UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Marine population structure in an anadromous fish : life-history influences patterns of neutral molecular… McLean, Jennifer Elizabeth 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1999-0371.pdf [ 6.77MB ]
Metadata
JSON: 831-1.0089038.json
JSON-LD: 831-1.0089038-ld.json
RDF/XML (Pretty): 831-1.0089038-rdf.xml
RDF/JSON: 831-1.0089038-rdf.json
Turtle: 831-1.0089038-turtle.txt
N-Triples: 831-1.0089038-rdf-ntriples.txt
Original Record: 831-1.0089038-source.json
Full Text
831-1.0089038-fulltext.txt
Citation
831-1.0089038.ris

Full Text

Marine Population Structure in an Anadromous Fish: Life-history Influences Patterns of Neutral Molecular Variation in the Eulachon, Thaleichthys pacificus by Jennifer Elizabeth McLean B . S c , The University of British Columbia, 1996 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S FOR T H E D E G R E E O F M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Zoology We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A July 1999 © Jennifer Elizabeth McLean, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, 1 agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date Owl DE-6 (2/88) Abstract Phylogeography, historical demography, conservation of biodiversity and the influence of life-history on intraspecific population structure may be investigated with molecular genetic data, in particular, mitochondrial DNA (mtDNA) and microsatellites. Due to the apparent decline in size of a number of populations, eulachon, Thaleichthys pacificus, have recently become the focus of a conservation movement in the northeast Pacific. Little is known of the marine life-history phase of this anadromous fish, and although it has been suggested that eulachon spawning in different rivers may form distinct populations, nothing is known of their population structure. Both mtDNA and microsatellite loci were employed to investigate population structure and possible management schemes. Mitochondrial DNA haplotypes, determined through RFLP analysis of four mitochondrial genes, were resolved in fish from several rivers throughout the geographic range of eulachon. My data support the idea that extant eulachon populations result from post glacial dispersal from a single Wisconsinan glacial refuge. Further, while three of the thirty-seven haplotypes recovered account for approximately 79% of the samples, many private haplotypes were observed, suggesting possible regional population structure. While a great deal of genetic variation was observed (37 haplotypes in 315 samples), an AMOVA showed that more than 97% of the total variation was detected within populations. Eulachon populations seem to be structured in a manner much more similar to marine species than anadromous species, which may be due to some unusual life-history characteristics. MtDNA results were tested against predictions made from hypotheses concerning the origin and persistence of sub-divided populations in marine species, and seem to be consistent with both the Member-Vagrant hypothesis and isolation-by-distance. Analysis of five microsatellite loci revealed that microsatellite DNA variation in eulachon is surprisingly low, yet a similar degree of population sub-division as seen with the mtDNA data was observed. Again, this level of structure was consistent with what is generally observed in marine species. Isolation-by-distance was not supported, but mtDNA data are more sensitive to population sub-division and expected to reveal patterns that nuclear DNA might not. Microsatellite data were consistent with a single Wisconsinan glacial refuge in the south. The extremely low variation observed at eulachon microsatellite loci may be due to the nature of the repeat arrays. Only 16% of the microsatellites sequenced contained perfect repeats: a small proportion compared to an average of 60% or higher in salmonids. Although eulachon numbers are declining throughout their range and especially in the southern portion of their distribution, levels of mtDNA diversity are high and microsatellite variation is consistent throughout their range. Recent demographic declines have apparently had little effect on the genetic diversity of eulachon. Despite the adequate levels of genetic diversity, population sub-division is only weakly developed in eulachon. This species represents a common marine situation of high gene flow and indeterminate population boundaries in which the definition of management units is difficult, and factors other than genetic variation must be considered in management or conservation plans. iv Table of Contents Abstract ii Table of Contents iv List of Figures vi List of Tables vii Acknowledgements ix Chapter 1: General Introduction 1 1.1 Population sub-division: patterns and processes 1 1.2 Population structure and conservation 3 1.3 Population structure and conservation in eulachon 4 1.4 Investigation of population structure in fish 9 1.5 Objectives of this thesis 15 Chapter 2: Phylogeography and micro geography of eulachon, Thaleichthys pacificus, as revealed by mitochondrial DNA analysis 16 2.1 Introduction 16 2.2 Materials and Methods 23 2.21 Sample Collection 23 2.22 DNA Extraction 23 2.23 PCR Amplification 25 2.24 Restriction Enzyme Digestion 26 2.25 Scoring of Variation 27 2.26 Data Analysis 27 2.3 Results 30 2.31 Mitochondrial DNA Restriction Fragment Variation 30 2.32 Mitochondrial DNA Haplotype Variation and Distribution 30 2.33 Rare and Private Hap lo types 34 2.34 Mitochondrial DNA Diversity Within and Among Sample Sites 35 2.35 Temporal Population Structure 35 2.36 Spatial Population Structure 36 2.37 Isolation-by-distance 38 2.38 Historical Demography 39 2.4 Discussion 40 2.41 Glacial Refugia and Post-glacial Dispersal of Eulachon 40 2.42 Genetic Population Structure of Eulachon 44 2.43 Life-history and Mechanisms of Population Sub-division in Eulachon 48 V Chapter 3: Characterization and analysis of five microsatellite loci in the eulachon, Thaleichthys pacificus, resolves unusually low levels of genetic variation 53 3.1 Introduction 53 3.2 Materials and Methods 60 3.21 Isolation and Cloning of 300-700 bp Fragments 60 3.22 Detection and Amplification of Positive Colonies 60 3.23 Sequencing 61 3.24 Primer Design 62 3.25 Primer Optimization 64 3.26 Primer End-labeling 64 3.27 Samples used and PCR Amplification 64 3.28 Electrophoresis and Scoring of Variation 67 3.29 Data Analysis 67 3.3 Results 73 3.31 Microsatellite Polymorphism 73 3.32 Microsatellite Allele Frequency Distribution 73 3.33 Heterozygosity 80 3.34 Hardy-Weinberg Equilibrium 81 3.35 Linkage Disequilibrium 81 3.36 Microsatellite Diversity Within Sample Sites 82 3.37 Microsatellite Diversity Among Sample Sites 83 3.37.1 Temporal Population Structure 83 3.37.2 Spatial Population Structure 85 3.38 Isolation-by-distance 88 3.39 Recent Population Bottlenecks 88 3.4 Discussion 90 3.41 Glacial Refugia and Post-glacial Dispersal 90 3.42 Eulachon Microsatellite DNA Variability and Genetic Population Structure 91 3.43 Life-history and Mechanisms of Population Sub-division 98 Chapter 4: General Conclusions and Conservation Implications 101 4.1 Conclusions from mtDNA and five microsatellite loci 101 4.11 Glacial Refugia and Post-glacial dispersal 101 4.12 Genetic Variability and Population Structure 101 4.13 Life-history and Mechanisms of Population Sub-division 103 4.2 Conservation implications 104 Literature Cited 107 vi List of Figures Figure 1.1 Geographic distribution of eulachon, Thaleichthys pacificus. 7 Figure 2.1 Sample sites. Sample designations are as follows: BR, Bering Sea; CK, Cook Inlet; NS, Nass River; GC, Gardner Canal; KK, Klinaklini River; FN, Franklin River; FR, Fraser River; QT, Queets River; COL, Columbia River 1997 and 1998, and Cowlitz River 1997 and 1998. 22 Figure 2.2 Mutational network showing evolutionary relationships among haplotypes. Each ellipse represents a haplotype, the lines connecting the haplotypes represent one mutation separating the haplotypes and the cross hatches represent additional mutations between the haplotypes. The size of each haplotype's ellipse is proportional to its relative abundance in the total sample. 32 Figure 2.3 Isolation-by-distance of eulachon sample sites. 38 Figure 2.4 Proportion of groups I (B-like haplotypes), represented by the lighter portion of the pies and II (A-like haplotypes), represented by the darker portion of the pies, at each sampling site. 50 Figure 3.1 Microsatellite gels. One representative gel from each locus is shown. Each lane represents an individual fish, and sample site designations are shown above the samples. 70 Figure 3.2 Allele frequencies at each sample site for each of five loci. For exact values, see Table 3.3. 77 Figure 3.3 Isolation-by-distance of eulachon sample sites. 89 List of Tables VI1 Table 2.1 Collection locations and descriptions, and mtDNA haplotype diversity and nucleotide diversity within sample sites. Haplotype diversity is estimated from the number of haplotypes and their frequencies within a putative population. Nucleotide diversity is an estimate which includes both the frequencies of haplotypes and the divergences among them. 24 Table 2.2 Composite haplotypes. The first column lists the composite haplotypes referred to in the text. Each letter in the following columns represents the banding pattern seen on the gel for the enzyme specified at the top of the column. 29 Table 2.3 Distribution of haplotypes at each sampling location. Sample designations are as in Table 2.1. 33 Table 2.4 Probabilities of monte carlo tests for population differentiation. The first column lists the "population" pairs tested, which were chosen based on geography. The second column is the P-value given by the test, and the third column is the Bonferroni corrected alpha, which is adjusted for multiple comparisons. 37 Table 2.5 Results of AMOVA test for partitioning of eulachon mtDNA variance. Collections were grouped geographically into 7 regions based on the chi-square tests for temporal and geographic heterogeneity. The regions are: Columbia and Cowlitz River (both years), Queets River, Fraser River, Knight Inlet, Gardner Canal, Nass River and Cook Inlet/Bering Sea. 37 Table 3.1 Eulachon microsatellite loci. Given for each locus are the locus name, repeat array, range of allele sizes and number of alleles. Repeat arrays are given for the species in which each microsatellite was developed. 63 Table 3.2 Collection locations and descriptions, including sample sizes. 65 Table 3.3 Primer sequences (5' to 3') and PCR annealing temperatures and MgCL; concentrations for on microsatellite loci. Asterisks (*) denote P-labelled primers. Primers sequences are unavailable for Osmo loci. 65 Table 3.4 Allele frequencies at each sample site for each of five microsatellite loci. 74 Table 3.5 Observed and expected heterozygosities for each microsatellite locus. Observed heterozygosities are direct counts, and expected heterozygosities are corrected, unbiased estimates. 80 Table 3.6 Probabilities for chi-square tests (Fisher's method) of linkage disequilibrium. Tests were performed for each locus pair (first column) across all sample sites. No tests were significant and no locus pairs exhibited linkage disequilibrium. 82 Table 3.7 Gene diversity and average gene diversity over loci for each sample site. Gene diversity is the expected heterozygosity summed over loci, and average gene diversity is the expected heterozygosity averaged over loci. 83 Table 3.8 Exact tests of population differentiation. Al l possible pairwise comparisons were performed (across loci) and P-values were compared to Bonferroni corrected alpha values. Asterisks (*) denote significant tests. The final column lists the number of loci that support the significant differences at the Bonferroni corrected alpha level. For each significant test, all five loci were individually supportive of the results at an alpha of 0.05. Sample designations are as in Table 3.2. 84 Table 3.9 Results of A M O V A test for partitioning of eulachon microsatellite D N A variance. Collections were grouped geographically into 7 regions based on the chi-square tests for temporal and geographic heterogeneity. The regions are: Columbia and Cowlitz River (both years), Queets River, Fraser River, Knight Inlet, Gardner Canal, Nass River and Cook Inlet/Bering Sea. 86 Table 3 .10 F S T and R S T values for each locus and overall. P-values show the probability that the F S T and RST values are significantly different from zero. "NT" means that the significance of the test was not determined. 87 Acknowledgements Without the encouragement and support of a number of people, my journey through this Masters thesis would not have gone as smoothly as it did. I would like to thank everyone who played a role in helping me achieve this goal. I would first like to thank my committee members — Doug Hay, Bob Devlin and Don McPhail — for their commitment to my success and advice throughout the project. Special thanks goes to my supervisor, Eric Taylor, both for introducing me to the wonderful world of molecular biology, and for his continual enthusiasm, guidance and confidence in me. I would also like to thank all of the other students associated with the Taylor-McPhail lab. By pooling our experiences with technical difficulties, and each of our limited knowledges of topics ranging from AMOVA to zoogeography, we've created an environment in the lab in which the total is greater than the sum of the parts. Lastly, I would like to thank my friends and family who have had to listen to me talk about nothing but eulachon for the past three years, and who have been extremely supportive in spite of it. Thank you. 1 Chapter 1 General Introduction 1.1 Population sub-division: patterns and processes Intraspecific population structure, or at least some degree of genetic differentiation among geographic locations, is exhibited by almost all species (Avise 1994). Plants, invertebrates and vertebrates all demonstrate examples of this phenomenon, which is often dependent on life-history characteristics as well as environmental factors. Fishes, the largest group of vertebrates, display various degrees of among-population differentiation, and provide examples in many different habitats of intraspecific diversity partitioned with respect to geography (Avise et al. 1987). The number and location of genetically distinct populations within a species, however, can vary enormously and depends on environmental factors and life-history traits (Avise 1994). Recent studies have made use of molecular genetic data to investigate issues such as phylogeography, historical demography, conservation of biodiversity and the influence of life-history on intraspecific population structure (Palumbi 1994, Nee et al. 1995, Rogers 1995, Avise 1998a, Avise 1998i>, Newton et al. 1999). Numerous DNA markers have been developed over the last 15 years, and questions concerning biogeography and microgeography are readily addressed with molecular techniques. In particular, mitochondrial DNA (mtDNA) and microsatellites have been useful in the investigation of the extent and distribution of intraspecific genetic diversity (Avise et al. 1987, Gold and Richardson 1998, Graves 1998, Estoup et al. 1998a, Shaw et al. 1999). Phylogeography investigates events in recent evolutionary history and determines 2 their role in creating current patterns of intraspecific variation in phylogenetic characters (Avise 1998a). Range expansion, range fragmentation and population bottlenecks may be induced by large-scale environmental events such as formation of land-bridges, glaciation, and sea level changes, and these population processes leave characteristic genetic patterns. Phylogeography links patterns in the current distribution of intraspecific phylogenetic variation to these historical events. Not only may the patterns of genetic variation seen in species today be connected to known historic geologic events, but the consequences of these events on effective population sizes may also be investigated. The examination of current genetic data, in particular the distribution of pairwise DNA sequence differences in mtDNA (Slatkin and Hudson 1991, Rogers 1995, Nee et al. 1995), allows the inference of historical demographic changes in populations. Coalescence theory developed for non-recombining DNA, such as mtDNA, has been useful in approximating population size fluctuations (Hudson 1990, Nee et al. 1995). Environmental factors, both past and present, play a large role in organizing population structure by creating or eliminating connections among populations. Historical environmental events relevant to the creation and maintenance of distinct fish populations include large-scale events such as glaciation, and present geographic barriers include thermoclines, nutrient poor currents, ocean gyres or separate drainages or lakes (Sinclair and lies 1989). Depending, in part, on whether they are freshwater, anadromous or marine species, fish species differ not only in their ability and potential for dispersal, but also in the manner in which their neutral genetic variation is partitioned. For instance, 3 Gyllensten (1985) summarized allozyme data and showed that the average amount of genetic variation partitioned among populations of fish was approximately 29.4%, 3.7% and 1.6% for freshwater, anadromous, and marine fish, respectively. Apparently, the lack of physical barriers in the ocean allows a great deal of mixing between fish from different locations, however, marine species do exhibit some genetic structure (Palumbi 1992, Palumbi 1994, Palumbi et al. 1997). How are marine species structured into populations when few physical barriers to reproduction exist in the ocean? Life-history strategies play a significant role in determining genetic variability and population structure of marine fish species (Jaglielo et al. 1996). Palumbi (1994) suggested five causes for marine species population sub-division: invisible barriers (such as oceanic currents), behavioural limits to dispersal, isolation-by-distance, selection and recent history. Although these hypotheses were originally proposed for marine fish, it may be possible to extend them to anadromous species with life-histories and/or population structures very similar to those of marine fish. In this thesis, I examine these marine population structure hypotheses with respect to the genetic population structure, and its implications for conservation, in the eulachon, Thaleichthys pacificus. 1.2 Population structure and conservation Molecular approaches to conservation include the assessment of current levels of genetic variation within species, and within populations of species. Two types of units for conservation have been defined in terms of genetic variation: evolutionarily significant units (ESUs) and management units (MUs) (Moritz 1994). Deep divergences 4 caused by historic population processes are seen in populations or groups of populations that have been isolated on an evolutionary time scale (as seen among ESUs) , while shallower divergences (MUs) are caused by population processes operating on an ecological time scale. B y identifying where and how deeply genetic diversity within a species is partitioned, decisions regarding conservation and management can be achieved. 1.3 Population structure and conservation in eulachon The smelt family, Osmeridae, comprises six genera and ten species which cover a range of life-history types, including freshwater, anadromous and marine (McAllis ter 1963). Osmerids have a boreal-subarctic distribution, and are found throughout the Pacific, the Atlantic, and the Arctic oceans and their drainages. Because all species and genera are found in the Pacific, and the majority of species are found in the eastern Pacific, this region is presumed to be the center of origin for osmerids (McAllis ter 1963). The eulachon, Thaleichthys pacificus, belongs to a monotypic genus and has an anadromous life-history. This species is distributed along the northeast Pacific coast of North America (Figure 1.1), ranging from the southern Bering Sea to the Columbia River (Hay 1996), and may spawn in only 20 or 30 rivers throughout this range (Hay 1996). Although little is known or understood of the basic biology of eulachon, spawning habits have been examined. The spawning migration occurs in the spring from mid-March to mid-May, and spawning takes place a short distance upriver in areas where the bottom is covered in coarse sand (Hart and McHugh 1944). Eulachon appear to prefer glacial-fed rivers (Hay 1996). Fertilized eggs attach to the river bottom until hatching, which occurs two to three weeks later, depending on water temperature (Shepherd and Vroom 1977). 