.5A). Bayesian assignment analysis reveals 5 major groups within the zebra mussel data set (Fig. 4.4A). One (colored blue in Fig. 4.4A) links most of the introduced Eurasian populations (sites 16-19) with the Volga River Russia samples (site 21). A second group (colored pink) differentiates the Lake Erie (site 10) population from all others. The third group (colored tan) links the Dnieper River (site 20) with samples from the Mississippi River drainage (sites 4-6), Oneida Lake (site 13), and the Hudson River (site 15, 2003). The fourth group (colored green) links samples from the upper Great Lakes (sites 7-9) and the earlier Hudson River sampling date (site 15, 1993). The final group (colored brown) assigns zebra mussel locations in the western United States (site 1-3) to a group that includes the Lake Ontario (site 11) and St. Lawrence River (site 14) samples, suggesting the eastern Great Lakes may have served as the source for this westward expansion. This grouping also has strong support in the 3D-FCA (Fig. 4.2A) and the neighbor-joining tree (Fig. 4.3A). All quagga mussel populations significantly differ from one another in the pairwise tests (Table 4.3B). Their AMOVA groupings also are significant (Table 4.4A), and the 107

122 populations show high self-assignment (Table 4.5B). The 3D-FCA (Fig. 4.2B), neighbor-joining tree (Fig. 4.3B), and Bayesian assignment test (Fig. 4.4B) are congruent in recovering 3 population groups. The first grouping links quagga mussel populations from the upper Great Lakes (sites C-D) with the invasive Volga River sample (site L). The second group comprises samples from the Dnieper River (site K), Oneida Lake (H), the St. Lawrence River (I) and some individuals from Lake Erie (E) in Quagga mussel sampling locations in the western United States (sites A, B) group with the Lake Ontario (F, G) samples in both the 3DFCA (Fig. 4.2B), the neighbor-joining tree (Fig. 4.3B), and the Bayesian assignment analysis (Fig. 4.4B); suggesting that Lake Ontario was a likely source for the quagga mussels in the Colorado River and southern California similar to the pattern shown in zebra mussels. Genetic diversity levels Individual zebra mussel sampling locations have total alleles (Table 4.1). Loci Dpo04 and Dpo272 have the fewest alleles (13), whereas locus DpolB8 has the most (58; Table 4.2A). The introduced population site in Spain (map location 16) possesses the fewest alleles, and the native Dnieper River (site 20) population has the greatest number. Great Lakes samples average 90 alleles, whereas native Eurasian populations average 113, which is a significant difference (t-test t = 2.168, df = 13, p = ). Samples from long established locations in the Mississippi River Basin (sites 4-6) average 105 alleles and the recent western expansion locations (1-3) average 106 alleles. Both are significantly higher than the average for the Great Lakes locations (p = and p = 0.002, respectively). Zebra mussels show a decrease in average number of alleles per 108

123 location from Eurasia to the Great Lakes, but subsequent expansion areas have higher numbers of alleles. However, these consist mostly of low frequency alleles (<0.01%) also found in various Great Lakes locations. Great Lakes zebra mussel samples have average observed heterozygosity (H O ) values that are slightly lower than those found in the native Eurasian range (0.67±0.08 vs. 0.74±0.09), which is significant (t-test p=0.018), although the heterozygosity distribution does not suggest a genetic bottleneck for the North American introduction (p=0.735) in the BOTTLENECK analysis. Zebra mussel populations outside the Great Lakes have somewhat lower H O (0.67±0.08) than those within the Great Lakes (0.73±0.04), which is not significant (t-test p=0.093). However, testing for genetic bottlenecks indicates that the three western expansion locations experienced bottlenecks (p= in each). In contrast, this analysis approach suggests that populations from Lake Superior (1995; p=0.011) and Lake Ontario (p=0.005) experienced continuous population expansion. Quagga mussel samples have total alleles (Table 4.1), with loci Dbu75 and Dbu92 having the fewest alleles (13) and locus Dbug05 having the most (40; Table 4.2B). The introduced site in California (site A) has the fewest number of alleles, and the introduced population in Lake Huron (site D) possesses the most. Great Lakes sites average 78 alleles and native locations average 74; whose means significantly differ (t = 4.763, df = 9, p = 0.001). The recently established location in California and Nevada averages 48 alleles, which is significantly lower than the Great Lakes locations (t = 4.014, df = 8, p = 0.004). Great Lakes populations of quagga mussels have an average H O that appears slightly lower than in their native Eurasian range (0.76±0.08 vs. 109