5 Larvae are passively transported downstream to the ocean, and often leave their river of origin within 24 hours of emergence. Although eulachon are an anadromous fish, certain life-history characters are similar to those of marine fish, especially those factors important in influencing the depth of population sub-division. Eulachon are currently a conservation concern, especially in the southern part of their range. Eulachon returns have apparently been declining since the 1950's (Hay 1996). In 1994, population crashes in a number of rivers were reported (Berry 1996, Gordon 1996, Lewis 1996). In March 1997, the eulachon fishery on the Fraser River was closed due to concern over the poor returns (Lee 1997), and has remained closed ever since. In the past two years, very few eulachon have returned to spawn in any river south of Alaska (Hay, pers. comm.). Eulachon runs have historically been very large, with catches of 2 to 3 million pounds not unusual in the Columbia River (Anon. 1997). Today, typical runs are described as "negligible," or even "no show" (Hume 1999). It is thought that the total number of eulachon may have declined from historical population sizes by up to 90% (Anon. 1998). Because of these reasons, it has been proposed that eulachon may deserve threatened or endangered status (Anon. 1998). Not only are eulachon of concern to conservationists, but in British Columbia, these fish have extraordinary cultural significance. Known as candlefish or salvation fish to British Columbia's coastal aboriginal people, these small oily fish represent an historically important food source, tradition and reason for ceremony, and status, wealth and trade. Because eulachon spawn so early in the spring, they were the first food source to come up-river, often returning just as stored winter foods were running out. Eulachon were smoked, salted and eaten fresh, but most importantly were used for rendering oil. 6 This oil was used in a number of ways, including as a condiment for various foods: "coastal native food without eulachon grease is like fries without ketchup" (Anon. 1998). Not only was the end result of the rendering important, but the actual process was meaningful as well, ".. .much like Christmas or Thanksgiving is to European Canadian culture" (Anon. 1998). In fact, separate eulachon villages were built especially for this process. When the rendering process was complete, the oil was used as a condiment for nearly every food, a gift, a status symbol, as the central component of a celebration, and as an important trade item. "Grease trails" connecting the BC coast with the interior formed trade routes for this precious commodity. It was by way of these grease trails that Alexander Mackenzie reached the BC coast. Eulachon are a fascinating species because life-history traits and meristic/ morphological evidence appear to contradict one another in terms of population structure. Historical evidence is consistent with what seems to be one large, panmictic population of eulachon. For instance, there are reports of eulachon spawning in certain rivers for tens of years and then not entering these rivers for a time, only to return a number of years later (Stacey 1998). By contrast, Hart and McHugh (1944) suggested that eulachon populations do not mix, as meristic traits, such as number of vertebrae, differ among rivers. Berry (1996) noted that of the two rivers sampled in Knight Inlet, Franklin River eulachon produce a poor quality red grease and their spawning migration peaks one week earlier than the Klinaklini eulachon, which produce a higher quality, less coloured grease. Many other species of fish with anadromous life-histories, especially those from the family Salmonidae, exhibit morphological and genetic differentiation among populations spawning in different locations (Taylor 1991). 7 Figure 1.1 Geographic distribution of eulachon, Thaleichthys pacificus. 8 The meristic characters which suggest eulachon populations may be structured in a manner similar to those of salmonids, however, may be under environmental regulation (Lindsey 1981). As well, certain life-history characteristics of eulachon differ from those of salmonids: the amount of time spent in freshwater after hatching is much shorter in eulachon, and while it has been determined that salmonids home (Quinn 1993), it is not known whether eulachon home, and if they do, what amounts of straying occur. If eulachon home to their natal rivers to spawn, then it is possible that a number of distinct populations exist. Homing by fish to their natal rivers, repeated over generations, limits gene flow and allows the accumulation of genetic differences among populations (Avise 1996). If eulachon do not spend enough time in freshwater to imprint, then they may have an opportunity to imprint in the ocean. Homing would then be to a coastal region rather than a river. If eulachon exiting a number of adjacent rivers entered the ocean and were retained together in a marine zone, then fewer populations than spawning rivers would exist. In addition to uncertainties concerning the microgeography of eulachon, the influence of recent glaciation on the phylogeography of eulachon is also unknown. McPhail and Lindsey (1970) hypothesized that eulachon were isolated in a single refuge during the last glaciation. Their current distribution is consistent with the idea that eulachon survived in the Pacific refuge south of the ice sheet, probably in the Columbia River system (McPhail and Lindsey 1986). Eulachon also exist, however, in areas (Bristol Bay, north coastal British Columbia) that were colonized fromBeringia (e.g. Oncorhynchus nerka: Taylor et al. 1996, Gasterosteus aculeatus: Thompson et al. 1997) and the possibility exists that they survived in Beringia or coastal refugia as well (Wood 1995, Taylor et al. 1996). MtDNA has confirmed the existence of two glacial refugia for another smelt, Osmerus mordax, which is found in northeastern North America (Bernatchez 1997). It is important to determine whether distinct populations of eulachon exist, and, if so, to characterize these populations. Fisheries management schemes in part are reliant on accurately defined population structure: continual exploitation of fish from mixed populations can lead to marked declines in single, less productive populations (Ovenden 1990, Schweigert and Withler 1990). As well, if distinct populations exist in different rivers, habitat degradation and declines in water quality at specific locations become significant to the continuation of these distinct populations. Conversely, populations should not necessarily be managed locally (ie. river by river) if dynamics are driven by a larger group of fish. Characterization of levels of microgeographic and phylogenetic variation within species is essential for conservation, management and the preservation of evolutionary and ecological potential (Ryman and Utter 1987). 1.4 Investigation of population structure in fish How can distinct populations within a species be identified and characterized? Since the early recognition of intraspecific variation, scientists have attempted to answer this question. It has long been recognized that within many species exist a number of distinct stocks or populations (Avise 1994). In certain cases, freshwater fishes for example, discrimination of different populations is straightforward. Distinct populations exist in unconnected lakes or drainages with clear interpopulation boundaries. The same is true for populations of organisms that exist on islands and that cannot access other such 10 populations within their species. When populations are separated in this manner, they are reproductively isolated, and have the opportunity to follow different evolutionary trajectories. With unambiguous boundaries, the identification of distinct populations is simple, and various traits within populations can be measured and compared to those traits in other populations. In species which consist of populations separated by indeterminate boundaries, such as marine species or terrestrial species with a high potential for dispersal, individual traits can be measured and populations can be defined based on geography and similarity of those traits. The choice of characters which can be used to study differences among populations within a species has increased dramatically since early studies of genetic population structure in the 1930s. Previous to the genetic approach, and often used today as a complement to it, is the measurement of meristic and morphometric traits to compare populations. Meristic traits include counts such as the number of vertebrae, number of gill rakers, and various fin ray counts. Because these traits have a genetic basis, many early studies used these measurements as an indirect estimate of genetic diversity within populations and genetic divergence among populations. It has since been shown, however, that many of these traits are subject to environmental variability, and therefore not appropriate for studies that attempt to measure intraspecific population structure (Lindsey 1981). The same problem of environmental regulation affects morphometric measurements, especially in fishes, which exhibit indeterminate growth capacity (Allendorf et al. 1987) and a high sensitivity to temperature variation which directly affects metabolic processes. An ideal character for intraspecific population structure 11 assessment is selectively neutral and therefore affected by population genetic processes including migration and random genetic drift, but excluding natural selection. The first selectively neutral markers used were an indirect assessment of genetic information. In the 1930s, human blood groups were used to describe intra- and interpopulation variation (Mourant 1956). Blood groups are an indirect measurement because, while they reflect information encoded at the D N A level, they are measured at the protein level. This method of assessing population structure worked well in humans and higher vertebrates, but when studies were attempted in other organisms, for example fishes, the results were not as good. Preliminary studies of blood groups in tuna and salmonids were accomplished, however, technical problems such as fish erythrocyte fragility precluded further blood group studies in fishes (Allendorf et al. 1987). The next selectively neutral markers developed to assess population structure were allozymes, which are another indirect assessment of genetic information. Protein electrophoresis was developed during the 1970s. This method of assaying genetic variation at the protein level makes use of the differences in net electrical charges of different versions of the same enzyme. These different versions are known as 'isozymes' when they are catalyzing reactions in the body, and 'allozymes' when they are run on electrophoretic gels. Detection of allozymes in an individual is accomplished by loading a solution of enzyme-containing tissue extract onto a polyacrylamide or starch gel, applying an electric current through the gel, and using specific histochernical staining to reveal the location in the gel of each form of the enzyme. A number of individuals can be run on the same gel, and a number of loci (each enzyme is a different locus) per gel can be examined. 12 The ease of obtaining a large amount of genetic information made allozymes a popular marker for population studies in the late 1970s and the 1980s, however, there are some limitations to this method. Sampling is generally lethal, and because allozyme electrophoresis relies on differences in electrical charge to separate different forms of an enzyme, a major drawback of this method is the high level of undetected variation. An assay at the protein level will not necessarily reflect changes that occur at the DNA level. For example, the redundancy in the genetic code means that each DNA base pair change does not always correspond to a change in the amino acid produced by that sequence. As well, many amino acids have the same electric charge, and it is possible that a substitution at the DNA level will result in a different amino acid, but one with the same charge. In both of these situations, the change would not be detected by allozyme electrophoresis. It has been estimated that two-thirds of the total amino acid substitutions that occur at loci assayed in allozyme studies go undetected in most investigations (Lewontin 1974). In order to detect higher levels of variation and achieve better resolution, direct assays of DNA sequences are necessary. In the late 1970s, the properties of mitochondrial DNA (mtDNA) were investigated, and by the mid-1980s, this molecule had become a leading marker in studies of intraspecific population structure. There are two methods of assessing mtDNA variation: restriction fragment length polymorphisms (RFLPs) and direct nucleotide sequencing. While nucleotide sequencing can reveal all base pair changes in the regions examined, it is much more time-consuming and costly than RFLP analysis, which can quickly survey the entire mitochondrial genome, but does not assay all nucleotides. Both of these methods, however, reveal higher levels of variation than does allozyme analysis. 13 MtDNA is a molecule well-suited to the study of intraspecific population structure for a number of reasons. The mitochondrial genome is a closed, circular molecule which is relatively small (approximately 15 000 to 18 000 base pairs) and can be easily isolated from nuclear DNA. Non-lethal sampling may be performed as only a small amount of tissue is required. This molecule contains little non-coding sequence, and 37 genes whose order is highly conserved across species. Of these 37 genes, 2 code for rRNAs, 22 code for tRNAs and 13 code for polypeptides. A non-coding, yet conserved, region is the d-loop, which is involved in replication (Borst and Grivell 1981). Greater resolution of genetic variation can be achieved with mtDNA than with allozymes because mtDNA has a comparatively rapid rate of evolution, and there is no hidden variation. Although each mitochondrial gene has its own mutation rate, the overall speed of evolution of the molecule is approximately 5 to 10 faster than for most nuclear DNA (Brown et al. 1979). Because the DNA sequence rather than protein sequence is being assessed, all base pair changes can be identified. The greater levels of resolution achieved with mtDNA allow investigations of more recent events (ie. post-glacial population sub-division) than does blood group or allozyme analysis. Another unique property of mtDNA is its haploid nature and maternal inheritance (Wilson et al. 1985). These properties allow investigation of maternal lineages and enhance the examination of population structure. Maternal lineages can be traced because there is less than 0.1% paternal contribution (Gyllensten et al. 1985). The ability to detect population structure is enhanced because at equilibrium, the effective population size of mtDNA is one-quarter that of nuclear markers, and therefore has a greater sensitivity to population divergence through random genetic drift (Birky et al. 1989). 14 Finally, because of the absence of recombination in this molecule (Wilson et al. 1985), evolutionary relationships between haplotypes (mitochondrial genotypes) may be inferred. This property allowed the development of the field of phylogeography (Avise et al. 1987), which describes the spatial distribution of haplotypes within and among populations when the evolutionary relationships of the haplotypes are known, and emphasizes historical aspects of current patterns (Avise 1998a). Recently, a number of studies have incorporated the use of microsatellites, a class of rapidly evolving, highly polymorphic repetitive nuclear DNA that consists of di-, tri-and tetra-nucleotide repeats (Beckmann and Weber 1992). Microsatellites have been able to resolve subtle levels of population structure that were not apparent at allozyme loci or using mtDNA in a variety of taxa (e.g. Bentzen et al. 1996, Estoup et al. 1998a, Shaw et al. 1999). Microsatellite loci are found approximately every 10 kilobases in the eukaryotic genome (Tautz 1989) and are flanked by unique non-repetitive regions in which primer sequences may be designed. Because microsatellites have a higher mutation rate and are more variable than other markers, they are presumably more sensitive than other nuclear markers to population genetic processes such as random genetic drift (Estoup et al. 1998a). This quality, combined with their selective neutrality, facilitates studies of intraspecific population structure, as well as hybridization, linkage mapping, paternity testing and pedigree analysis. Mutation rates for microsatellites are thought to be between 10"2 and 10"5 (Weber and Wong 1993), and mutation is presumed to occur through slipped-strand mispairing which causes variation in the number of repeat units at a particular locus (Schlotterer and Tautz 1992). Different alleles are identified by their different lengths, which are often resolved by electrophoresis on polyacrylamide 15 gels. Microsatellites are bi-parentally inherited and are exhibited in a codominant manner, and so, unlike haplotypes in mtDNA, two alleles are expressed in each individual. 1.5 Objectives of this thesis Due to the recent success of studies involving mtDNA and microsatellites in answering phylogeographic and microgeographic questions, I used both of these genetic markers to assess intra- and interpopulation diversity in eulachon in order to meet three objectives. First, because mtDNA is widely used for phylogeographic analyses, I used this marker to investigate previous theories of glacial refugia for eulachon (McPhail and Lindsey 1970). With mtDNA data, I tested the hypothesis that eulachon were isolated in a single refuge, the Columbia system, during the most recent glaciation. Second, I assessed the genetic population structure of eulachon throughout their geographic distribution. Both mtDNA and microsatellites have been used to resolve population structure in species with recent, post-glacial population sub-division. Third, because eulachon populations did not appear to be structured in a pattern typical of anadromous species, I used the information gained about population structure in combination with life-history information to investigate hypotheses describing the origin and maintenance of marine species population sub-division. 16 Chapter 2 Phylogeography and microgeography of eulachon, Thaleichthys pacificus, as revealed by mitochondrial D N A analysis 2.1 Introduction For many years, scientists have attempted to explain the present distributions of species, populations within species, and genetic variation within species by linking current patterns to historical processes (reviewed in Avise 1998a). Examinations of the phylogeography and historical demography of a particular species can lead to evolutionary explanations for its present-day intraspecific population structure, and thus will afford a reliable basis for conservation and management decisions. Further, these investigations can promote the understanding of the biology of the species in question, especially if life-history is poorly characterized. An excellent species through which to link evolutionary processes, intraspecific population structure and life-history are eulachon, Thaleichthys pacificus. Eulachon are currently a conservation concern, have a poorly described life-history, and little is known of the basic biology of this species (Hay 1996). A description of the extent and distribution of genetic variation in this species would contribute to both the characterization of current intraspecific population structure and the description of evolutionary processes important to this and other species with similar life-histories and geographic distributions. As well, conservation and management plans would be enhanced by a full description of intraspecific genetic diversity. Previous methods used to examine population sub-division and answer evolutionary questions include the measurement and comparison of meristic and 17 morphological traits, blood group proteins and allozymes. Each of these methods has contributed to the field of evolutionary biology, however, each of these methods has certain limitations. Meristic counts are prone to selection and tend to vary with environmental conditions such as water temperature (Lindsey 1981). It is difficult to use body measurements to characterize distinct populations of fish because fish have an indeterminate capacity for growth which is highly connected to environmental variables (Allendorf et al. 1987). Blood protein groups were found to be useful in humans and higher vertebrates, but technical limitations made them a poor marker for characterization of fish populations. Allozyme analysis has been extremely useful in the characterization of genetic variation in some species of fish, and was the main nuclear marker used for this purpose for over two decades (Hedrick 1999). Due to the high levels of undetectable variation and lethality of sampling procedures, however, allozymes are not ideal. Other molecular genetic data, particularly mitochondrial DNA (mtDNA), have recently been used successfully in studies of fish population characterization (Avise et al. 1987, Ovenden 1990, Schweigert and Withler 1990, Mulligan et al. 1992, Gold and Richardson 1994). There are a number of reasons for the utility of mtDNA in studies of population structure: sampling is non-lethal and only a small amount of tissue is needed; the mitochondrial genome is relatively small and can be isolated from nuclear DNA with little difficulty; it has a comparatively rapid rate of evolution which is necessary for defining recently diverged populations; it is haploid in nature and does not recombine; and it is predominantly maternally inherited (Wilson et al. 1985, Avise et al. 1987). These characteristics of mtDNA result in an effective population size at equilibrium that is reduced to one-quarter the effective population size of nuclear markers, and therefore 18 has a greater sensitivity to divergence of distinct genetic populations by random genetic drift (Birky et al. 1989). As well, evolutionary relationships between haplotypes may be inferred. The mitochondrial genome contains 37 genes, which, on average, evolve 5 to 10 times faster than most nuclear genes (Brown et al. 1992). Different mitochondrial genes evolve at different rates. Of these 37 genes, there are a number of regions typically used in studies of intraspecific population structure, each with an associated mutation rate. Generally, the mitchondrial regions used in studies of phylogeography and microgeography include the d-loop, which is a non-transcribed region that evolves rapidly compared to the rest of the mitochondrial genome; genes coding for NADH 5 and 6, which also evolve comparatively quickly; the gene which codes for cytochrome B; and more slowly evolving regions such as rRNA subunits 12 and 16. Regardless of the number of regions used in a mtDNA analysis study, the mitochondrial genome must be considered as one locus only, because recombination does not occur (Wilson et al. 1985). The lack of recombination of the mtDNA molecule affords unique possibilities for the study of historical demography. Evolutionary relationships between each haplotype pair can be inferred, and because genetic distances reflect the genealogical distances among individuals, mtDNA analysis allows characterization of historical population sizes and historical changes of those population sizes. Distinctions can be made among populations of constant size, exponentially growing populations, populations currently experiencing bottlenecks and populations of constant size that suddenly expand to a larger constant size (Slatkin and Hudson 1991). 19 At least three studies have addressed the concern that mtDNA may not reflect the same population structure patterns as those obtained by other methods. Ferris (1983) examined mtDNA in several subspecies of the mouse subgenus Mus, and corroborated the previous findings of Marshall and Sage (1981) who used protein electrophoresis and comparative morphological analysis to define groups. Avise and Smith (1974) examined allozymes in two subspecies of bluegill sunfish, Lepomis macrochirus, and discovered similar results with mtDNA ten years later (Avise et al. 1984). Avise et al. (1979) compared results from mtDNA and allozymes in pocket gophers, Geomys pinetis. Enough allozyme variation was discovered to distinguish eastern from western forms, however, mtDNA data was not only able to confirm the two forms, but also revealed an order of magnitude more variation and resolved the genetic relationships of groups within the two forms. Many recent studies have proved the utility of mtDNA in answering microgeographic questions (Bernatchez and Dodson 1990, Mulligan et al. 1992, Gold et al. 1993, Bremer et al. 1995), and its unique ability to anwer phylogeographic questions (reviewed in Avise 1998a). Mitochondrial DNA variation was examined in anadromous populations of cisco (Coregonus artedii) in James and Hudson Bays to determine if fish originating in different rivers were reproductively isolated from one another (Bernatchez and Dodson 1990). Not only were significantly discrete stocks found in Hudson Bay, but two distinct mitochondrial lineages, representative of populations from two glacial refugia, were found to exist in this area. Mulligan et al. (1992) assayed mtDNA in walleye pollock (Theragra chalcogramma) from the eastern Bering Sea and Shelikof Strait. The results suggested 20 the presence of several distinct populations in that region, in contrast to the allozyme results of Grant and Utter (1980). The allozyme study suggested that only one population exists in the southeastern Bering Sea. A mtDNA analysis of red drum {Sciaenops ocellatus) from the Gulf of Mexico and the Atlantic Ocean revealed weak population sub-division and a high level of gene flow in this species. Calculations of evolutionary effective female population size indicated that long-term female effective population sizes were extremely high: the estimates revealed values of almost 100 thousand females. Bremer et al. (1995) investigated the phylogeography and population structure of swordfish (Xiphias gladius) in the Atlantic, Pacific and Mediterranean Oceans with an emphasis on the Atlantic. Two mitochondrial clades, representing two distinct glacial populations, were discovered but little population sub-division was revealed. The evolutionary relationships of the haplotypes indicated that both historical gene flow and current mixing between oceans were responsible for the lack of observed population differentiation. I answer three main questions by examining the extent and distribution of mtDNA diversity in eulachon. Because the regions in which species have existed for the longest times tend to have the highest level of genetic variation (Crandall and Templeton 1993), McPhail and Lindsey's postulation of a single southern refuge for eulachon can be tested. I predict that the Columbia system, because it remained ice-free during the Wisconsinan glaciation and is the presumed glacial refuge for eulachon, will contain the highest mtDNA diversity. Also, examination of the relationships among the haplotypes should reflect a pattern characteristic of range expansion from one geographic region, and not 2 1 show any deep divergence among groups of haplotypes, which would be characteristic of more than one glacial refuge. Although eulachon are an anadromous species, they exhibit some unusual life-history characteristics such as inconsistency in choice of spawning rivers. If previously measured meristic and morphological traits are indicative of the true population sub-divisions in eulachon, then mtDNA will show haplotype frequency differences among rivers. If life-history information has more bearing on their structure, and eulachon populations are divided in a manner more similar to marine species than anadromous species, this will be reflected in few or no differences in haplotype distributions and a low FST value. If little division is found among eulachon populations and it is determined that anadromy does not establish population structure, then mtDNA data can be used to determine which processes are important for population sub-division in eulachon. Isolation-by-distance predicts a dependency of genetic distance on geographic distance, and the Member-Vagrant Hypothesis predicts that oceanic barriers will have more consequence on population differentiation, and geographic clusters of samples sites will form populations. •22 Figure 2.1 Sample sites. Sample designations are as follows: BR, Bering Sea; C K , Cook Inlet; NS, Nass River; G C , Gardner Canal; K K , Klinaklini River; F N , Franklin River; FR, Fraser River; QT, Queets River; C O L , Columbia River 1997 and 1998, and Cowli tz River 1997 and 1998. 23 2.2 Materials and Methods 2.21 Sample Collection Samples were collected from a number of locations throughout the distributional range of the species (Figure 2.1). These locations include eleven freshwater sites and one marine site. Replicate year samples were collected at two sites (Table 2.1). The marine sample was taken in the Bering Sea, and may be a mixture of eulachon originating from different freshwater environments. Sample sites are referred to as "populations" throughout this paper, however, this term refers only to geographic location and does not imply demographic or genetic distinction. The samples collected in 1995 were originally frozen at -20° C, and fins were later removed from these frozen samples and placed in 95% ethanol for storage. It was subsequently found that this is not a reliable way to store tissue destined for use in the Polymerase Chain Reactions (PCR), as the DNA becomes highly degraded. The samples from the remaining locations were either frozen whole, or fresh livers were removed and stored in 95% ethanol. 2.22 D N A Extraction Genomic DNA was obtained from tissues by Pronase digestion/phenol-chloroform extraction (Taggart et al. 1992). The DNA was resuspended in 20-100 uL of TE solution (10 mM Tris, ImM EDTA in H20; pH 8.0) and stored at -20° C. 24 Table 2.1 Collection locations and descriptions, and mtDNA haplotype diversity and nucleotide diversity within sample sites. Haplotype diversity is estimated from the number of haplotypes and their frequencies within a putative population. Nucleotide diversity is an estimate which includes both the frequencies of haplotypes and the divergences among them. C O L L E C T I O N Y E A R S A M P L E S A M P L E H A P L O T Y P E N U C L E O T I D E L O C A T I O N DESIGNATION SIZE DIVERSITY (SE) DIVERSITY(xlOO) Columbia River 1997 C O L 30 0.6915 (0.0485) 0.112 Columbia River 1998 C L 29 0.7351 (0.0375) 0.109 Cowlitz River 1997 cow 22 0.7357 (0.0670) 0.196 Cowlitz River 1998 cw 30 0.7887 (0.0404) 0.143 Queets River 1998 QT 28 0.6286 (0.0598) 0.093 Fraser River 1995 FR 28 0.5662 (0.0722) 0.083 Knight Inlet: K N Franklin River 1995 F N 17 0.6346 (0.0851) 0.079 Klinaklini River 1995 K K 30 0.7209 (0.0556) 0.153 Gardner Canal: GC Kowesas River 1995 K O 16 0.5565 (0.0793) 0.076 Kitimat River 1995 KI II 0.6061 (0.0542) 0.067 Kemano River 1995 K E 11 0.4675 (0.1223) 0.063 Nass River 1997 NS 19 0.5121 (0.0887) 0.073 Cook Inlet 1997 C K 14 0.6455 (0.0526) 0.089 Bering Sea 1997 BR 29 0.7646 (0.0547) 0.154 25 2.23 P C R Amplification Four genes in the mitochondrial genome were amplified by PCR. Mitochondrial primers and genomic DNA were used to amplify two fragments, each containing two genes. The first fragment contained NADH-5 and NADH-6 genes (approximate molecular weight 2.2 kilobase pairs), and the other contained the 12S and 16S rRNA genes (approximate molecular weight 2.0 kilobase pairs). The ND5/6 region was amplified using the primers C-GLU (5'CAACGGTGGTTCTTCAAGTC3') and C-LEU-3 (5'GGAACCAAAAACTCTTG GTGCAACTCC3') (Park et al. 1993). Amplifications were carried out in 25 uL volumes containing (final concentrations): 200 uM each of dATP, dGTP, dCTP and dTTP, 600 nM of each primer, 2.0 U of Taq polymerase, IX reaction buffer(BRL), 2.0 mM MgCl2 and between 200 and 400 ng genomic DNA template. The first cycle was performed with denaturation for 120 seconds at 95° C, primer annealing for 90 seconds at 55° C, and primer extension for 150 seconds at 72° C. Five further cycles were performed with denaturation for 60 seconds at 94° C, annealing for 60 seconds at 55° C, and extension for 150 seconds at 72° C. The final 35 cycles were performed with denaturation for 60 seconds at 92° C, annealing for 60 seconds at 55° C, and extension for 150 seconds at 72°C. The 12S/16S fragment was amplified with primers 12SF (5AAACTGGGATTA GATACCCCACTAT3') and 12SR (5AGATAGAAACCGACCTGGAT3') (Palumbi 1996). Amplifications were carried out in 30 uL volumes containing (final concentrations): 200 uM each of dATP, dGTP, dCTP and dTTP, 667 nM of each primer, 2.0 U of Taq polymerase, IX reaction buffer (BRL), 1.5 mM MgCi2, and 26 between 200 and 400 ng genomic DNA template. The first cycle was performed with denaturation for 120 seconds at 95° C, primer annealing for 90 seconds at 50° C, and primer extension for 120 seconds at 72° C. The next five cycles were performed with denaturation for 60 seconds at 94° C, annealing for 60 seconds at 50° C, and extension for 120 seconds at 72° C. The final 30 cycles were performed with denaturation for 60 seconds at 94° C, annealing for 60 seconds at 48° C, and extension for 120 seconds at 72° C. 2.24 Restriction Enzyme Digestion Three uL of each PCR product from ND5/6 and 12/16S amplifications were digested separately with 15 restriction endo nucleases: nine enzymes recognizing 4-base nucleotide sequences (Aci I, Alu I, Dpn II, Hae 111, Hha I, Nla III, Rsa I, Taq I), four multi-pentameric enzymes (Ava II, Dde I, Hinf I, Nci I) and two multi-hexameric enzymes (Ban I, Sty I). The digestions took place overnight under the conditions specified by the vendors (New England Biolabs). Separation of restriction fragments was accomplished by horizontal 2% agarose gel electrophoresis. Gels were stained with ethidium bromide and photographed under ultraviolet light. Molecular weights of the fragments were estimated by comparison to a 1 KB ladder (GibcoBRL). 27 2.25 Scoring of Variation All restriction fragment length polymorphisms (RFLPs) could be explained by between one and five restriction site changes. For each RFLP observed for each enzyme, a presence/absence restriction site matrix was created. Each fragment pattern (RFLP) was assigned a single letter code (e.g. Aci I - A, B, etc.) in order of appearance. Composite haplotypes were designated by a 15 letter code; the letters of the code corresponding to the total pattern of site presence/absence data for all 15 enzymes combined (e.g. AAAAAAAAAAAAAAA, BAAAAAAAAAAAAAAA, etc. See Table 2.2). A few fragments smaller than 100 base pairs could not easily be resolved, and as a consequence, some fragments were inferred in order to complete the restriction site matrix. 2.26 Data Analysis A presence-absence matrix of restriction sites for each composite haplotype was constructed using the REAP (McElroy et al. 1992) software package. This matrix was then used for a number of analyses in the REAP program: the number of base substitutions per nucleotide site, d, and standard errors among composite haplotypes (Upholdt 1977, Nei and Li 1979), haplotype diversity for each sample site, h (Nei 1987), nucleotide diversity and nucleotide divergence among populations (Nei and Tajima 1981, 1983). Geographic heterogeneity in the frequency distributions of composite haplotypes was assessed by the chi-square randomization procedure of Roff and Bentzen (1989) as implemented in REAP. The sequential Bonferroni correction (Rice 1989) was used to 28 achieve a "tablewide" alpha-level of 5% when multiple simultaneous tests were performed. An Analysis of Molecular Variance, AMOVA (Excoffier et al. 1993), was used to partition the observed genetic variation into variance components within and among geographically-defined populations (sampling sites) and regions. These calculations were accomplished using the program Arlequin (version 1, Schneider et al. 1996). AMOVA calculations were based on presence/absence of restriction site information rather than genetic distances among haplotypes because drift probably is a much greater factor than mutation in populations derived postglacially (in approximately the last 10,000 years). The model used in the AMOVA therefore assumes that interhaplotypic distances are equal. Arlequin was also used to investigate the historical demography of eulachon. By examining the distribution of pairwise differences between haplotypes, I tested whether eulachon population sizes have remained constant over time, or if they have undergone rapid expansion and exponential growth (Slatkin and Hudson 1991). The historical female effective population size was calculated as in Avise et al. (1988). A test of sample size sufficiency (Bernatchez et al. 1989, Bernatchez and Dodson 1990, Brown et al. 1992, Epifanio et al. 1995) was performed to ensure that a high proportion of the genetic variation present in the sample was detected. I tested isolation-by-distance as an indicator of emerging population structure in two ways: I plotted genetic distance versus geographic distance among sample sites for a visual examination of the data, and I used a Mantel test for correlation of matrices to test the significance of this relationship (Mantel 1967, Sokal 1979). 29 Table 2.2 Composite haplotypes. The first column lists the composite haplotypes referred to in the text. Each letter in the following columns represents the banding pattern seen on the gel for the enzyme specified at the top of the column. C O M P . H A P L Q Acil Alul Avail Banl Ddel Dpnll Haelll Hhal Hinfl Hpall Neil Nlalll Rsal Styl Taql A A A A A A A A A A A A A A A A B A A A A B A A A A A A A A A A D A A A A B A A A A B A A A A A E B A A A A A A A A A A A A A A F B A A A B A A A A A A A A A A G A B A A A A A A A A A A A A A H A A A A A C A A A A A A A A A I A A A A A A A C A A A A A A A J A A A A A D A A A A A A A A A K A A A A A A A A C A A A A A A L A A A A A A A A A A A A A A B N A B A A B A A A B A A A A A A o A A A A C A A A A A A A A A G P D A A A A A A A A A A A A A A Q E A A A A A A A A A A A A A A R B A A A B A A A A A A A A A C S A A A A B E A A A A A A A A A T C A A A B B A A A A A A A A A U A A A A A F A A A A A A A A A V B A A A A A A A A B A A A A A w A A A A A A A A A C A A A A A X A C A A A A A A A A A A A A A Y A A A A B A A A B A A A A A A Z F. A A A B A A A A A A A A A A A A G A A A A A A A A A A A A A A A B A A A A A A A A A A A A A A D A C A A A A A A A A A A A A A A C A D A A A A A A A A A A A A A A E A E A A A A B A A A A A A A A A G A F B A A A B A A A C A A A A A A A G B A A A B B A A A A A A A A A A H A A A A A B A A A A A A A A A AI A A A A A A A A A A A A A A F A J A D A A A A A A A A A A A A A A K A A A A A A A A A A A A A A H A L A A A A B B A A A A A A A A A A M A A A A A F A A A C A A A A E 30 2.3 Results 2.31 Mitochondrial D N A Restriction Fragment Variation Of the fifteen enzymes used, eight provided restriction fragment length polymorphisms (RFLPs): Aci I and Hpa II with the 12/16S fragment, and Alu I, Dde I, Dpn II, Hha I, Hini I and Taq I with the ND5/6 fragment. Each of these RFLPs was generated by either a gain or a loss of an endonuclease recognition site, indicating that nucleotide substitutions were responsible for all of the site changes. No significant mtDNA fragment size variation was observed, therefore, there were probably no additions or deletions in the genes surveyed. The mean fragment size for the ND5/6 region was 2528 (± 19) bp, and the 12S/16S region was 1999 (± 4) bp. From the four mitochondrial genes sampled as two fragments, a total of 142 restriction sites was found: 86 of these were in the ND5/6 fragment and 56 in the 12S/16S fragment. The average number of bases surveyed was 546, representing approximately 13% of the combined size of the two fragments. 2.32 Mitochondrial D N A Haplotype Variation and Distribution In my sample of 315 eulachon, 37 mtDNA composite haplotypes were resolved (Table 2.2). Haplotype A, the most abundant haplotype, was found in each population and was present in 50% of the total sample. Haplotypes A and B (B was also present at each sampling location) together accounted for 67% of all fish sampled. Haplotype A was the prevalent haplotype at every location sampled except the Bering Sea, which was approximately 38% haplotype F. Haplotype F was the third most abundant haplotype, 31 found at all other sites, at varying frequencies (3-28%). Haplotypes A, B and F together accounted for 79% of the total sample. The remaining 34 haplotypes were found at low frequencies in one or some of the populations. No other haplotype was found in all populations, and 19 haplotypes were "private" (found only in one location, Table 2.3). The relationships among the haplotypes in terms of site gains and losses (Figure 2.2) indicated that all haplotypes are closely related with the number of single nucleotide substitutions separating the haplotypes between one and thirteen. There were 48 recognition site losses, and 19 site gains from the putative ancestral haplotype, A. Haplotype divergences ranged from 0.0906% to 1.2354%, with a mean of 0.3876% (standard error 0.00828) and median of 0.3686%. The parsimony mutational network (Figure 2.2) suggests that the haplotypes fall into two major clusters: one such cluster is formed around haplotype B, and consists of B, D, F, H, N, O, R, S, T, Y, Z, AE, AF, AG and AL (Group I), and the other, formed around haplotype A, comprises all other haplotypes (Group II). All populations except the Bering Sea sample and the Cook Inlet sample contain a majority of Group II haplotypes (67%-91%), while Cook Inlet contains 50% and the Bering contains 41% Group II haplotypes, respectively (Figure 2.4). The presence of two mtDNA groupings, however, is not strongly supported; only one site change separates these clusters (a loss from A to B) and the two groups were not resolved following bootstrap and consensus tree analyses of the restriction site matrix (dendogram not shown). Figure 2 .2 Mutational network showing evolutionary relationships among haplotypes. Each ellipse represents a haplotype, the lines connecting the haplotypes represent one mutation separating the haplotypes and the cross hatches represent additional mutations between the haplotypes. The size of each haplotype's ellipse is proportional to its relative abundance in the total sample. 33 Table 2.3 Distribution of haplotypes at each sampling location. Sample designations are as in Table 2.1. H A P L O C O L C L C Q W C W QT FR F N K K K O KI K E NS C K BR T O T A L A 15 9 11 12 16 18 10 15 10 5 8 13 7 8 157 B 6 11 2 2 4 4 2 5 4 5 1 3 3 4 56 D 1 1 1 3 E 4 1 1 1 1 8 F 3 3 6 5 1 1 2 1 1 4 1 1 38 G 1 1 2 1 1 1 7 H 1 1 I 1 1 J 2 1 3 K 1 3 1 1 1 7 L 1 1 N 1 1 O 1 1 P 1 1 Q 1 1 R 1 1 S 1 1 T 1 1 U 1 1 V 1 1 W 1 1 2 X 2 2 Y 1 1 Z 1 i . 34 A A 1 A B 1 A C 1 A D 2 1 A E 1 A F 1 A G 1 A H 1 AI 1 A J 1 A K 1 A L 1 A M 1 T O T A L 30 29 22 28 28 28 17 30 16 11 11 19 14 29 2.33 Rare and Private Haplotypes Eleven of the fourteen samples contained at least one private haplotype, and all sample sites except Cook Inlet contained one or more haplotypes found in low frequency (Table 2.3). Due to the relatively small sample sizes for each river, it is possible that the rare and private haplotypes would not be confined to a single sampling area if a larger sample from each location were analyzed. A test for sample size sufficiency, however, indicated that with a sample size of close to twenty for each sampling location (my average sample size was 22), on average, approximately 90% of haplotypes present at each location were found. 