124 0.74±0.05), but is not significant (t-test p=0.496). Quagga mussel samples outside the Great Lakes have lower H O than found within (0.76±0.08 vs. 0.60±0.18), which is significant (t-test p=0.039). Great Lakes populations of quagga mussels have a greater than average number of alleles, but do not differ in observed heterozygosity from native Eurasian populations, with their heterozygosity distribution refuting a genetic bottleneck (p=0.338) However, recent western expansions of the quagga mussel have lower average numbers of alleles and lower observed heterozygosities, supporting a founder effect. Moreover, heterozygosity excesses detected by the BOTTLENECK analysis for Lake Matthews (p=0.018) and Lake Mead (p=0.040) samples support a genetic bottleneck in the western expansion of the quagga mussel. The reduction appears greatest in the California sample, which traces to an origin from Lake Mead via the Colorado River aqueduct. Two samples show signs of population expansion detectable in BOTTLENECK analysis: Lake Huron (p=0.003) and western Lake Ontario (p=0.024); which suggest rapid, continuous population growth following introduction. Temporal genetic change Temporal comparisons among three zebra mussel sampling locations reveal significant differences among years using both F ST and contingency tests (Table 4.3A), with strong self-assignment using GENECLASS (Table 4.5A). Thus all samples are most similar to their original population compositions, despite some temporal changes. Temporal comparisons of zebra mussel samples from Lake Erie (-16 alleles; 17%; ST =0.091, p<0.0001) and Lake Superior (-22 alleles; 21%; ST =0.154, p<0.0001) show a net loss of alleles (Table 4.1), whereas the Hudson River population reveals a net gain of alleles 110

125 (+28 alleles; 33%; ST =0.140, p<0.0001). According to the BOTTLENECK analysis, the Lake Superior population lost the signal of rapid population expansion between 1995 and None of the other time pair comparisons show demographic signals in the BOTTLENECK analysis, suggesting that those changes did not result from changes in the size of the reproducing population. For quagga mussels, only the Lake Erie location was tested for temporal change, and the two sampling years show a significant difference, with a large gain in alleles (+41 alleles; 84%; ST = , p<0.0001). This renders the 2007 Lake Erie samples more similar to those from the St. Lawrence River and from the Black Sea, possibly due to population mixing caused by additional introductions into Lake Erie (Fig. 4.2B, 4.3B). Discussion Multiple founding sources contributed to significant genetic structure among dreissenid mussel populations of both species across their North American invasive ranges, which likely reflect some maintenance of native population differences due to differential colonization and expansion. In addition, the relatively small founder effects shown in the invasive dreissenid populations highlight the ease of introducing these organisms in large numbers especially during their veliger larval stage. In both species, samples from Lake Erie appear genetically distinct, suggesting the likelihood of separate introductions from Eurasia. This appears especially likely given that both Mills et al. (1996) and Carlton (2008) found evidence of Lake Erie samples of both species (including shells on beaches, reports from divers on natural gas wellheads in northern Lake Erie, on water 111

126 treatment filters in Ontario, and on the hull of a fishing vessel) preceding the formal discovery of dreissenid mussels in the Great Lakes. Boileau & Hebert (1993) also found that the Lake Erie population of the zebra mussel appeared genetically distinct from other Great Lakes and European populations using allozymes (which may have been influenced by the cryptic presence of the quagga mussel). Like Boileau & Hebert (1993), the Lake Erie zebra mussels we sampled do not show strong connections with any of the analyzed European locations in the present study. However, Stepien et al. (2002) found that Lake Erie samples were most similar to those from Poland and the Netherlands using randomly amplified nuclear polymorphic DNA. Our samples from those regions grouped with the other Great Lakes populations, not with Lake Erie. Additional sampling and analyses from possible source regions thus is warranted. Previously invaded locations often serve as hubs for further spread of an invader (Bossenbroek et al. 2009; Floerl et al. 2009; Kelly et al. 2009), which appears likely for the zebra mussel. These genotypes may be pre-adapted for invasive success. For example, the bloody red shrimp Hemimysis anomala is another Ponto-Caspian invader whose previously invaded Northern European locations served as sources for their establishment in the North American Great Lakes (Audzijonyte et al. 2008). Likewise, the Chinese mitten crab has been introduced to the Great Lakes from several European ports, but has failed to establish (Tepolt et al. 2007). Similarly, the Eurasian ruffe Gymnocephalus cernuus invaded the Great Lakes from the Elbe River via the Baltic Sea region (Stepien et al. 2005). The Great Lakes populations then served as such a hub for 112