35 2.34 Mitochondrial D N A Variation Within and Among Sample Sites Haplotype diversity, a measure of the number of different haplotypes within a sample site, was varied and moderately high, ranging from 0.49 (Kemano River 1995) to 0.80 (Cowlitz River 1998), with an average of 0.66. Nucleotide diversity (xlOO), which measures the number of haplotypes within a sample site as well as the amount of divergence among them, was fairly low, ranging from 0.06 (Kemano River 1995) to 0.20 (Cowlitz River 1997), and averaging 0.11 (Table 2.1). Nucleotide divergence among sample sites ranged from less than zero (i.e. fish in one river with haplotypes more similar to fish in rivers other than in their own) to 0.000307 (between the Bering Sea and Columbia River 1997 samples) and averaged 0.000041. 2.35 Temporal Population Structure Temporal differences at the two sites that were sampled for two years was tested with a chi-squared randomization test (Columbia River, P=0.067; Cowlitz River, P=0.065) with the Bonferroni correction for multiple comparisons. The P-values shown correspond to the probability of observing as extreme a result if the test were repeated many times (in this case, 1000 permutations) and the null hypothesis of no genetic differentiation among populations were true. The mitochondrial haplotype distributions for the two years in each river were compared, and the null hypothesis of genetic homogeneity was not rejected at either of the two sample sites. This result allowed me to combine the Columbia River 1997 data with the Columbia 1998 data, and the Cowlitz 1997 data with the Cowlitz 1998 data in the geographic population structure analysis. 36 2.36 Spatial Population Structure Similar chi-square randomization tests for geographic structure were performed, also with a Bonferroni correction (Table 2.4). These tests were performed on subsets of the data to examine populations geographically clustered within inlets or other small areas. For example, the populations in three different rivers within Gardner Canal were tested for heterogeneity and were found to be statistically homogeneous (P=0.367). It has been suggested, on the basis of morphological characters, that the Franklin River and the Klinaklini River contain distinct runs of eulachon. Genetic information, however, shows almost complete homogeneity (P=0.946). The Alaskan populations, those in Cook Inlet and the Bering Sea, were also not significantly different (P=0.934). A surprising result, however, was found in a comparison between the Cowlitz River and the Columbia River (the Cowlitz is tributary to the lower Columbia): while these rivers do not show a significant genetic difference with the corrected alpha (P=0.02), the P-value is low relative to comparisons among populations a similar or even greater geographic distance apart. Performing an AMOVA with hierarchical structure revealed that approximately 97% of the variation observed was found within populations (P=0.018) and only about 2% is found among regions (regions consist of a number of sample sites — those that could be grouped according to the Monte Carlo simulation and geography, Table 2.5). While each population displays high genetic variation, there is very little genetic differentiation among the putative populations ("regions" in the AMOVA): the F S T value is 0.023. 37 Table 2.4 Probabilities of monte carlo tests for population differentiation. The first column lists the "population" pairs tested, which were chosen based on geography. The second column is the P-value given by the test, and the third column is the Bonferroni corrected alpha, which is adjusted for multiple comparisons. POPULATION P - V A L U E ADJUSTED COMPARISON (TEST) A L P H A Franklin River, Klinaklini River 0 946 0.05 Cook Inlet, Bering Sea 0 934 0.025 Kemano River, Kowesas River, Kitimat River 0 367 0.017 Columbia River 1997, Columbia River 1998 0 067 0.013 Cowlitz River 1997, Cowlitz River 1998 0 065 0.010 Columbia River, Cowlitz River 0 020 0.008 Table 2.5 Results of AMOVA test for partitioning of eulachon mtDNA variance. Collections were grouped geographically into 7 regions based on the chi-square tests for temporal and geographic heterogeneity. The regions are: Columbia and Cowlitz Rivers (both years), Queets River, Fraser River, Knight Inlet, Gardner Canal, Nass River and Cook Inlet/Bering Sea. SOURCE OF VARIATION PERCENTAGE OF VARIATION P-VALUE Among Regions 1.75 0.192 Among populations within regions 0.64 0.268 Within populations 97.61 0.018 2.37 Isolation-by-distance The graph of genetic distance (d) versus geographic distance (Figure 2.3) shows a trend similar to that expected in a typical isolation-by-distance model (r2=0'.22, P=0.0001). A Mantel test for correlation of matrices also supports isolation-by-distance among sample sites (P=0.012). 0.00035 -, 0.0003 0.00025 g 0.0002 c £ 0.00015 c U 0.000-1 H 0.00005 ^=0.2239 500 1000 1500 Geographic Distance (km) 2500 Figure 2.3 Isolation-by-distance of eulachon sample sites. 39 2.38 Historical Demography The distribution of pairwise differences between haplotypes (Slatkin and Hudson 1991) was examined for evidence of either an evolutionarily constant population size or an exponentially growing population size. This test can only be performed when the variance of the distribution of pairwise comparisons is larger than the mean. This requirement was met in three sample sites (Cowlitz River, Klinaklini River and Nass River) and all tests rejected the null hypothesis of exponential growth. When all samples were pooled together, the variance was larger than the mean, and the null hypothesis of exponential growth was again rejected. The historical female effective population size was estimated to be approximately 98 000, with a mutation rate of 1% per million years and an eulachon generation time of 4 years (after Gold et al. 1993). 40 2.4 Discussion 2.41 Glacial Refugia and Post-glacial Dispersal of Eulachon At various times throughout the last million years, the Pleistocene Epoch, much of North America was covered with ice (McPhail and Lindsey 1970). The Pleistocene was characterized by cycles of cooling and warming, during which glaciers advanced and retreated over the landscape. Of the major cold periods during this era, the most recent was the Wisconsinan glacial period. The Wisconsinan, starting approximately 50,000 years ago and ending approximately 10,000 years ago, had a striking effect on the distribution of fish species throughout northwestern North America (McPhail and Lindsey 1970, 1986). Freshwater fish with pre-glacial distributions were strongly affected by the advance of the ice sheets: they either went extinct, or were forced into ice-free glacial refugia (McPhail and Lindsey 1970, 1986). Anadromous fish like eulachon, although able to survive in the ocean, require a freshwater habitat for spawning and therefore utilized these refugia as well. When the ice sheets began to receed, post-glacial dispersal along a number of routes was possible, and fish dispersed throughout North America and gradually recolonized their present distributions. Three major refugia contributed to the group of fish that currently exist on the Pacific coast of North America: the Pacific refuge, consisting mainly of the southern two-thirds of the Columbia River system; the Bering refuge; and an area south of the ice, between the Appalachians and the Rockies (McPhail and Lindsey 1970, 1986). The latter area did not contribute many species to the coast, and did not contain eulachon because it was not connected to their marine environment. Many of the freshwater fish that have 41 recolonized the coast were isolated in the Bering and Columbia refugia. Two routes of post-glacial dispersal from the Columbia refuge were possible: an inland route, which was probably taken by many freshwater species intolerant to saltwater, and a coastal route, which could be utilized by euryhaline species such as eulachon. The Bering refuge also contained a number of fish that are currently found on the coast (McPhail and Lindsey 1970). McPhail and Lindsey (1970) hypothesized that eulachon were isolated only in the Pacific refuge. Coastal dispersal would be likely for eulachon as they are an anadromous species; not only are they capable of dispersing through saltwater, they spend a large portion of their lives in a marine environment. It is possible, but unlikely, that eulachon survived glaciation in the Bering refuge; while they are found in Bristol Bay and near the Pribiloff Islands, they have not crossed to the Siberian side of the Bering Strait which was also part of Beringia (McPhail and Lindsey 1986). Many species of fish that survived glaciation in the Bering refuge are found in that region (e.g. Taylor et al. 1996, Redenbach and Taylor 1999). The mtDNA data collected in my study are consistent with the single refuge hypothesis. All of the haplotypes are closely related and little divergence is present, suggesting no extensive isolation of gene pools. There is no deep split in the tree of haplotypes (Figure 2.2) as would be expected if two gene pools had been kept separate for at least 40,000 years, or even much longer (ie. pre-Wisconsinan glaciations). In fact, this "star phytogeny" with few prevalent haplotypes and many closely related rare haplotypes is consistent with haplotype phylogenies observed in a number of marine fishes including haddock, Atlantic cod and cape hake (see Shields and Gust 1995). 42 Star phylogenies are typical of populations that have undergone range expansion or exponential population growth (Slatkin and Hudson 1991, Rogers 1995). Given the geological history of the current range of eulachon, it is obvious that they have undergone range expansion in their recent evolutionary history. Populations would be expected to reflect this in distributions of pairwise comparisons of haplotypes. For populations that have remained at a constant size, this distribution is geometric. For populations that have undergone exponential growth, this distribution is Poisson. All tests of exponential growth in eulachon rejected the null hypothesis of expanding populations. This may seem at odds with the glaciation affected history of eulachon, however, the model of exponential growth was developed for human population expansion over the last 100 thousand years. The Pleistocene glaciations have caused multiple range contractions and expansions for eulachon, and this may be why they do not exhibit the expected Poisson distribution correponding to exponential growth. A single refuge is also suggested by the pattern of haplotype and nucleotide diversity within sample sites. Haplotype diversity takes into account the frequencies of the haplotypes, but not the amount of divergence among the haplotypes. Nucleotide diversity calculations consider both of these features. Eulachon populations tend to exhibit high haplotype diversity and low nucleotide diversity (Table 2.1); this phenomenon is often observed in species with many haplotypes that are closely related. Many, closely related haplotypes is a pattern often associated with a single ancestral population that has recently undergone rapid expansion (Rogers and Harpending 1992) like those of sardine, Sardinops spp. (Bowen and Grant 1997). If eulachon had persisted in more than one refuge, the haplotypes would be much more distantly related, increasing 43 the nucleotide diversity. For example, the two glacial races of rainbow smelt, Osmerus mordax, described by Bernatchez (1997) have high nucleotide diversity (0.821 and 0.864) and are separated by 0.7% nucleotide sequence divergence in mtDNA. By contrast, my haplotype "groups" are separated by only one mutation and differed by less than 0.4% in sequence. Nucleotide and haplotype diversity are generally higher in older populations (Crandall and Templeton 1993). A prediction can therefore be made concerning which rivers will contain eulachon populations with the highest diversities. If eulachon were isolated in a single southern refuge as postulated, then the Columbia and Cowlitz rivers would be expected to exhibit a higher level of diversity than the Bering population. This was not observed (Table 2.1). The haplotype and nucleotide diversities in the Bering are as high as those in the Columbia/Cowlitz. If the rivers in the postulated southern refuge are grouped, however, the haplotype and nucleotide diversities increase and exceed those in the Bering. These southern populations have probably existed for a much longer time, and are likely descendents of the source population for more northerly populations. Haplotype A was abundant at every sampling site. Haplotypes B and F were also present at each location. According to coalescent theory, older alleles are found in higher frequency, occur in the interior of a phylogeny, and possess greater numbers of mutational connections (Crandall and Templeton 1993). If populations are sub-divided, then the older alleles are the most widespread. Haplotypes A and B have all of these features, and are likely the "oldest" haplotypes, with many of the more rare haplotypes resulting from new mutations. 44 2.42 Genetic Population Structure of Eulachon Pacific salmon typify the life cycle and population structure of anadromous fish species: each river supports a distinct spawning population (for examples of population structure in chum salmon, Oncorhynchus keta, see the following. Allozyme data: Phelps etal. 1994. MtDNA data: Park et al. 1993. Minisatellite data: Taylor et al. 1994, Beacham 1996). Perhaps because salmon have been so intensively studied, this type of sub-divided population structure is generally expected of anadromous fish. This type of structure, however, may not apply to all anadromous fish species and eulachon may be one such exception. Reproductive isolation is necessary for the creation and maintenance of genetically distinct populations in the absence of physical barriers. One of the most important ways that separate populations can be maintained is through imprinting at early life-history stages and homing for reproduction (Horrall 1981). Repeated cycles of imprinting and homing allow the accumulation of differences among populations. These differences may be morphological, meristic, behavioural or genetic (Taylor 1991). Although meristic and morphological traits appear to differ in eulachon from different rivers (Hart and McHugh 1944, Berry 1996), my analysis of mitochondrial DNA suggests that there is little genetic differentiation among eulachon from distinct freshwater locations throughout their range. The evidence for distinct populations — grease type, gill raker and vertebral counts, spawning and larval release times — are largely influenced by specific environmental conditions at early life-history stages. For example, water temperature at the larval stage has an effect on the number of gill rakers the fish will eventually have (Lindsey 1981). Each river in which eulachon spawn has a 45 distinct temperature regime, and differences among rivers in quantitative traits (ie. body form, grease quality) may result from environmental variation in conditions such as water temperature. The chi-square tests for geographic heterogeneity within inlets showed no significant genetic heterogeneity among populations. Heterogeneity was tested in these situations so that samples from rivers within the same inlet could be grouped, and power to detect geographic structure would be enhanced. Rivers within Gardner Canal (Kemano, Kitimat and Kowesas rivers) did not show significant heterogeneity. This was also true for the samples within Knight Inlet: Franklin River and Klinaklini River. These results are not surprising because these rivers have the same outlet to the ocean, and eulachon, which do not spend much time in freshwater, may not home to a particular river, but to an inlet instead. Consequently, there is probably a great deal of mixing between fish from different rivers within an inlet. The low P-value in the comparison between the Columbia and the Cowlitz relative to other comparisons may result from the older (ie. pre-glacial) age of these populations. Perhaps they contain more highly differentiated populations because eulachon have been present in the southern part of their geographic range for a much longer time and have, therefore, had a much longer time in which to diverge from one another. Ward et al. (1994) discussed average and median allozyme-based FST values for 113 different species of fish from three different life-history types: freshwater, anadromous and marine. For freshwater species, the average FST was 0.222 and the median value was 0.144; for anadromous species, the average was 0.108 and the median was 0.081; and for marine species, the average FST value was 0.062, and the median was 46 0.020. Because eulachon are anadromous, the expectation is that their F S T value will be intermediate and match that of other anadromous species. This means that the "among regions" portion or the genetic variance is predicted to be fairly high, and the "within regions" portion will be low. The AMOVA showed that only a small amount of genetic variation exists among population groups or regions (sample sites grouped by the chi-square test for geographic heterogeneity), and most (>97%) of the variation exists within these regions. Very little differentiation in the frequencies of the common haplotypes is present among populations. Because mtDNA has proven an effective marker for detecting distinct populations created since the Pleistocene glaciations (Avise et al. 1987, Taylor and Bentzen 1993, Bernatchez and Martin 1996), my inability to uncover strong population structure was not due to limited sampling or limited genetic variation. I found a great deal of genetic variation, but this variation was not strongly geographically partitioned. In addition to the genetic data, which indicate that little mtDNA population structure apparently exists in this species, several aspects of the eulachon's behaviour are consistent with this idea. Eulachon can be observed spawning in a specific river for a number of years, disappear from that area for the next few years, and then reappear and spawn in that area again (Lewis 1996). As well, disturbances at the river mouth can cause the fish to refrain from entering the river to spawn and it is possible that they detour to neighbouring rivers. If it is true that eulachon do not return to their natal rivers to spawn, then defining populations by spawning river is an incorrect approach. There are a number of reasons that eulachon may not exhibit population structure as revealed by mtDNA: the populations may not yet have reached equilibrium, the 47 proportions of fish that are homing and straying may not be what is typically expected for anadromous fish, or eulachon may be structured in a manner more similar to marine fish species. Eulachon population sub-division may not be apparent because the time since glaciation has not been enough to allow the accumulation of differences among them; however, many other anadromous fish populations of the same age exhibit strong genetic structure, including other osmerids (Taylor and Bentzen 1993, Bernatchez and Martin 1996). Another possibility is that the historical high effective population sizes have not yet allowed differentiation as measured by neutral genetic traits. Alternatively, it may be that the high number of migrants and resulting large amount of genetic mixing constrains the formation of distinct populations. If eulachon were isolated in a single refugium during the last Pleistocene glaciation, then they have had 10-15,000 post-glacial years to diverge into discrete populations. This evolutionarily short period, although allowing enough time for other species to diverge into populations, may not have been long enough for separate eulachon populations to form, at least when assessed with mtDNA, particularly with straying and high female effective population sizes such as 98 000. Long-term female effective population sizes have been determined for a number of marine fish species, and range from approximately 5 000 (black sea bass, Avise 1992) to 800 000 (Atlantic menhaden, Avise 1992). Estimates closest to that for eulachon are Gulf and Atlantic red drum (95 000 and 93 000, respectively) and Gulf menhaden (250 000). Red drum have a higher mean haplotype divergence but a shorter generation time (Gold et al. 1993), and menhaden have a haplotype divergence value that is almost three times larger than that in eulachon and a similar generation time (Gold et al. 1993). It has been estimated that, for marine fish, census female population sizes are up to two orders 48 of magnitude larger than effective female population sizes (Gold et al. 1993). If this is true for eulachon, and there is a 1:1 ratio of males to females, then the expected census size would be approximately 20 million. This number is much smaller than what is observed, most likely because eulachon populations have been subject to repeated bottlenecks during the Pleistocene glaciation. These bottlenecks severely reduced haplotype divergences, thereby decreasing the effective population size. I have found that eulachon populations appear to be structured more like those of marine fish than anadromous fish, with a so-called star phylogeny and <2% of the total genetic variation found among the populations. As well, the mtDNA FST value was 0.023, which was similar to Ward et al.'s (1994) median FST value, 0.020, for marine fish species. This suggests that processes typically associated with population sub-division in the marine environment may be more influential on the genetic structure of eulachon than those associated with anadromous species. 2.43 Life-history and mechanisms of population sub-division in eulachon How is reproductive isolation created and maintained in marine environments? Of the five mechanisms described by Palumbi (1994), two are probably of little importance for eulachon. The recent geologic history of the current geographic range of eulachon is well-characterized, and while the Pleistocene glaciations certainly had an effect on eulachon populations, isolation in distinct refugia probably did not contribute to the current genetic population structure of eulachon. Because mtDNA was chosen for this analysis, and it is assumed to be neutral, the effects of selection of eulachon genetic population structure will not be assessed. The observed lack of molecular differences, 49 however, does not necessarily mean that there are no genetic differences among populations in quantitative traits (e.g. morphology, behaviour, life-history) as a response to selection (Clayton 1981). Of the remaining three mechanisms, two may be discussed in combination: behavioural limits to dispersal and invisible barriers. The Member-Vagrant Hypothesis (lies and Sinclair 1982, Sinclair and lies 1989) as applied to eulachon predicts the existence of a small number of geographically-separated, genetically distinct populations of eulachon, each occurring in an area bounded by some geographic or oceanographic feature (a retention zone). The invisible barriers are formed by these retention zones or ocean currents, and the behavioural component involves homing to the retention zones for reproduction. The final possibility is isolation-by-distance, which predicts a large-scale pattern of genetic similarity among eulachon originating in rivers that are geographically close to one another, and a gradual decrease in similarity as geographic distances increase. My data support a weak pattern of isolation-by-distance (r2=0.22, P=0.0001) when genetic distance is plotted against geographic distance (Figure 2.3), and when a Mantel test is performed (P=0.012). The low r2 value indicates that the trend observed in the graph only partially accounts for the pattern observed in nature. I contend that my data support "nascent" population structure in this species, consistent with its recent post-glacial recolonization and large evolutionarily effective population sizes. 50 Figure 2.4 Proportion of groups I (B-Hke haplotypes), represented by the lighter portion of the pies and II (A-like haplotypes), represented by the darker portion of the pies, at each sampling site. 51 Further support for emerging population sub-division comes from the trend observed in Figure 2.4. If the haplotypes at each sampling site are grouped into "A-like" and "B-like" haplotypes and their relative proportions are plotted on a map, a north-south trend in proportions of each group per site becomes apparent. Although these two "groups" of haplotypes are only separated by a single restriction site change, this shallow division is consistent with the recency of recolonization and large effective population sizes. The larval retention zones predicted by the Member-Vagrant Hypothesis are difficult to define. Alternative clusterings of sample sites (i.e. over a larger geographic range than inlets) do not increase the proportion of total genetic variation that is found among populations. If all of the samples collected south of the Aleutian Peninsula are grouped against the Bering samples, however, the percentage of variation among regions becomes 11.5%. It is possible that two larval retention zones exist within the range of eulachon: the Bering Sea and the North Pacific Ocean. A recent study of the beluga whale, Delphinapterus leucas, suggests that the peninsula may be a barrier to genetic exchange between populations of this species (O'Corry-Crowe et al. 1997). As well, the peninsula is thought to be a partial barrier to migration of walleye pollock larvae (Bailey et al. 1997) and chum salmon (Winans et al. 1998). The differences in both the percentage of variation among populations, and the high proportion of "B-like" haplotypes in the Bering is driven mainly by one haplotype, F, which is found in all populations, but at the highest percentage in the Bering. In comparison to the average value for all other sample sites combined (6%), the frequency of the F haplotype in the Bering is unusual (38%). This is probably a result of the effects 52 of random genetic drift, and partial population isolation due to the Aleutian Peninsula. It is likely that post-glacially, eulachon dispersed northward along the coast from the Pacific refuge, and that this dispersal resulted in the founding of some number of populations in different coastal areas. At some point after glaciation, eualchon dispersed past the Aleutian Peninsula and into the Bering Sea where they are found today. Since that time, my data suggest that migration between the population in the Bering and populations south of the Peninsula has been restricted, allowing genetic drift to increase the frequency of the F haplotype in the Bering while the frequency of the A haplotype remains high throughout the rest of the range. While no deep divergences were found among putative eulachon populations, the data are consistent with emerging population sub-division throughout the geographic range of the species. An isolation-by-distance pattern was found, and the Member-Vagrant Hypothesis with two major retention zones was also supported. Although my data suggest very weak population sub-division in eulachon, the behavioural and evolutionary characteristics of this species, in particular, their historically high effective population sizes, may require more sensitive assays of intraspecific population structure. In Chapter 3,1 employ microsatellite DNA assays to provide more powerful tests of population sub-division in eulachon. 53 Chapter 3 Characterization and analysis of five microsatellite loci in the eulachon, Thaleichthys pacificus, resolves unusually low levels of genetic variation 3.1 Introduction Microsatellite loci form a class of highly polymorphic, and therefore highly informative, nuclear DNA. This class of DNA consists of di-, tri- and tetranucleotide repeats, although the repeats may be longer or one base pair only (Beckmann and Weber 1992). Part of the utility of microsatellites stems from their abundance in the genome; they are found approximately every 10 kilobases in the eukaryotic genome (Tautz 1989) and are flanked by unique non-repetitive regions in which primer sequences may be designed. Mutation rates for microsatellites are thought to be between 10~2 and 10"5 (Weber and Wong 1993), although there is some variation in this estimate (Schug et al. 1999). Because microsatellites have a higher mutation rate and are more variable than other loci, they are considered more sensitive than other nuclear markers to population genetic processes such as random genetic drift (Estoup et al. 1998a). Another quality important for studies of population structure is selective neutrality. Due to the random distribution of microsatellites throughout the genome, they are occasionally found within coding regions, however, the majority of microsatellite loci are found in neutral regions (Jarne and Lagoda 1996) and selection can be discounted as an important evolutionary force. The combination of high variability, selective neutrality and the abundance of possible loci makes microsatellites a popular and effective class of markers for studies of intraspecific population structure, as well as hybridization, linkage mapping, paternity 54 testing and pedigree analysis (Amos et al. 1993, Queller et al. 1993, Bowcock et al. 1994, Roy et al. 1994). Microsatellite mutation is presumed to occur through slipped-strand mispairing during replication, which causes variation in the number of repeat units at a particular locus (Levinson and Gutman 1987). Different alleles are identified by their different lengths (measured by the number of base pairs), and are usually resolved by electrophoresis on polyacrylamide gels. Microsatellites are bi-parentally inherited and are exhibited in a co-dominant manner, and so, unlike one haplotype expressed per individual for mtDNA, two alleles, one from each parent, are expressed. Repeat arrays may be pure, compound or interrupted. Pure, or perfect, arrays contain one set of repeat units only (ie. C AC AC AC AC A), compound arrays contain more than one set of repeat units (ie. CACACACACACTCTCTCT) and interrupted arrays contain a set of repeat units that are interrupted in one or more locations by "extra" base pairs (ie. CACACAGGCACACA). When mutation rates are high, the mutation process becomes relevant to data analysis. There is direct evidence for polymerase slippage during replication (Strand et al. 1993), and this is the mechanism regularly used to explain mutations at microsatellite loci. The studies that have shown slippage to cause mutations, however, have also noted that occasional larger mutations may occur by other processes including unequal crossing-over. Given that the mechanism generally occurring to create new microsatellite alleles is slippage, four mutation models have been proposed for the creation of new alleles at a locus: the stepwise mutation model (SMM, Ohta and Kimura 1973), the 55 infinite alleles model (IAM, Kimura and Crow 1964), the two-phase stepwise mutation model (TPM, DiRienzo et al. 1994) and the /^ -allele model (Jarne and Lagoda 1996). The SMM suggests that new microsatellite alleles differ from their previous state by one repeat unit (Valdes et al. 1993). Addition and subtraction of repeat units to create new alleles occur with equal probability, and only one unit is added or subtracted at a time. Because the relationships between alleles can be determined (alleles with one repeat unit between them are more closely related than alleles with a two repeat unit difference), the SMM has a "memory" of past allele size. The IAM proposes that new alleles are formed independent of their previous states, and so alleles differ equally from one another, with alleles that are close in size no more related than alleles with great size disparity. While the IAM seems to apply in the case of slightly larger repeat units such as minisatellites, a modified version of the SMM seems to apply to microsatellites (DiRienzo et al. 1994). The TPM (DiRienzo et al. 1994) is similar to the SMM because new allelic states depend on previous states and most mutations are the result of one-step mutations, however, this two-phase model allows for occasional larger mutational steps as well. The ^ f-allele model limits the number of possible allelic states, and instead of creating new alleles, mutations change one allele into another already existing allele. Which model best fits microsatellite evolution depends on the length of the repeat, the type of repeat and the type of repeat array, however, the SMM, or the modified version of the SMM, the TPM, works well for repeat lengths of 1 to 5 base pairs (Jarne and Lagoda 1996). 56 The analysis of microsatellite data is affected by the choice of mutation model considered most appropriate for the data. Estimates of population sub-division such as F S T values assume that all alleles are equally related (IAM, /^ -alleles model), however, this may not be true for microsatellites in which the high mutation rate makes mutation an important evolutionary factor. An analogue of F S T , RST, has been developed (Slatkin 1995). R S T values also describe the extent of population sub-division, but take the relationships between alleles at a locus into account as do the SMM and the TPM. Microsatellites are known as a class of "high resolution" markers due, in part, to their ability to resolve intraspecific population structure in cases where other methods (allozymes, mtDNA) have been unsuccessful. Recent microsatellite investigations have revealed previously undetected intraspecific population structure in a number of species, ranging from Atlantic salmon, Salmo salar (McConnell et al. 1995), brown trout, Salmo trutta (Estoup et al. 1998a) and Arctic char, Salvelinus alpinus (Brunner et al. 1998) to black bear, Ursus americanus (Paetkau and Strobeck 1994). Low levels of genetic variation are typical of Atlantic salmon when assessed with allozymes, nuclear ribosomal DNA markers and mtDNA (Davidson et al. 1989). These markers have been able to determine the continent of origin of Atlantic salmon caught in the Greenland mixed fishery, but no finer-scale resolution has been achieved. Microsatellites, however, have revealed more extensive variation in both the number of alleles and the observed heterozygosity, and are able to distinguish Bay of Fundy versus eastern shore populations in Nova Scotia (McConnell et al. 1995). Estoup et al. (1998a) conducted a comparative analysis of microsatellite and 57 r allozyme loci in brown trout, which have a complex evolutionary history and exhibit many phenotypic forms and variation in life-history traits. Previous allozyme and mtDNA studies were able to resolve population differentiation in brown trout in western Europe (Bernatchez et al. 1992), but the patterns of genetic variation observed did not appear to be related to geographic distribution. Microsatellites were successfully employed as a higher resolution marker to enhance the clarity of the distribution of genetic variation. The higher level of polymorphism increased the power of tests of population differentiation, allowed the assignment of individual fish to different populations, and revealed a pattern of isolation-by-distance (Estoup et al. 1998a). Arctic char is another species that exhibits a great deal of variation in morphology, colouration, ecology and life-history traits (Brunner et al. 1998). Studies of genetic heterogeneity in allozymes and mtDNA, however, have been unable to detect genetic differences among different morphs, different populations and different subspecies (Kornfield et al. 1981) due to the extremely low levels of genetic variation in these markers. Brunner et al. (1998) examined 6 microsatellite loci in Arctic char from central Alpine lakes in Europe and revealed high levels of heterozygosity, many alleles per locus and enough variation to resolve strong genetic differentiation. Both allozyme loci and mtDNA have been used to examine levels of genetic variation in the three North American bear species, but have been unable to detect population differentiation in widely dispersed populations, and in some cases, any genetic variation at all (Cronin et al. 1991). Black bears exhibit low levels of genetic variation at allozyme loci and in mtDNA throughout their range. Paetkau and Strobeck (1994) examined 4 microsatellite loci in black bears to investigate the extent of genetic variation 58 in the species for conservation purposes. They found that while bears in most geographic regions exhibited high levels of polymorphism at microsatellite loci, one of the regions contained a population of bears with a low amount of variation. Studies of this type can aid in the characterization of populations that are at risk with respect to genetic diversity. It is important to use more than one class of marker in any study of molecular genetic variation. One marker alone may be affected by selection, hybridization or introgression, and the result of the study will be a gene tree rather than a population tree. The five microsatellite loci in this study, as well as adding a second class of marker, make the number of independent genetic markers up to six (including mtDNA). This increases the power and the robustness of the overall analysis. Finally, microsatellites are not restricted to describing maternal lineages or population structure, but can describe total population structure. I analyze five smelt microsatellite loci to describe the extent and distribution of microsatellite DNA variation in eulachon. The mitochondrial DNA results in the previous chapter were consistent with McPhail and Lindsey's (1970) proposal of a single glacial refuge for eulachon during the most recent glaciation, and I wanted to confirm those results with microsatellite data. Again, I expect that the highest diversity will be found in the Columbia system, and no significant north-south split will be detected in allele frequencies. After analysis of mitochondrial DNA data, the population structure of eulachon throughout their geographic range is still unclear. Microsatellites have been successful in a number of cases in which mitochondrial DNA was unable to resolve intraspecific 59 population structure, and I use them in a further attempt to identify distinct populations of eulachon. Given the life-history and mtDNA population structure of eulachon, it is unlikely that microsatellites will resolve the type of population structure that is typically seen in anadromous species. If structure is resolved but there does not appear to be a distinct population within each river, then microsatellites may help to differentiate between the possible processes that are responsible for population structure in the marine environment: isolation-by-distance and the Member-Vagrant Hypothesis. 60 3.2 Materials and Methods 3.21 Isolation and Cloning of 300-700 bp Fragments Two eulachon DNA samples were chosen for cloning of microsatellites: one each from the Columbia River and the Bering Sea. Geographically widespread samples were used to detect a maximum amount of variation. Approximately 4 ug of total genomic DNA was digested with a cocktail of restriction enzymes: Hae III, Rsa I, Alu I and Hinc II. Twenty ug of Bluescript II vector was cut with Sma I and dephosphorylated. The products of both restrictions were run on 1.5% low melting point agarose gels. Eulachon DNA restriction fragments between 300 and 700 base pairs were removed from the gel and purified using beta-agarase I and the protocol recommended by the manufacturer (New England Biolabs). The purified DNA was ligated into the Bluescript vector, and this vector was then transformed into E. coli supercompetent cells following the protocol outlined by the manufacturer (Stratagene) and Glenn (1995). Product of this transformation was then placed on LB agar plates containing the antibiotic ampicillin and X-gal to facilitate blue-white colony selection, and incubated at 37°C overnight. 3.22 Detection and Amplification of Positive Colonies Putative microsatellite-containing positive colonies (only colonies containing the plasmid can grow on media containing antibiotic and only colonies containing plasmids with inserted eulachon DNA will be the correct colour) were picked with sterile pipette tips and replated onto grid petri dishes containing LB agar and incubated overnight at 37 °C. Colonies were then transferred from the grid petri dishes to Hybond nylon 61 hybridization membranes (Amersham) . The membranes were probed with 3 2P end-labeled synthetic oligonucleotide probes ((GA)n, (CA)n, (GATA)n and (AAG)n) in Westneat buffer (7% SDS, 1 mM EDTA, 0.263 M Na2P04, 1% BSA) at 55°C. After stringency washes (2XSSC/0.1% SDS; 5 minutes at 20°C, 20 minutes at 55°C) to remove excess radio-labeled probe, membranes were exposed to X-ray film (Kodak) for approximately 48 hours. Positive colonies were revealed as exposure sites on the films. Positive colonies were picked with sterile pipette tips into 25 uL of sterile distilled water in 0.2 mL PCR tubes. Plasmid insert DNA was PCR amplified using the M13 primers complementary to the plasmid regions flanking the insert site. Primary test PCR reactions were performed in 40 uL volumes containing: 5 uL of the DNA solution picked from the positive colonies, 800 nM dNTPs, IX PCR reaction buffer (BRL), 0.5 U of Taq DNA polymerase (BRL), 3.0 mM MgCl2, and 5 pM each of the M13 forward (5'TGTAA AACGACGGCCAGT 3') and reverse (5'CAGGAAACAGCTATGACC3') primers. PCR reaction conditions were as follows: 1 cycle of 120 second denaturation at 95°C, 60 second annealing at 52°C and 90 second extension at 72°C; 4 cycles of 60 denaturation at 94°C, 60 second annealing at 52°C and 90 second extension at 72°C; 25 cycles of 60 second denaturation at 94°C, 60 second annealing at 50°C and 90 second extension at 72°C; and 5 minute extension at 72°C. After a successful test PCR, reactions were scaled up to 150 uL for sequencing. 