127 dreissenid expansion into the rest of North America (this study), and are predicted to serve a hub for future further expansion into other parts of the world (Kelly et al. 2009). In contrast to zebra mussels, North American quagga mussels appear to have arrived directly from their native Ponto-Caspian region, with their invasive populations tracing to the Southern Bug and Dnieper Rivers. Congruent with the present study, Stepien et al. (2002) analyzed nuclear RAPD markers to discern that quagga mussels from western Lake Erie were distinct from other North American quagga mussels, with all tracing back to a Dnieper River source (that study lacked the Southern Bug River samples we analyzed here). Similarly, the North American round goby Neogobius melanostomus invasion traced back to the Dnieper River (Brown & Stepien 2009). The Dnieper River watershed thus served as a common source population for invasive species that became successfully established in the Great Lakes. The geographic difference in founding sources discerned between the two dreissenid species suggests that zebra and quagga mussels each arrived in the Great Lakes in unrelated introduction events - the zebra mussels originating on ships from the Baltic Sea and northern Europe, and the quagga mussels on ships from the Black Sea. These correspond to two of the main invasion pathways out of the Ponto-Caspian identified previously (Ricciardi & MacIsaac 2000; Grigorovich et al. 2002). The observed heterozygosities we found for North American quagga mussels (average H O = 0.73) appears higher than the diversity Therriault et al. (2005) found for invasive quagga mussel locations along the Volga River in Europe (average H O = 0.50; although 113

128 the latter used fewer microsatellite markers than we did 6 compared to our 9). The number of alleles per population is similar between the two studies when populations are adjusted for sample size, suggesting that introductions in North America and the Volga River each had large founding populations. Neither species experienced a genetic bottleneck when it was introduced to North America suggesting that a large number of propagules were introduced into the Great Lakes, which likely contributed to their successful establishment. Likewise, the brown mussel Perna perna introduction into the Gulf of Mexico did not experience a bottleneck and successfully expanded its range following ballast water introduction from multiple sources in South America and southern Africa (Holland 2001). Similarly, the round goby did not undergo a genetic bottleneck during its introduction into the Great Lakes (Stepien and Tumeo 2006; Brown and Stepien 2009). All of these invasion examples thus were founded by large numbers of successful colonists. The high volume of ship ballast tanks facilitates massive introductions (Kelly et al. 2009), providing early life history stages of species introduced by this vector a good chance of establishment (Bax et al. 2003; Drake & Lodge 2004; Holeck et al. 2004). We also detected fine scale patterns of divergences among North American populations of both dreissenid mussel species. Their distinct population genetic structures likely resulted from long distance jump dispersal patterns from divergent sources via ballast water or recreational boating, followed by local spread. Many of these populations have high genetic variation, likely maintained by large founding numbers or continued 114

129 migration events from commercial transport and recreational boating traffic. The genetic bottlenecks shown in the western populations of both species suggest that they likely were founded by relatively small numbers of individuals, and that there has been little additional gene flow since. This fits the model of low-frequency transport via recreational boats predicted for dreissenid expansions into the western regions of North America (Johnson et al. 2001; Bossenbroek et al. 2007). This stands in contrast to the samples from the edges of the Great Lakes, which show evidence for a period of rapid growth, likely fueled by large introductions via commercial shipping. The significant difference between the genetic composition of the northern and southern Mississippi River locations of zebra mussels found here supports the gradient observed by Elderkin & Klerks (2001) using allozymes and Elderkin et al. (2004) using nuclear AFLP (amplified frequency length polymorphism) data, who attributed the observed genetic structure to boat-mediated introductions and the difficulty for free-spawning mussels to maintain consistent recruitment in lotic systems. Stepien et al. (2002) also found a marked genetic difference between the southern Mississippi River samples of the zebra mussel versus those from the Great Lakes. This genetic difference may lead to adaptive variation of genotypes in the south versus the north, which should reinforce neutral population genetic differences and merits further analysis of quantitative adaptive traits and their heritability. The significant temporal change in genetic structure of the tested populations suggests that there was some allelic turnover at those locations, and that some recruits contributed to the gene pool that originated from other areas. However, the colonies retained their 115