3.23 Sequencing PCR products were purified for sequencing using Quiagen "Quiaquick" PCR purification columns and eluted into 50 uL of sterile distilled water. Sequencing PCRs 62 were performed in 20 uL volumes containing Taq termination premix (8 uL containing Taq DNA polymerase, reaction buffer and ddNTPs) obtained from the NAPS Unit at UBC, 1 uL 0.5% DMSO, 3.2 pM M13 forward primer, approximately 100 ng purified DNA and sterile distilled water to make up the total volume. Sequencing PCR conditions were as follows: 1 cycle of denaturation for 4 minutes at 95°C and 25 cycles of denaturation for 30 seconds at 95°C, annealing for 15 seconds at 55°C and extension for 4 minutes at 60°C. Excess primers and ddNTPs were removed from the PCR product with Centri-Sep (Applied Biosystems) columns. Product was sent to a sequencing lab (NAPS Unit, UBC) for sequencing of amplified eulachon insert DNA. 3.24 Primer Design Twenty-two putative microsatellite sequences were obtained. Of these, 12 were rejected: 4 sequences were unusable (sequence too poor to design reliable primers, many ambiguous bases), 2 were false positives (no microsatellite in sequence), one was too small (two non-tandem repeats), and 5 sequences contained microsatellites that were too close to the insert site (not enough flanking region to design primers). Ten primer sets were designed. Primers for eulachon microsatellite amplification were designed through a combination of three methods: the program OSP (Hillier and Green 1991), the program Primer 3 (Rozen 1998) and "by eye." Primers pairs were designed to minimize self-complementation and primer-dimer formation, and to have similar annealing temperatures. Standard nomenclature for microsatellite loci is as follows. The first letter of the locus name is the first letter of the species' genus. The second two letters of the 63 locus name are the first two letter of the species. The end number or letter is the name of the original clone. Three primer sets used in this study were not isolated from eulachon DNA. Osmol6, OsmoS5 and Osmo\51 were contributed by Robert St-Laurent (Department de Biologie, Laval Universite), who developed primer sequences and refined PCR conditions for a microsatellite assay of rainbow smelt, Osmerus mordax. Table 3.1 Eulachon microsatellite loci. Given for each locus are the locus name, repeat array, range of allele sizes and number of alleles. Repeat arrays are given for the species in which each microsatellite was developed. LOCUS REPEAT TYPE OF SIZE #OF N A M E A R R A Y REPEAT RANGE (bp) A L L E L E S Osmo\6 (GT) I 5 dinucleotide 85-101 9 Osmo 85 (CA) 1 0 AAA(CA),4 dinucleotide 146-156 6 OsmolSl (GT) 3 3 AAGA(GT) 8 dinucleotide 89-93 3 TPaU (GT) 4(CA) 2GACT(G) 2(GT) i 2(G) 4GT single nucleotide 129-144 10 TPa26 (AC) 6AA(AC) 7TG(AC)4AT(C)9(AC)7 dinucleotide 273-287 8 64 3.25 Primer Optimization Cold (non-radiolabeled primers) PCRs were used to find the optimal annealing temperatures and extension times, as well as the ideal MgCl2 concentrations for each primer set. Cold PCR products were run on 1% agarose gels in 0.5X TBE buffer. When cold PCR products were acceptable (no extra bands or smearing), 3 2P end-labeled primers were used in a hot PCR reaction. Of the 10 primer sets designed, 6 gave unresolvable PCR products, 2 worked inconsistently and were not included, and 2 worked well (see Table 3.1 for microsatellite array type and Table 3.3 for primer sequences). 3.26 Primer End-labeling For the hot PCR, one of the primers of each set was end-labeled. The forward primer was labeled first, and if the product was unsatisfactory, the reverse primer was then labeled instead (as in the case of TPal4). The terminal phosphate group on the 5' end of the primer was replaced with 3 2P in a one microlitre reaction (per sample) containing: IX reaction buffer (New England Biolabs), 0.25 U PNK (New England Biolabs), 0.5 uM primer, 0.024 uL 3 2P and 0.802 uL water. The reaction was incubated at 37°C for one hour and then at 65°C for five minutes to denature the enzyme. 3.27 Samples used and PCR Amplification The samples used in the microsatellite assay were the same as the samples used in the mtDNA assay, except for some of the British Columbian samples (FR, FN, KK, GC). These DNA extractions were used up in the mtDNA assay and there was no extra tissue for re-extraction. The total sample size for the microsatellite assay was 293 (Table 3.2). 65 Table 3.2 Collection locations and descriptions, including sample sizes. C O L L E C T I O N Y E A R S A M P L E S A M P L E L O C A T I O N D E S I G N A T I O N SIZE Columbia River 1997 C O L 30 Columbia River 1998 C L 30 Cowli tz River 1997 C O W 30 Cowli tz River 1998 C W 30 Queets River 1998 QT 30 Fraser River 1995 F R 10 Knight Inlet K N Franklin River 1995 F N 10 Klinakl in i River 1995 K K 18 Gardner Canal 1995 G C 15 Nass River 1997 NS 30 Cook Inlet 1997 C K 30 Bering Sea 1997 B R 30 Table 3.3 Primer sequences (5' to 3') and P C R annealing temperatures and M g C l 2 concentrations for microsatellite loci. Asterisks (*) denote 3 2P-labelled primers. Primer sequences are unavailable for Osmo loci. L O C U S P R I M E R A N N E A L I N G [MgCl 2 ] N A M E S E Q U E N C E T E M P E R A T U R E (°C) (mM) Osmo\6 *p R 55 1.5 Osmo&5 *p R 53 1.5 Osmo 157 *p R 55 1.5 TPaU F A G A G G C G C A G A T G A A G A G *R C A C G T T G C C G T G G T A A T G C 62 2.5 TPa26 *F A G G A C T G G C G T G G G A A A T R C T G C A C T G C T G T C T G G A G A A 68 2.5 66 P C R amplification was performed in a 10 uL total volume consisting of the following: 100 ng eulachon genomic D N A , I X P C R buffer ( B R L ) , 0.5 U Taq D N A polymerase, varying concentrations of M g C l 2 (see Table 3.1), 200 u M each of d A T P , dCTP, d G TP and dTTP, 6 p M unlabeled primer, 0.5 p M 3 2 P end-labeled primer and 2.5 p M of the unlabeled primer that was subject to end-labeling. P C R cycling varied with the primer set used (see Table 3.1), but followed a general pattern of one cycle of 3 minutes denaturation at 95°C, 60 seconds at a locus-specific annealing temperature and 60 seconds of extension at 72°C; five cycles of 60 seconds denaturation at 94°C, 60 seconds at a locus-specific annealing temperature and 60 seconds extension at 72°C; 25 to 35 cycles of 60 seconds denaturation at 92°C, 60 seconds at a locus-specific annealing temperature and 60 seconds extension at 72°C; and five minutes extension at 72°C. Osmerus microsatellite loci were amplified in a total reaction volume of 12.5 u L containing I X P C R reaction buffer (BRL) , 1.2 m M M g C l 2 , 150 u M each of d A T P , dCTP, d G T P and dTTP, 6 p M of the unlabeled primer, 0.5 p M 3 2 P end-labeled primer and 2.5 p M of the unlabeled primer that was subject to end-labeling, 0.25 U Taq D N A polymerase, 100 ng template D N A and sterile distilled water. P C R cycling was as follows: one cycle of five minutes denaturation at 96°C; five cycles of 60 seconds denaturation at 95°C, 30 seconds annealing at a locus-specific annealing temperature (see Table 3.1) and 60 seconds extension at 72°C; 25 cycles of 30 denaturation at 96°C, 30 seconds annealing at a locus-specific annealing temperature and 60 seconds extension at 72°C; and one 5 minute extension at 72°C. 67 3.28 Electrophoresis and Scoring of Variation Approximately 7 uL of formamide based loading buffer (US Biochemicals sequencing stop buffer) was added to each completed 10 uL PCR reaction. Samples were heat denatured for 15 minutes and kept on ice until loaded. Five uL of each denatured sample was loaded onto a 6% denaturing TBE urea-polyacrylamide gel. As there were 62 lanes on each gel, and 4 were reserved for the ladder in order to determine allele size, 58 samples could be loaded on each gel. Some samples were re-run on more than one gel to ensure consistency in scoring. Gels were electrophoresed vertically at 55-60 W for 2-5 hours depending on the size of the fragment. Larger fragments were run for longer times to increase resolution of alleles. The gels were dried on 3 M M Whatman filter paper and exposed overnight to X-ray film (Kodak). Alleles were scored and named according to size, which was determined by the M l 3 sequencing reaction ladder. 3.29 Data Analysis The GENEPOP (version 3.0) software package (Raymond and Rousset 1995) was used to calculate allele frequencies at each locus, total genotype frequencies across loci, linkage disequilibrium among loci, and exact tests of population differentiation. Conventional F-statistics (Weir and Cockerham 1984), variance components and diversity per population were computed using the program "Arlequin" (version 1, Schneider et al. 1996). F-statistics follow the infinite alleles model of microsatellite evolution and so do not take alleles sizes into account, but simply identity versus non-identity. R-statistics, analogues of F-statistics, follow the stepwise mutation model and 68 make use of allele size information. R S T values were calculated with the program " R S T CALC" (Goodman 1997), and the statistical significance of the FSTand R S j values was calculated using the program "FSTAT" (Goudet 1995). The "groups" used for comparison in these analyses were the same as those defined in the mtDNA section and were tested for population differentiation of microsatellite data with the GENEPOP software package. Arlequin was also used to calculate gene diversity (the expected heterozygosity summed over loci) and average gene diversity (the expected heterozygosity averaged over loci). The program "TFPGA" (Miller 1997) was used to calculate observed and expected heterozygosities for each locus, departures from Hardy-Weinberg equilibrium at each locus and the significance level of a Mantel test (Mantel 1967, Sokal 1979) for matrix correlation. The program "Bottleneck" tests the frequency distributions of alleles at microsatellite loci for recent (within the past several dozen generations) bottlenecks, or reductions in effective population size (Cornuet and Luikart 1996) and was used to test whether eulachon populations have undergone recent bottlenecks. Each sample site was tested individually, each "group" of sample sites was tested, and all samples combined were tested. The basis of this test is the equilibrium between mutation and random genetic drift, and the resulting allele frequency distribution. Mutation-drift equilibrium at selectively neutral loci such as microsatellite loci results in an allele frequency distribution that is "L-shaped" or that has many alleles at low frequencies. These low frequency alleles (less than 10%) will be more abundant than alleles of intermediate frequency, regardless of the mutation model or rate assumed. If a population experiences 69 a genetic bottleneck, low frequency alleles will be rapidly lost and the frequency distribution will no longer be L-shaped. Instead, intermediate frequency alleles become the most common. Deviations from the L-shape are detected by "Bottleneck". 70 TPa26 Osmo85 OS/770157 Figure 3.1 Microsatellite gels. One representative gel from each locus is shown. Each lane represents an individual fish, and sample site designations are shown above the samples. 73 3.3 Results 3.31 Microsatellite Polymorphism All microsatellite loci screened in this study were polymorphic (four at 0.95 criterion and one at 0.99 criterion) and each had a distinct repeat array, although all but TPalA were dinucleotide repeats (Figure 3.1). The number of alleles at each locus varied from 3 to 10, and averaged 7.2. The size ranges at each locus did not exceed 16 base pairs (Table 3.1). 3.32 Microsatellite Allele Frequency Distribution The number and frequency of alleles at each of the 5 loci were similar at all sample sites. At all five loci, all sample sites shared the most common allele. Private alleles were observed at two loci, TPalA and TPa26, but not in the same populations. Interesting trends were observed at two loci: at Osmol6, the largest allele was only found in Alaskan populations (CK and BR); and at Osmo\51, the smallest allele was found only in the Columbia and Cowlitz rivers in the 1997 samples. At two loci, Osmo%5 and Osmo\51', some populations contained all alleles. The number of alleles per population at each locus was: Osmol6, 4-8; OsmoS5, 2-6; Osmo\51, 1-3; TPalA, 2-6; and TPa26, 1-4 (Table 3.4 and Figure 3.2). The highest diversity, both by number of alleles and levels of heterozgosity, was found in the Columbia system (Columbia River both years and Cowlitz River both years combined). 74 Table 3.4 Allele frequencies at each sample site for each of five microsatellite loci. Osmo 16 C O L C L C O W C W Q T FR F N K K G C N A S S C K B R A L L E L E SIZE 85 0.093 0.09 0.11 0.04 0.03 0 0.11 0 0.07 0 0 0 87 0.037 0.09 0.15 0 0.06 0 0 0.03 0.10 0.04 0 0.09 89 0 0 0 0.02 0 0 0 0 0.04 0 0 0.11 91 0.556 0.57 0.57 0.46 0.42 0.50 0.78 0.75 0.68 0.52 0.70 0.52 93 0.185 0.09 0.06 0.11 0.17 0.15 0 0.09 0.04 0 0.04 0.05 95 0.056 0.07 0.06 0.07 0.10 0.05 0.05 0 0 0.17 0.14 0.07 97 0.056 0.05 0.04 0.20 0.15 0.25 0.05 0.13 0.07 0.26 0.12 0.11 99 0 0 0 0.09 0.02 0.05 0 0 0 0.02 0 0.05 101 0.019 0.05 0.02 0 0.06 0 0 0 0 0 0 0 OsmoH5 C O L C L C O W C W Q T FR F N K K G C N A S S C K B R A L L E L E SIZE 146 0.019 0 0 0 0 0.05 0.05 0 0 0 0 0 148 0.019 0 0 0.03 0 0 0 0 0.03 0 0.02 0 150 0.019 0 0.02 0.02 0.02 0.05 0.3 0.14 0.10 0 0 0.02 152 0.885 0.96 0.96 0.90 0.93 0.85 0.6 0.81 0.77 0.93 0.94 0.94 154 0.019 0 0.02 0.05 0.02 0.05 0 0 0.03 0.07 0.02 0.02 156 0.038 0.04 0 0 0.03 0 0.05 0.06 0.07 0 0.02 0.02 75 Osmo 157 C O L C L C O W C W Q T FR F N K K G C N A S S C K B R A L L E L E SIZE 89 0.02 0 0.02 0 0 0 0 0 0 0 0 0 91 0.98 0.96 0.93 0.95 1 1 1 0.97 0.97 1 0.98 1 93 0 0.04 0.05 0.05 0 0 0 0.03 0.03 0 0.02 0 TPal4 C O L C L C O W C W Q T FR F N K K G C N A S S C K B R A L L E L E SIZE 129 0 0 0 0 0.02 0 0 0 0 0 0 0 130 0 0 0.02 0 0 0 0 0 0 0 0 0 134 0.087 0 0.28 0.22 0.13 0.05 0.10 0.08 0.10 0.13 0 0.19 135 0.67 0.50 0.70 0.78 0.69 0.90 0.90 0.91 0.90 0.57 0.09 0.74 136 0.22 0.47 0 0 0.15 0 0 0 0 0.30 0.63 0.02 137 0.02 0 0 0 0.02 0.05 0 0 0 0 0.24 0.02 138 0 0 0 0 0 0 0 0 0 0 0.02 0 141 0 0 0 0 0 0 0 0 0 0 0.02 0.02 143 0 0 0 0 0 0 0 0 0 0 0 0.02 144 0 0.03 0 0 0 0 0 0 0 0 0 0 76 TPa26 C O L C L C O W C W Q T FR F N K K G C N A S S C K B R A L L E L E SIZE 273 0 0.05 0 0 0 0 0 0 0 0 0 0.02 275 0 0 0 0 0 0 0 0 0 0.02 0 0.04 277 0.02 0 0 0 0 0 0 0 0 0 0 0 279 0 0 0.02 0 0 0 0 0 0 0 0 0 281 0 0 0.02 0.02 0 0 0 0.03 0.03 0 0.02 0 283 0.94 0.86 0.96 0.98 0.98 0.9 1 0.94 0.97 0.89 0.90 0.94 285 0.04 0.07 0 0 0.02 0.1 0 0.03 0 0.07 0.08 0 287 0 0.02 0 0 0 0 0 0 0 0.02 0 0 79 Alleles at Osmo 157 locus Figure 3.2 Allele frequencies at each sample site for each of five microsatellite loci. For exact values, refer to Table 3.4. 80 3.33 Heterozygosity The observed heterozygosity (over all sample sites) ranged from 4.6% (Osmo\51) to 61% (Osmo\6) with an average over all loci of 28%. The expected heterozygosity ranged from 4.6% to 64% with an average of 29%. Heterozygosities at three loci were slightly lower than expected (Osmol6, TPalA and TPa26), and were higher than expected at two loci (0^085 and OsmolSl), although no locus showed a significant deviation from expected values. Over all samples and all loci, the observed and expected heterozygosities were one percent different (Table 3.5). Table 3.5 Observed and expected heterozygosities for each microsatellite locus. Observed heterozygosities are direct counts, and expected heterozygosities are corrected, unbiased estimates. LOCUS N A M E OBSERVED HETEROZYGOSITY EXPECTED HETEROZYGOSITY Osmo\6 0.6080 0.6445 Osmoc\5 0.2083 0.1950 Osmo\51 0.0469 0.0460 TPaU 0.4598 0.4622 TPa26 0.0952 0.1219 Overall 0.2836 0.2939 81 3.34 Hardy-Weinberg Equi l ibr ium Tests for Hardy-Weinberg equilibrium (HWE) were performed for each locus and for each population (60 tests), and revealed that only two of the 60 tests performed were significantly different from HWE expectations (results not shown). These were the Columbia River 1998 and the Nass River samples, both for the Tpa26 locus. With a Bonferroni correction for multiple comparisons, however, the P-values (0.0276 and 0.0164, respectively) were not significant. All loci in all populations exhibited HWE. 3.35 Linkage Disequilibrium Linkage disequilibrium occurs when loci are not segregating independently. For example, alleles at two loci found very close together on the same chromosome may be transmitted in association from parent to offspring. This does not allow recombination to occur, and alleles at these loci will remain in the same combinations over time. A test for linkage disequilibrium among eulachon microsatellite loci showed that for all locus pairs, there was no linkage disequilibrium (Table 3.6). All loci, therefore, appear to be independently inherited. 82 Table 3.6 Probabilities for chi-square tests (Fisher's method) of linkage disequilibrium. Tests were performed for each locus pair (first column) across all sample sites. No tests were significant and no locus pairs exhibited linkage disequilibrium. LOCUS PAIR P-VALUE 0577*016 and Osmo85 0.737 Osmo 16 and Osmo 157 0.683 Osmo 16 and 7Pal4 0.937 Osmo 16 and TP aid 0.841 Osmo85 and OsmoXSl 0.958 Osmo%5 and TPalA 0.245 OsmoS5 and TPa26 0.914 Osmo\51 and TPalA 0.808 OsmolSl and TPa26 0.714 TPalA and TP old 0.976 3.36 Microsatellite Diversity Within Sample Sites Gene diversity (Nei 1987) at each sample site was high, ranging from 0.8032 (Klinaklina River 1995) to 0.9605 (Columbia River 1998), with an average of 0.8973 (Table 3.8). Average gene diversity over loci ranged from 0.1312 (Bering) to 0.2726 (Fraser River) and averaged 0.232 (Table 3.7). 83 Table 3.7 Gene diversity and average gene diversity over loci for each sample site. Gene diversity is the expected heterozygosity summed over loci, and average gene diversity is the expected heterozygosity averaged over loci. S A M P L E SITE G E N E D IVERSITY (SD) A V E R A G E G E N E D I V E R S I T Y (SD) C O L 0.9019 (0.0342) 0.2513 (0.1787) C L 0.9605 (0.0115) 0.2032 (0.1812) C O W 0.8955 (0.0306) 0.2283 (0.1667) C W 0.9468 (0.0141) 0.2518 (0.1788) Q T 0.9383 (0.0220) 0.2240 (0.1643) FR 0.8684 (0.0540) 0.2726 (0.1962) F N 0.9368 (0.0427) 0.2711 (0.2077) K K 0.8032 (0.0579) 0.1990 (0.1528) G C 0.8161 (0.0725) 0.2386 (0.1749) N A S S 0.9260 (0.0184) 0.2631 (0.1848) C K 0.8539 (0.0445) 0.2512 (0.1787) B R 0.9198 (0.0235) 0.1312 (0.1235) 3.37 Microsatellite Diversity Among Sample Sites 3.37.1 Temporal Population Structure Temporal differences at the two sites that were sampled for two years were tested using an exact test of population differentiation (Excoffier et al. 1992) with a Bonferroni correction for multiple comparisons. Neither null hypothesis of no genetic differentiation was rejected (Columbia P=0.009, Cowlitz P=0.0243), which allowed these samples to be grouped to increase the power of further analysis (Table 3.8). 84 Table 3.8 Exact tests of population differentiation. Al l possible pairwise comparisons were performed (across loci) and P-values were compared to Bonferroni corrected alpha values. Asterisks (*) denote significant tests. The final column lists the number of loci that support the significant differences at the Bonferroni corrected alpha level. For each significant test, all five loci were individually supportive of the results at an alpha of 0.05 Sample designations are as in Table 3.2. C O M P A R I S O N P - V A L U E S U P P O R T I N G L O C I F N X G C 0.8836 FR X G C 0.8827 C O L X Q T 0.7747 F N X K K 0.7464 FR X K K 0.6641 N A S S X C K 0.4753 Q T X FR 0.4527 C O L X G C 0.3862 FR X B R 0.3857 C O L X FR 0.3657 G C X B R 0.3318 FR X F N 0.2738 C O L X C K 0.2731 C W X B R 0.2085 C W X F R 0.1913 Q T X B R 0.1410 FR X C K 0.1270 Q T X C K 0.1023 G C X B R 0.0959 Q T X G C 0.0860 C O W X B R 0.0682 C O W X F N 0.0680 K K X B R 0.0478 C O L X F N 0.0411 C O L X K K 0.0395 Q T X K K 0.0303 C W X F R 0.0300 C W X Q T 0.0298 C O W X C W 0.0243 C O W X FR 0.0206 Q T X F N 0.0198 F N X B R 0.0197 Q T X N A S S 0.0194 85 FR X N A S S 0.0164 C O W X G C 0.0156 C O W X FR 0.0130 G C X C K 0.0104 C O L X C L 0.0090 C W X G C 0.0062 C W X Q T 0.0053 C L X G C 0.0044 C L X C K 0.0044 C O L X B R 0.0035 C W X F N 0.0033 C O L X C O W 0.0031 K K X C K 0.0020 C O L X N A S S 0.0020 C K X B R 0.0011 C L X FR 0.0009* 2 F N X C K 0.0009* 3 C L X C O W 0.0008* 1 C L X Q T 0.0006* 2 C W X C K 0.0002* 2 C L X B R 0.0001* 1 C L X N A S S 0.0001* 1 C O L X C W 0.0001* 1 N A S S X B R 0.0001* 2 F N X N A S S 0.0001* 3 C L X F N 0.0000* 3 C L X C W 0.0000* 2 C O W X N A S S 0.0000* 2 C O W X C K 0.0000* 2 C O W X N A S S 0.0000* 1 C W X N A S S 0.0000* 1 K K X N A S S 0.0000* 2 G C X N A S S 0.0000* 1 3.37.2 Spatial Population Structure Similar tests with a Bonferroni correction were performed to see if sample sites within inlets could be grouped. The Columbia samples (both years) could be grouped with the Cowlitz samples (both years) (P=0.0115, not shown above). Samples within 86 Knight Inlet (KK and FN) were not statistically different (P=0.7464) and the Alaskan samples (CK and BR) could also be grouped (P=0.0011). Sample sites within Gardner Canal were not tested with microsatellite data because different samples were used than in the mtDNA study, and there were so few of them. An AMOVA revealed that approximately 0.85% of the total genetic variation was partitioned into the "among region" category, while more than 95% of the total genetic variation was "within populations" or sample sites (P=0.000; Table 3.9). Although little variation was partitioned "among populations within regions," this value was significant (P=0.000; Table 3.9). Alternate groupings of the sample sites (for example the Bering Sea versus all other sites, which for mtDNA increased the percent of variation existing among regions) did not give significant values, or values higher than 3.6% for the "among region" category. Table 3.9 Results of AMOVA test for partitioning of eulachon microsatellite DNA variance. Collections were grouped geographically into 7 regions based on the chi-square tests for temporal and geographic heterogeneity. The regions are: Columbia and Cowlitz River (both years), Queets River, Fraser River, Knight Inlet, Gardner Canal, Nass River and Cook Inlet/Bering Sea. SOURCE OF PERCENTAGE OF P-VALUE VARIATION VARIATION Among regions 0.85% 0.301 Among populations 3.75% 0.000 within regions Within populations 95.40% 0.000 87 F S T values were calculated for each locus independently and over all loci. Values ranged from 0.0013 {OsmolSl) to 0.0939 {TPalA) and averaged 0.0470 when each sample site was considered as a population (Table 3.10). When samples were pooled into the groups defined above, the overall F S T value was 0.045. FSTcalculations assume that all alleles are equally related to one another, and follow the IAM. Because microsatellites may follow a mutation model more like the SMM, in which the relationships between alleles are determined by their length, R S T values are often calculated to determine the extent of population sub-division for microsatellite loci. Table 3.10 FST and R S T values for each locus and overall. P-values show the probability that the F S T and R S T values are significantly different from zero. "NT" means that the significance of the test was not determined. LOCUS NAME F S T VALUE P-VALUE RST VALUE P-VALUE Osmol6 0.0242 0.005 0.0894 NT Osmo&5 0.0467 0.005 -0.0070 NT Osmo 157 0.0013 0.285 0.0009 NT TPalA 0.0930 0.005 0.1121 NT TPa26 0.0015 0.410 0.0080 NT OVERALL 0.0470 0.005 0.0798 0.0000 3.38 Isolation-by-distance A graph of genetic distance (distance-based, Reynolds et al. 1983) versus geographic distance (km) (Figure 3.3) does not show a trend similar to that expected in a typical isolation-by-distance model. In fact, the trend appears to be slightly negative. A Mantel test for correlation of matrices does not show significance (P=0.56). This drift-based measure of genetic distance does not take the mutation model into account when calculating genetic distance. Other measures of genetic distance based on the mutation models, the IAM and SMM, revealed similar trends and were not significant either. 3.39 Recent Population Bottlenecks A normal L-shaped distribution of allele frequencies was observed at all sample sites, groups of sample sites, and in the total combined sample (details not shown). This result indicates that eulachon have not undergone recent genetic bottlenecks, at least of severe enough magnitude to be resolved by shifts in the allele frequency distribution. 89 0.15 £ 0.05 1 -0.05 Geographic Distance (km) Figure 3.3 Isolation-by-distance of eulachon sample sites. 90 3.4 Discussion 3.41 Glacial Refugia and Post-glacial Dispersal The single glacial refuge proposed by McPhail and Lindsey (1970) is supported by the microsatellite data. Eulachon likely survived the Wisconsinan glaciation in a single, southern refuge: the Columbia system and possibly further south. Older populations, or regions in which species have existed for longer times, generally contain the highest neutral genetic diversity throughout the current geographic range of the species (Crandall and Templeton 1993). The highest diversity, both in terms of heterozygosity and highest number of alleles per locus, is found in the Columbia system (including two years of samples from the Columbia River and two years of samples from the Cowlitz River, which is tributary to the Columbia). If the southern refuge was the only region in which eulachon remained during the extensive ice coverage, then post-glacial dispersal was northwards up the coast, around and through the Aleutian Peninsula and island chain, and into the Bering Sea. Fish species which survived glaciation in two distinct refugia, the Columbia and Beringia, often show a marked latitudinal split in their allele frequencies. These splits usually occur along the coast of British Columbia, where two distinct glacial "populations" have made post-glacial contact (for example, see Wilson et al. 1987, chinook salmon; Taylor et al. 1994, chum salmon; Varnavsakya and Beacham 1992, pink salmon; Taylor et al. 1996, sockeye and kokanee salmon). The only trends of this type that were observed were at the microsatellite loci Osmol6 and OsmolSl. The largest allele (101) at the locus Osmol6 was only present in the Alaskan populations, but at extremely low frequencies (Table 3.4 and Figure 3.2). This minor bias does not warrant the suggestion of a second refuge in Beringia. At the locus Osmo\51', the smallest allele (89) was only present in the Columbia River and Cowlitz River samples from 1997. Again, the frequencies were extremely low, and may have more bearing on the recent population dynamics of eulachon than on their glacial history. 3.42 Eulachon Microsatellite D N A Variability and Genetic Population Structure Genetic variation in eulachon at five microsatellite loci was surprisingly low relative to other species in which microsatellite DNA has been examined. The average number of alleles per locus was 7.2 and ranged from 3 to 10. Studies of microsatellite variation in other salmonids have shown the number of alleles observed at each locus to be much higher (up to 45 in Atlantic cod, Ruzzante et al. 1997, up to 30 in Pacific herring, O'Connell et al. 1998), however, similar values to those seen in eulachon have been observed in turbot (4-11, Estoup et al. l99Sb), bluegill sunfish (4-8, Colbourne et al. 1996) and cutthroat trout (7-13, Condrey and Bentzen 1998). The maximum size range of alleles at any locus was 16 base pairs, which is approximately one-fifth the size of the average maximum size range per locus in a number of other studies (75 base pairs is the approximate average maximum size range from the following studies: Brooker et al. 1994, Colbourne et al. 1996, Morris et al. 1996, Ruzzante et al. 1996, Ruzzante et al. 1997, Condrey and Bentzen 1998, Estoup et al. 1998/3, O'Connell et al. 1998, Wenburg et al. 1998). The average observed and expected heterozygosity over loci was 28% and 29%, respectively. The range of heterozygosity values among loci was 4% to 64%. 92 Heterozygosity values generally range from 65% to 95% in other species (Brooker et al. 1994, Colbourne et al. 1996, Morris et al. 1996, Ruzzante et al. 1996, Ruzzante et al. 1997, Condrey and Bentzen 1998, Estoup et al. 1998b, O'Connell et al. 1998, Wenburg et al. 1998), although some comparably low values have been observed (as low as 14%, but average of 88% at other five polymorphic loci, Brooker et al. 1994; 25% to 35% at 10 loci in various salmonids, Presa and Guyomard 1996). Although the levels of heterozygosity and allele number were so low, there was enough variation present to examine any potential genetic differences among sample sites and attempt to define populations. Because all microsatellite loci were in Hardy-Weinberg equilibrium and there was no evidence of linkage disequilibrium between any locus pairs, all five loci could be used to investigate the genetic population structure of eulachon. Exact tests of population differentiation (Table 3.8) were used to test whether samples sites among years within sites, and among rivers within an inlet could be grouped. The microsatellite data supported the same seven groups as were used for the mtDNA analysis: the Columbia system, Queets River, the Fraser River, Knight Inlet, Gardner Canal, the Nass River and the Alaskan samples. Significant values in Table 3.8 were generally supported by three or fewer loci, and because they did not make geographic sense, biological significance was not inferred. Further support for two of these groups was mentioned in the previous section. The Alaskan samples, Cook Inlet and the Bering Sea, share an allele at the Osmo 16 microsatellite locus that is not present in any other sample. Also, the Columbia system contains an allele at locus Osmo 157 that is not present in any other river. 93 The AMOVA showed that an extremely small amount of variation was partitioned into the "among regions" category, and that this value was not statistically significant. More than 95% of the total neutral genetic variation was partitioned within populations, or sample sites (Table 3.10), and this value was significant. These results correspond with both the mitochondrial DNA results and 'wandering' habits of eulachon, and show that eulachon populations seem to be structured like marine fish populations. An unusual result, however, is that the small amount of variation partitioned into the "among populations within regions" category was statistically significant. Although it is significant, it is less than 4% of the amount of variation found within regions, and is likely not biologically significant. How does one decide whether the SMM (and RST) or the IAM (and F S T ) is most appropriate for microsatellite data analysis? Differences in allele frequency distribution, repeat array type (perfect versus imperfect), repeat unit size (di-, tri- or tetranucleotide) and number of repeats are all important (O'Connell et al. 1997). Allele frequency distributions tend to be more similar to a normal distribution under the SMM, and a uniform distribution under the IAM. Eulachon microsatellite allele frequency distributions are more normal than uniform, with one or two very common alleles (Figure 3.2), and from one to nine less frequent alleles. O'Connell et al. (1997) examined 5 loci, all of which were perfect repeats, and determined that the IAM was a better fit. Shriver et al. (1993) determined that tri- and tetranucleotide repeats were served better by the SMM than were dinucleotide repeats and minisatellites (larger repeats). Because the microsatellite loci examined in eulachon don't fit into either of these categories (most are imperfect and dinucleotide repeats), I report and consider both values. 94 Except for Osmoo5, F S T and R S T values give generally the same information at each locus. Excluding the negative R S T value, both estimates of population structure give the same order and same order of magnitude except TPalA for which Rsi-is an order of magnitude higher. Overall the same order of magnitude of population differentiation is seen (Table 3.10). The per locus F S T values ranged from 0.0013 {Osmo\51) to 0.0930 (TPaU). A range of values was expected due to differing levels of polymorphism at different loci. The smallest F S T values were calculated from loci with the fewest number of alleles and lowest heterozygosity (Osmo\51 and TPa26), and these are the values that were not significant. All other F S T values were significant, as was the overall value, 0.0470 (Table 3.10). R S T values per locus include one negative value (OsmooS), and ranged from 0.0009 {Osmo\51) to 0.1121 (7Pal4). The R S T value over all loci, 0.0798, was significant (Table 3.9). Six out of 12 sample sites contained private alleles, which were observed at both TPa loci, but no Osmo locus. No sample contained more than one private allele except Cowlitz River 1997, which contained two. No private alleles exceeded a frequency of one allele per population. Due to the lack of pattern and low number of private alleles observed, these alleles were not useful in determining population sub-division. With respect to the level of population sub-division observed in eulachon, microsatellite DNA analysis is consistent with mtDNA analysis. Both have revealed that eulachon populations are characterized by weak levels of structure, comparable to the levels of sub-division seen in marine species, such as Atlantic cod and Pacific herring, rather than anadromous species such as the coastal cutthroat trout. 95 Population genetic analysis of microsatellite markers has revealed structure in other anadromous species that have colonized this coast post-glacially, for example, coastal cutthroat trout, Oncorhynchus clarki clarki. Microsatellite DNA variation in coastal cutthroat trout was examined at six microsatellite loci by Wenburg et al. (1998). The number of alleles per locus across all samples ranged from 4 to 45 and averaged 24. The maximum size range of alleles was more than 100 base pairs, and the observed and expected heterozygosity ranged from 34% to 81% and 40% to 83%, respectively. Averages for these statistics were 67% (observed heterozygosity over all samples and all loci) and 71% (expected heterozygosity over all samples and all loci). F S T and R S T values were 0.121 and 0.093, respectively. Cutthroat trout have a much higher amount of genetic diversity at their microsatellite loci than do eulachon, as do many other species. Because coastal cutthroat trout are anadromous, one would expect to see a higher degree of structure than in eulachon, and this is true for both F S T and R S T , although the RST value for cutthroat is on the same order of magnitude as the R S T in eulachon. Microsatellite data has been used to examine population structure in marine species also. Microsatellite DNA variation was assayed the same five loci in Atlantic cod, Gadus morhua, in two recent studies (Ruzzante et al. 1996, 1997). A larger sample size was used in the first study, but with the combined data, the number of alleles per locus ranged from 10 to 45 and averaged 30.2. The average observed and expected heterozygosities over all loci were 0.847 and 0.873, respectively. Significant F S T and R S T values in both studies revealed evidence of population sub-division between inshore and offshore populations of Atlantic cod off Newfoundland. Even though eulachon have an extremely low level of variation at the five microsatellite loci examined, measures of population sub-division ( F S T and R S T ) were significant, and similar to what has been observed in marine species such as Atlantic cod and Pacific herring. Pacific herring from the Bering Sea and southeast Alaska were examined at five microsatellite loci (O'Connell et al. 1998). Levels of heterozygosity were again, much higher than those observed in eulachon, and ranged from 86% to 97%, with an average over loci of 89%. Up to 26 alleles were found per locus per population, which is three times higher than the number of alleles observed in any eulachon population at a particular locus. The R S T value for Pacific herring over the geographic range described above was 0.0542, which is on the same order of magnitude as that for eulachon. There are a number of reasons why eulachon may not exhibit the type of genetic population structure typically expected in anadromous species. The recency of population differentiation and historically high effective population sizes (past gene flow); lack of homing or large straying rates (current gene flow); or the fact that populations may not yet be at equilibrium, are all possibilities. Whatever the reason for the lack of river-by-river structure in eulachon, there does appear to be some emerging level of structure and not complete panmixia. Another possible explanation for the lack of expected population structure in eulachon is the extremely low levels of variation observed at microsatellite loci. Why is eulachon microsatellite diversity so low? In a recent examination of the mutation rate of microsatellite loci in fruit flies, Drosophila melanogaster, Schug et al. (1997) discovered suprisingly low mutation rates; up to three orders of magnitude lower than microsatellite mutations rates in humans, mice, pigs and rats (no direct study of mutation accumulation has been completed for fish yet, but values for heterozygosity and 97 estimated mutation rates seem to be generally fairly high). Studies with large numbers of loci have put estimates for the range of microsatellite mutation rates in humans, mice, pigs and rats between 10"3 and 10"5 (Schug et al. 1997), but in a study designed specifically to determine Drosophila mutation rates, a value of 6.3 X 10"6 was discovered. Despite this low mutation rate, however, the levels of variation appear to as high in natural populations of Drosophila as they are in the previously mentioned mammals. Schug et al. (1997)'s explanation for the low mutation rates observed was the length of the repeats in Drosphila, which are, on average, shorter than repeat arrays in other species. The average dinucleotide repeat observed in their study was 13 repeat units (the range was from 7 to 16) which was considered possibly below the minimum threshold for high mutability. Of the five eulachon microsatellite loci assayed, however, the locus with the highest number of repeats (41) had the fewest number of alleles (3). For eulachon, the low levels of variation observed may be explained by the nature of the repeats examined, rather than as a characteristic of the species. All but one locus examined had an interrupted, and in some cases multiply interrupted, repeat array (Table 3.1). The only perfect array was Osmo\6, which also had the highest heterozygosity. Of all the clones that were sequenced during the process of primer creation, only 16% were perfect repeats. It is possible that eulachon microsatellite repeat arrays are more stable than those in other species due to the high number of interruptions, and therefore have a lower mutation rate, leading to fewer numbers of alleles and low heterozygosity. Stabilization of microsatellite loci by interruptions has been suggested in number of studies (see Taylor et al. in press, and references within). In other salmonids, the percentage of perfect repeats observed in the clones sequenced is approximately 60% 98 (Rainbow trout, Oncorhynchus mykiss, Brooker et al. 1994 and O'Connell et al. 1997; Atlantic salmon, Salmo salar, McConnell et al. 1995); much higher than the 16% observed for eulachon. Whatever the reason for the low variation observed at eulachon microsatellite loci, it is not because eulachon have experienced a recent genetic bottleneck. A test for recent genetic bottlenecks showed that eulachon populations have not experienced one, although they have been observed to be undergoing an ecological bottleneck (actual numbers are decreasing, rather than genetic variation). This could be true because either the population reduction is too recent to be detected in the test, or the pattern of historical exponential growth is robust to subsequent population disturbances. 3 . 4 3 Life-history and mechanisms of population sub-division in eulachon Clearly, eulachon population structure does not follow that of a typical anadromous fish species, but is much more like structure in marine species. Due to some unusual features of the life-history of eulachon, populations may be structured by processes associated more with species that have a marine life-history than those than exhibit anadromy. Eulachon spend very little time in freshwater, and because they are poor swimmers as larvae, are washed out of rivers within 24 hours of emergence (Hay 1996). Marine processes may be the key to clarifying eulachon population structure. Reproductive isolation in the ocean can originate and be maintained by five previously discussed mechanisms (Palumbi 1994). The two competing hypotheses for eulachon are isolation-by-distance, under which geographically proximate populations are genetically 99 similar; and the Member-Vagrant Hypothesis, under which marine retention zones form the foundation of distinct populations, and homing occurs to a marine region. Although the mtDNA data supported isolation-by-distance of eulachon groups, the microsatellite data do not. In a cluster analysis of "populations", the most geographically widespread populations, the Columbia and the Bering Sea, were grouped together as the most genetically similar (dendogram not shown). The Mantel test for correlation between genetic and geographic distance did not support isolation-by-distance either. The graph of genetic distance versus geographic distance showed a trend opposite to that expected under an isolation-by-distance model (but was not significant), and the highest percentage of variation partitioned among groups yielded by alternative groupings of the sample sites was still extremely low given the amount of variation that was partitioned within groups. The microsatellite data are not consistent with isolation-by-distance of eulachon populations. Although lack of support for one hypothesis is not evidence for the truth of another, the absence of a pattern consistent with isolation-by-distance leaves the question, how are eulachon populations structured, unanswered. Because the samples within inlets could be grouped (exact population differentiation tests), and the AMOVA and F S T / R S T calculations were significant and consistent with some differences among populations, it is possible that the microsatellite data support the Member-Vagrant Hypothesis. Investigations of this hypothesis in other species have made use of additional biological data such as larval salinity and temperature tolerance, and oceanographic data such as water salinity and temperature, and the location of features such as gyres, all of which create boundaries for the larval retention zones (Bernatchez and Dodson 1990, 100 Bernatchez and Martin 1996). Before these types of studies are performed in eulachon, it may not be possible to confirm the Member-Vagrant Hypothesis for this species, or to determine the geographic location of the larval retention zones. 101 Chapter 4 General Conclusions and Conservation Implications 4.1 Conclusions from analysis of m t D N A and five microsatellite loci 4.11 Glacial refugia and Post-glacial dispersal Both mtDNA and microsatellite results indicate that eulachon survived glaciation in a single southern refuge, as was proposed by McPhail and Lindsey (1970) based on their geographic distribution. Neutral genetic diversity is highest in the Columbia system for both DNA marker classes, suggesting the presence of the species in the south for a longer time than in the north, where eulachon arrived post-glacially. Neither mtDNA nor microsatellites show the sharp break in allele frequencies that generally reflect a pattern of post-glacial recolonization from two refugia. 