130 original population genetic structure, thus exhibiting genetic resilience. Haag & Garton (1995) found significant genetic differences among zebra mussel larvae and adults from western Lake Erie (the same location as our Lake Erie samples), indicating possible selection. Additionally, the temporal differences observed in our Hudson River samples of zebra mussel likely are linked to the cycling population dynamics observed in that system, where large year classes dominate recruitment on a 2-4 year cycle, with population size fluctuating over an order of magnitude (Strayer & Malcolm 2006). We detected evidence for rapid population expansion in our 1995 Lake Superior samples, but by 2006 the population had stabilized, and no such signal remained. These signals are ephemeral (Cornuet & Luikart 1996) and would be predicted to fade after a decade. Although we did not have same temporal sampling regime in other populations, there is additional evidence that 10 years is sufficient for zebra mussel populations to return to equilibrium after demographic events. When Boileau & Hebert (1993) examined the newly established Oneida Lake zebra mussel population, it had much lower heterozygosity than the other North American populations they examined. Our Oneida Lake sample a decade later has a heterozygosity level very similar to that of other Great Lakes sites, suggesting that the initial bottleneck has faded with time, diversifying with new recruitment. Similarly, an initial reduction in heterozygosity associated with a founder effect returned to equilibrium values within two decades of the Drosophila pseudoobscura invasion of New Zealand, although that invasive population unlike dreissenids - retained very low genetic diversity (Reiland et al. 2002). The 100 th Meridian Initiative 116

131 Western expansions populations of both dreissenid species did not originate from the closest geographic sources which would have been from either the Mississippi River basin or the southwestern Great Lakes. Instead, they appear to trace to population origins in the eastern Great Lakes. This represents a possible mixed success for the 100 th Meridian Initiative (Bossenbroek et al. 2009), with our study results suggesting that few individuals colonized from the Mississippi River drainage or the western Great Lakes, where 100 th Meridian control resources were concentrated (Mangin 2001). Instead, it appears that the successful colonists originated from areas much further east in Lake Ontario and the St. Lawrence River. This sort of long-distance dispersal event is hard to predict and observe (Buchan & Padilla 1999, Leung et al. 2006), but can be detected using genetic methods, as we have shown in this study. Under the 100 th Meridian Initiative, more easterly population areas may not have received the program s intensive education programs because the focus was on the frontier of the zebra mussel s distribution, rather than its core. The effectiveness of these education programs is supported by the apparent lack of spread from target regions in our data and by the slowing spread of infestations to inland lakes in areas with intense education campaigns (Johnson et al. 2006; J. Bossenbroek, personal communication, 2009). Thus, education efforts carried out to prevent the spread of dreissenid mussels likely were effective (Johnson et al. 2001) in those target areas but should have been more comprehensive. Conclusions 117

132 The dreissenid introductions to North America are similar in some ways to each other and to other Great Lakes invasions, but show key differences. Both quagga and zebra mussels were introduced from multiple source locations in Europe. Zebra mussels that became established in North America colonized from multiple, previously invaded locations within northern Europe and the Baltic Sea region. Quagga mussels instead colonized directly from native Southern Bug and Dnieper River estuaries, and preserved these historic genetic differences among their colonized regions in the Great Lakes. The invasive populations of both species are genetically heterogeneous across their ranges, showing considerable population genetic structure across North America. Genetic diversity levels in both species reflect no overall founder effects comparing the Great Lakes to likely Eurasian source locations, indicating very high numbers of introduced individuals. Zebra mussel locations outside the Great Lakes have similar levels of genetic diversity compared to those within the Great Lakes, whereas quagga mussel locations show reduced genetic diversity in newly established western populations. Once established, the genetic composition of given populations genetically changed over time, but retained their original genetic signature; suggesting external recruitment and long distance gene flow contribute to population genetic diversity. Since their genetic signatures remained distinctive, showing closest assignment with themselves, the genotypes that become established first retain genetic population resilience over time; leading to differentiation among sites. These population factors thus significantly shape the genetic identity and provide the raw material governing the genetic adaptation of the invasion. 118

133 Tables and Figures Table 4.1 Locations of A. zebra mussel and B. quagga mussel samples, with latitude, longitude, date of first sighting, sample size (N), microsatellite number of alleles (N A ), heterozygosity measures (H E, expected; H O observed), F IS. Year of first discovery obtained from USGS ( for North America, and DAISIE ( for Europe. A. Watershed/Region Sampling Location Latitude Longitude Year discovered N N A H E H O F IS Western expansion 1. San Justo L., CA Miss. R. drainage 2. Walnut River, KS ~ Timber Creek, KS Upper Miss. R. 4. Lake Pepin, WI Lower Miss. R. 5. s. Miss. R., LA Ohio R. drainage 6. Tippecanoe R., IN Great Lakes L. Superior 7. Duluth, MN L. Michigan 8. Mackinac Straits, MI L. Huron 9. Alpena, MI L. Erie 10. Gibraltar Is., OH L. Ontario 11. Cape Vincent, NY Erie Canal 12. Clyde, NY Oneida Lake 13. Bridgeport, NY St. Lawrence R. 14. Becancour, QC

134 Hudson R. 15. Catskill, NY Mediterranean S. drainage 16. Ebro R., ESP North S. drainage 17. L. Ijsselmeer, NLD Baltic S. drainage 18. Piasnica R., POL Black S. drainage 19. Danube R., HUN Dnieper R., UKR Native Caspian S. drainage 21. Volga R., RUS Native B. Watershed/Region Sampling Location Latitude Longitude Year DiscoveredN N A H E H O F IS Western expansion A. L. Matthews, CA B. L. Mead, NV Great Lakes L. Michigan C. Grand Haven, MI ~ L. Huron D. Charity Is., MI ~ L. Erie E. Gibraltar Is., OH ~ L. Ontario F. Olcott, NY ~ G. Cape Vincent, NY ~ Oneida Lake H. Bridgeport, NY ~ St. Lawrence R. I. Montreal, QC ~ Black Sea J. Bug R., UKR Native K. Dnieper R., UKR Native Caspian Sea L. Volga R., RUS

135 Table 4.2 Microsatellite primers used for amplification of A. zebra mussel, B. quagga mussel DNA, with source, annealing temperature (TA), the range of repeat numbers (R N ) and sizes (R S ; bp), number of alleles (N A ), average observed heterozygosity (H O ), and average F IS and F ST for each microsatellite locus. A. Locus Source T A ( o C) R N R S (bp) N A H O F IS F ST Dpol A6 Naish and Boulding Dpol B Dpol B Dpo04 Feldheim et al. in review Dpo Dpo Dpo Dpo Dpo Dpo Dpo B. Locus Source T A ( o C) R N R S (bp) N A H O F IS F ST Dbug1 Wilson et al Dbug

136 Dbug Dbu74 Feldheim et al, in review Dbu Dbu Dbu Dbu Dbu

137 Table 4.3 Pairwise tests for A. zebra mussel and B. quagga mussel locations. Below diagonal = F ST, above diagonal = contingency test. * = p<0.05, = sig. after Bonferroni correction (Rice 1989), NS = not significant. A ' ' San Justo L., CA ~ 2. Walnut River, KS Timber Creek, KS Lake Pepin, WI s. Miss. R., LA Tippecanoe R., IN Duluth, MN Mackinac Straits, MI Alpena, MI Gibraltar Is., OH Cape Vincent, NY Clyde, NY Bridgeport, NY Becancour, QC Catskill, NY Ebro R., ESP L. Ijsselmeer, NLD Piasnica R., POL Danube R., HUN ~ NS ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 123

138 Dnieper R., UKR ~ 21. Volga R., RUS ~

139 B. A B C D E '02 E '07 F G H I J K L A. L. Matthews, CA ~ B. L. Mead, NV ~ C. Grand Haven, MI ~ D. Charity Is., MI ~ E. Gibraltar Is., OH ~ ~ F. Olcott, NY ~ * G. Cape Vincent, NY ~ H. Bridgeport, NY ~ I. Montreal, QC ~ J. Bug R., UKR ~ K. Dnieper R., UKR ~ L. Volga R., RUS ~ 125

140 Table 4.4 Analysis Of MOlecular VAriance (AMOVA; Excoffier et al. 1992) results showing partitioning of genetic variation within and among A. zebra mussel and B. quagga mussel sampling sites using microsatellites (μsat). Φ-statistics (F-statistic analogues) significant at P < A. Comparisons among five primary population groups: (1-3, 11, 14); (12, 15-19, 21); (4-7 '95, 13, 15 '03); (10, 20); (7 06, 8, 9). B. Comparisons among three primary quagga mussels population groups: (A, B, F, G, J); (C, D, H, I, L); (E, K). A. Division of variation partitioning tested Measure of variation partitioning Among 5 primary population groups: Proportion of variation 0.06 (1-3, 11, 14); (12, 15-19, 21); (4-7 '95, 13, 15 CT 0.06 '03); (10, 20); (7 06, 8, 9) Among sampling sites within the groups Proportion of variation 0.08 SC 0.08 Within sampling sites Proportion of variation 0.86 ST 0.14 B. Division of variation partitioning tested Among 3 primary population groups: (A, B, F, G, J); (C, D, H, I, L); (E, K) Measure of variation partitioning Proportion of variation 0.02 CT 0.05 Among sampling sites within the groups Proportion of variation 0.10 SC

141 Within sampling sites Proportion of variation 0.89 ST

142 Table 4.5 Assignment test using GENECLASS2 (Piry et al. 2004). A. Zebra Mussels, B. Quagga Mussels. Self-assignment is in bold down diagonal, percent correctly assigned is on the right. Most individuals assigned to population of origin, and those that did not usually assigned to a nearby location. A ' ' % Correct 1. San Justo L., CA Walnut River, KS Timber Creek, KS Lake Pepin, WI s. Miss. R., LA Tippecanoe R., IN Duluth, MN Mackinac Straits, MI Alpena, MI Gibraltar Is., OH Cape Vincent, NY Clyde, NY Bridgeport, NY Becancour, QC Catskill, NY Ebro R., ESP L. Ijsselmeer, NLD Piasnica R., POL Danube R., HUN Dnieper R., UKR Volga R., RUS

143 B. A B C D E '02 E '07 F G H I J K L % Correct A. L. Matthews, CA B. L. Mead, NV C. Grand Haven, MI D. Charity Is., MI E. Gibraltar Is., OH F. Olcott, NY G. Cape Vincent, NY H. Bridgeport, NY I. Montreal, QC J. Bug R., UKR K. Dnieper R., UKR L. Volga R., RUS

144 Figure Legends Fig. 4.1 Maps of current range and sampling locations for zebra mussels and quagga mussels. A. North America (USGS; B. Eurasia (DAISIE; Fig. 4.2 Three dimensional factorial component analysis (GENETIX v.4.05; Belkhir et al. 2004) for A. zebra mussels and B. quagga mussels. A. Zebra mussels in L. Erie were distinct from all other samples. B. Quagga mussel samples separated into 3 distinct groups, with western North American samples clustering with those from L. Ontario. Fig. 4.3 Genetic distance tree among A. zebra mussel and B. quagga mussel population sites constructed in PHYLIP v3.6 (Felsenstein 1989) with Cavalli-Sforza chord distances (Cavalli-Sforza and Edwards, 1967) based on microsatellite data. Bootstrap percentage values from 1000 pseudoreplicates. Branch lengths are proportional to genetic divergence. Fig. 4.4 Bayesian STRUCTURE analysis (Pritchard et al. 2000; Pritchard and Wen 2004) for A. zebra mussels and B. quagga mussels. Analysis discerned 5 groups (posterior probability = 0.91) for zebra mussels and 3 groups (posterior probability = 0.89) for quagga mussels. Both species showed evidence of multiple introductions. Recent populations established in western North America link to the eastern portions of the Great Lakes. 130

145 Fig. 4.1A 131

146 Fig. 4.1B 132

147 Fig. 4.2A. 133

148 Fig. 4.2B 134

149 Fig. 4.3A. 135

150 Fig 4.3B. 136

151 Fig. 4.4A. 137

152 Fig 4.4B. 138