4.12 Genetic Variability and Population Structure Neither class of DNA marker, of which both are able to resolve post-glacial population structure in anadromous species, revealed typical anadromous structure in eulachon. Instead, both classes of genetic marker gave evidence consistent with the level of structure that would be expected in a marine species. There are a number of possible reasons for this result which have been mentioned previously: eulachon may be homing to natal rivers, but the recency of the Wisconsinan glaciation and high effective population sizes of eulachon populations may have not yet allowed an equilibrium to be reached; eulachon may home to natal rivers but with an high proportion of random straying; eulachon may home to natal rivers with some degree of non-random straying, 102 for example to a river proximate to their natal river (isolation-by-distance); or eulachon may home to an area other than a river, such as a larger retention zone in the ocean (Member Vagrant Hypothesis). Of all Pacific salmon, the pink salmon (Oncorhynchus gorbusha) has a life cycle most similar to that of eulachon. Pink salmon fry emigrate from the rivers and enter the estuary shortly after emergence from their gravel nests. Although these fish do exhibit genetic population structure, they tend to stray at a higher rate (Horrall 1981) than other species of Pacific salmon. This may be because they do not spend as much time imprinting in their early freshwater environment. Eulachon leave their freshwater habitat even earlier than do pink salmon. This would leave them less time for imprinting (if imprinting does occur) and possibly cause imprecise homing, resulting in a degree of population sub-division that matches the levels observed in marine species. Eulachon exhibit a genetic signature of population expansion on an evolutionary time scale (a star phylogeny of haplotypes and a Poisson distribution of pairwise divergences in mtDNA), yet on an ecological timescale, their numbers seem to be declining. A marine species that also exhibits this pattern is the coconut crab, Birgus latro (Lavery et al. 1996). The coconut crab inhabits islands in the Pacific Ocean and has a marine planktonic larval stage. Recent ecological population declines have been recorded for this species, but mtDNA data are consistent with population expansion. The authors suggest two possible explanations for this apparent contradiction, which also apply to eulachon. First, the population reductions may be too recent to be detected in the genetic signatures of eulachon and coconut crab. Drastic eulachon population declines have occurred in the last 10 years, which may be too recent to be seen in 103 mtDNA. Second, the historical population expansion that occurred after the Pleistocene glaciations as eulachon re-colonized the coast may have been so significant that its resultant genetic legacy is robust to current population size fluctuations. The microsatellite data also do not reflect recent population declines. A test for recent genetic population bottlenecks did not reveal that any had occurred within the last few dozen generations. As well, eulachon appear to be declining more rapidly in the southern extent of their range, but the number of alleles and heterozygosity at microsatellite loci does not increase in more northerly populations. 4.13 Life-history and mechanisms of population sub-division Because the life-history of eulachon is perhaps more similar to that of marine species than anadromous species (very little time in freshwater, large population sizes, possibly high dispersal capability and potential for population mixing) and the degree of observed population structure is quite low, mechanisms of population sub-division in the marine environment were considered. Two applicable explanations were selected from Palumbi's (1994) five suggestions and compared. These were isolation-by-distance and the Member-Vagrant Hypothesis. Mitochondrial DNA data were consistent with both hypotheses: a pattern of isolation-by-distance was strongly supported in a graph of genetic versus geographic distance, and some evidence was consistent with the Member-Vagrant Hypothesis. Although a statistically significant level of population sub-division was resolved with microsatellite data, the microsatellite analysis did not support isolation-by-distance. Lack of support for isolation-by-distance at nuclear markers when mitochondrial evidence is supportive, however, is not surprising given that mtDNA has 104 one-quarter the effective population size of nuclear DNA, and is therefore more significantly affected by population processes such as random genetic drift. An allozyme and mtDNA study of gene flow in red drum, which have a marine life-history, exposed a strong isolation-by-distance pattern with mtDNA that was not resolved with nuclear information (Gold et al. 1993). The authors suggested two reasons for this that are also true for eulachon: the isolation-by-distance pattern has a very slight effect, and mtDNA is more sensitive than nuclear genes in reflecting the genetic impact of population sub-division. Although the Member-Vagrant Hypothesis can not be disqualified, the strong pattern of isolation-by-distance seen in mtDNA and the large female effective population sizes are suggestive of isolation-by-distance in eulachon. 4.2 Conservation implications In order to address conservation concerns, I examined the levels of genetic variability present at each sampling site. The high level of genetic variation observed in mtDNA and consistent levels of diversity at microsatellite loci throughout their range show that eulachon have maintained their levels of neutral genetic variation despite the decrease in their numbers. Eulachon populations have apparently been declining since the 1950's (Hay 1996) and in 1994, population crashes in a number of British Columbian rivers were reported (Berry 1996, Gordon 1996, Lewis 1996). These decreases in population size, while reducing the numbers of fish, did not apparently significantly reduce levels of genetic variation, although the link between neutral molecular variation and adaptive variation is weak (Lynch 1996). The high variation found suggests that the 105 long-term adaptive potential of this species has remained unharmed, possibly due to historically large effective population sizes. Moritz (1994) defined two types of units for conservation: evolutionarily significant units (ESUs) and management units (MUs). ESUs are historically isolated from one another and show reciprocal monophyly with mtDNA and a significant divergence in nuclear allele frequencies. Clearly, eulachon consist of a single ESU. Management units, on the other hand, comprise populations that share similar haplotype frequencies, are connected by gene flow, and show significant divergence in haplotype or allele frequencies from other such units. Although MUs represent much shallower divergences than do ESUs, eulachon may only comprise one or a few MUs. No rivers exhibited haplotype or microsatellite allele frequencies from one another, rather, a pattern of isolation-by-distance was observed. The data clearly imply that population structure is only weakly developed in eulachon, a result that represents a challenge to conservation programs based solely on assays of molecular/biochemical genetic variation (cf. Avise 1998&, Waples 1998). Although more sensitive analyses (e.g. larger samples, more variable loci) may well produce more statistically significant genetic differentiation among populations in local geographic areas, the general pattern of relatively subtle population structure is likely to persist given the post-glacial origin of most extant populations of eulachon and their apparent "wandering" behaviour. The potential treatment of eulachon as only a few genetic MUs throughout their range (compared to the number of rivers in which they spawn), however, must be considered in light of the demographic consequences of features of their life history. The apparently high levels of gene flow and unpredictability 106 in determining which rivers will be used for spawning in any particular year suggest that, regardless of the level of genetic structure, habitat preservation of current and recent spawning areas must still be a high priority. Consequently, animals like eulachon represent a classic case where separate management regimes (at least as they relate to habitat preservation) should be applied to individual spawning runs despite the apparent lack of strong genetic population structure. The results also highlight the need for studies to further our understanding of the basic biology and life history of eulachon. For instance, relatively low amounts of gene flow can constrain population divergence at neutral loci, yet these low levels of exchange amongst populations are unlikely to resuscitate declining populations over a human lifespan (Waples 1998). Further, although eulachon appear to be characterized by high levels of genetic (i.e. mtDNA) variation, the absolute numbers of eulachon are declining at an alarming rate throughout their range (Hay 1996). These observations suggest that critical needs for conservation of eulachon, and similarly characterized species, involve knowledge of population sub-division at selected loci that may influence recruitment and productivity in specific environments as well as more direct estimates of exchange rates among populations to supplement inferences from molecular data (Waples 1998). 107 Literature Cited Allendorf F, Ryman N, Utter F (1987) Genetics and Fisheries Management. In: Population genetics and fishery management, (ed. Ryman N, Utter F, eds. University of Washington Press, Seattle. Amos B, Schlotterer C, Tautz D (1993) Social structure of pilot whales revealed by analytical DNA profiling. Science, 260, 670-672. Anonymous (1997) Columbia River and Tributary Smelt Commercial Landings (in thousands of pounds), 1938-1997. Anonymous (1998) Extension Note #32: Eulachon: A significant fish for First Nations communities. Ministry of Forests, Prince Rupert Forest Region. Avise JC, Smith MH (1974) Biochemical genetics of sunfish. I. Geographic variation and subspecific intergradation in the bluegill, Lepomis macrochirus. Evolution, 28, 42-56. Avise JC, Giblin-Davidson C, Laerm J, Patton JC, Lansman RA (1979) Mitochondrial DNA clones and matriarchal phylogeny within and among geographic populations of the pocket gopher, Geomys pinetis. Proceedings of the National Academy of Sciences USA, 76, 6694-6698. Avise JC, Bermingham E, Kessler LG, Saunders NC (1984) Characterization of mitochondrial DNA variability in a hybrid swarm between subspecies of bluegill sunfish (Lepomis macrochirus). Evolution, 38, 931-941. Avise JC, Arnold J, Ball RM, Bermingham E, Lamb T, Neigel JE, Reeb, CA, Saunders, NC (1987) Intraspecific phylogeny: the bridge between population genetics and sytematics. Annual review of Ecology and Systematics, 18, 489-522. Avise JC, Ball MR, Arnold J (1988) Current versus historical population sizes in vertebrate species with high gene flow: a comparison based on mitochondrial DNA lineages and inbreeding theory for neutral mutations. Molecular Biology and Evolution, 5(4), 331-344. Avise JC (1992) Molecular population structure and the bio geographic history of a regional fauna: a case history with lessons for conservation biology. Oikos, 63, 62-76. Avise JC (1994) Molecular markers, natural history, and evolution. Chapman and Hall, New York. 108 Avise JC (1996) Mitochondrial DNA polymorphism and a connection between genetics and demography of relevance to conservation. Conservation Biology, 9, 686-690. Avise JC (1998a) The history and purview of phylogeography: a personal reflection. Molecular Ecology, 7, 371-379. Avise JC (1998Z?) Conservation Genetics in the Marine Realm. Journal of Heredity, 89, 377-382. Bailey KM, Stabeno PJ, Powers DA (1997) The role of larval retention and transport features in mortality and potential gene flow of walleye pollock. Journal of Fish Biology, 51 (Supplement 1), 135-154. Beacham TD (1996) The use of minisatellite DNA variation for stock identification of chum salmon, Oncorhynchus keta. Fisheries Bulletin of the United States, 94, 611-626. Beckmann JS, Weber JL (1992) Survey of human and rat microsatellites. Genomics, 12, 627-631. Bentzen P, Taggart CT, Cook D (1996) Microsatellite polymorphism and the population structure of Atlantic cod (Gadus morhua) in the northwest Atlantic. Canadian Journal of Fisheries and Aquatic Sciences, 53, 2706-2424. Bernatchez L, Dodson JJ, Boivin S (1989) Population bottlenecks: influence on mitochondrial DNA diversity and its effect in coregonine stock discrimination. Journal of Fish Biology, 35 (Supplement A), 233-244. Bernatchez L, Dodson JJ (1990) Mitochondrial DNA variation among anadromous popualtions of cisco (Coregonus artedii) as revealed by restriction analysis. Canadian Journal of Fisheries and Aquatic Sciences, 47, 533-543. Bernatchez L, Guyomard R, Bonhomme F (1992) DNA sequence variation of the mitochondrial control region among geographically and morphologically remote European brown trout populations. Molecular Ecology, 1, 161-173. Bernatchez L, Martin S (1996) Mitochondrial DNA diversity in anadromous rainbow smelt, Osmerus mordax Mitchill: a genetic assessment of the member-vagrant hypothesis. Canadian Journal of Fisheries and Aquatic Sciences, 53, 424-433. Bernatchez L (1997) Mitochondrial DNA analysis confirms the existence of two glacial races of rainbow smelt Osmerus mordax and their reproductive isolation in the St Lawrence River estuary (Quebec, Canada). Molecular Ecology, 6, 73-83. Berry M (1996) Meeting of the BC Eulachon Research Council minutes, November 6. 109 Birky CW, Fuerst P, Maruyama T (1989) Organelle gene diversity under migration, mutation, and drift: equilibrium expectations, approach to equilibrium, effects of heteroplasmic cells, and comparison to nuclear genes. Genetics, 121, 613-627. Borst P, Grivell LA (1981) Small is beautiful - portrait of a mitochondrial genome. Nature, 290, 443-444. Bowcock AM, Ruiz-Linares A, Tomfohrde J, Minch E, Kidd JR, Cavalli-Sforza LL (1994) High resolution of human evolutionary trees with polymorphic microsatellites. Nature, 368, 455-457. Bowen BW, Grant WS (1997) Phylogeography of the sardines (Sardinops spp.): assessing biogeographic models and population histories in temperate upwelling zones. Evolution, 51(5), 1601-1610. Bremer JR, Baker AJ, Mejuto J (1995) Mitochondrial DNA control region sequences indicate extensive mixing of swordfish (Xiphias gladius) populations in the Atlantic Ocean. Canadian Journal of Fisheries and Aquatic Sciences, 52, 1720-1732. Brooker AL, Cook D, Bentzen P, Wright JM, Doyle RW (1994) Organization of microsatellites differs between mammals and cold-water teleost fishes. Canadian Journal of Fisheries and Aquatic Sciences, 51, 1959-1966. Brown WM, George M, Wilson AC (1979) Rapid evolution of animal mitochondrial DNA. Proceedings of the National Academy of Sciences USA, 76, 1967-1971. Brown JR, Beckenbach AT, Smith MJ (1992) Influence of Pleistocene glaciations and human intervention upon mitochondrial DNA diversity in white sturgeon (Acipenser transmontanus) populations. Canadian Journal of Fisheries and Aquatic Sciences, 49, 358-567. Brunner PC, Douglas MR, Bernatchez L (1998) Microsatellite and mitochondrial DNA assessment of population structure and stocking effects in Artie charr Salvelinus alpinus (Teleostei: Salmonidae) from central Alpine lakes. Molecular Ecology, 7, 209-233. Colbourne JK, Neff BD, Wright JM, Gross MR (1996) DNA fingerprinting of bluegill sunfish using (GT)n microsatellites and its potential for assessment of mating success. Canadian Journal of Fisheries and Aquatic Sciences, 53, 342-349. Condrey MJ, Bentzen P (1998) Characterization of coastal cutthroat trout microsatellites and their conservation in other salmonids. Molecular Ecology, 7, 783-792. 110 Cournet JM, Luikart F (1996) Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics, 144, 2001-2014. Crandall KA, Templeton AR (1993) Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics, 134, 959-969. Cronin MA, Amstrup SC, Garner GS, Vyse ER (1991) Interspecific and intraspecific mitochondrial DNA variation in North American bears (Ursus). Canadian Journal of Zoology, 69, 2985-2992. Davidson WS, Birt TP, Green JM (1989) A review of genetic variation in Atlantic salmon Salmo salar L., and its importance for stock identification , enhancement programmes and aquaculture. Journal of Fish Biology, 34, 547-560. DiRienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin M, Freimer NB (1994) Mutational processes of simple sequence repeat loci in human populations. Proceedings of the National Academy of Sciencees USA, 91, 3166-3170. Epifanio JM, Smouse PE, Kobak CJ, Brown CL (1995) Mitochondrial DNA divergence among populations of American shad (Alosa sapidissima): how much variation is enough for mixed-stock analysis? Canadian Journal of Fisheries and Aquatic Sciences, 52, 1688-1702. Estoup A, Rousset F, Michalakis Y, Cornuet JM, Adriamanga M, Guyomard R (1998a) Comparative analysis of microsatellite and allozyme markers: a case study investigating microgeograpic differentiation in brown trout (Salmo trutta). Molecular Ecology, 7, 339-354. Estoup A, Gharbi K, SanCristobal M, Chevalet C, Haffray P, Guyomard R (1998Z?) Parentage assignment using microsatellites in turbot (Scophthalmus maximus) and rainbow trout (Oncorhynchus mykiss) hatchery populations. Canadian Journal of Fisheries and Aquatic Sciences, 55, 715-725. ExcoffierL (1993) Analysis of molecular variance. Version 1.55. Genetics and Biometry Laboratory, University of Geneva. FelsensteinJ (1993) PHYLIP (phylogeny inference package). Version 3.5. University of Washington, Seattle. Ferris SD, Sage RD, Huang CM, Nielsen JT, Ritte U, Wilson AC (1983) Flow of mitochondrial DNA across a species boundary. Proceedings of the National Academy of Sciences USA, 80, 2290-2294. Glenn TC (1995) Microsatellite manual version 5. Available electronically via ftp:onyx.si.edu/protocols/MsatManV5.rtf. Gold JR, Richardson LR, Furman C, King TL (1993) Mitochondrial DNA differentiation and population structure in red drum (Sciaenops ocellatus) from the Gulf of Mexico and Atlantic Ocean. Marine Biology, 116, 175-185. Gold JR, Richardson LR (1994) Genetic distinctness of red drum (Sciaenops ocellatus) from Mosquito Lagoon, east-central Florida. Fishery Bulletin, 92, 58-66. Gold JR, Richardson LR (1998) Mitochondrial DNA diversification and population structure in fishes from the Gulf of Mexico and Western Atlantic. The Journal of Heredity, 89(5), 404-414. Goodman SJ (1997) Rst CALC: A collection of computer programs for calculating unbiased estimates of genetic differentiation and determining their significance for microsatellite data. Molecular Ecology, 6, 881-885. Gordon M (1996) Meeting of the BC Eulachon Research Council minutes, November 6. Goudet J (1995) Fstat, a program for IBM PC compatible to calculate Weir and Cockerham's (1984) estimators of F-statistics. Instiut de Zoologie et d'Ecologie Animale, Batiment de Biologie, Universite de Lausanne, CH-1015 Doringy, Switzerland. Graves JE (1996) Conservation genetics of fishes in the pelagic marine realm. In: Conservation Genetics: Case histories from nature, (ed. Avise JC, Hamrick JL), pp 471-501. Chapman and Hall, New York. Graves JE (1998) Molecular insights into the population structures of cosmopolitan marine fishes. The Journal of Heredity, 89(5), 427-437. Gyllensten U (1985) The genetic structure of fish: differences in the intraspecific distribution of biochemical genetic variation between marine, anadromous, and freshwater species. Journal of Fish Biology, 26, 691-699. Hart JL, McHugh JL (1944) The smelts (Osmeridae) of British Columbia. Research Board of Canada Bulletin, 64. Hay DE (1996) Meeting of the BC Eulachon Research Council minutes, November 6. Hedrick PW (1999) Highly variable loci and their interpretation in evolution and conservation. Evolution, 53(2), 313-318. Hillier L, Green P (1991) OSP: A computer program for choosing PCR and DNA sequencing primers. PCR Methods and Applications, 1, 124-128. 112 Horrall RM (1981) Behavioral stock-isolating mechanisms in Great Lakes fishes with special reference to homing and site imprinting. Canadian Journal of Fisheries and Aquatic Sciences, 38, 1481-1496. Hume S (1999) 'Soul food' fish of native Indians is all but off menu. In: The Vancouver Sun, May 14. lies TD, Sinclair M (1982) Atlantic herring: stock discreteness and abundance. - Science, 215, 627-633. Jageilo TH, Leclait LL, Vorderstrasse BA (1996) Genetic variation and population structure of lingcod. Transactions of the American Fisheries Society, 372-385. Jarne P, Lagoda PJL (1996) Microsatellites, from molecules to populations and back. Trends in Ecology and Evolution, 11(10), 424-429. Kimura M, Crow JF (1964) The number of alleles that can be maintained in a finite population. Genetics, 49, 725-738. Kornfield I, Beland K, Moring J, Kircheis F (1981) Genetic similarity among endemic arctic char and implications for their management. Canadian Journal of Fisheries and Aquatic Sciences, 38, 32-39. La very S, Moritz S, Fielder DR (1995) Changing patterns of population structure and gene flow at different spatial scales in Birgus latro (the coconut crab). Heredity, 74,531-541. Lavery S, Moritz C, Fielder DR (1996) Genetic patterns suggest exponential population growth in a declining species. Molecular Biology and Evolution, 13, 1106-1113. Lee (1997) Oolichan fishery cancelled on Fraser. In: The Vancouver Sun, April 3. Levinson G, Gutman GA (1987) Slipped-strand mispairing: a major mechanism for DNA sequence evolution. Molecular Biology and Evolution, 4, 203-221. Lewis (1996) Meeting of the BC Eulachon Research Council minutes, November 6. Lewontin RC (1974) The genetic basis of evolutionary change. Columbia University Press, New York. Lindsey (1981) Stocks are chameleons. Canadian Journal of Fisheries and Aquatic Sciences, 38, 1497-1506. Luikart G, Allendorf FW, Cornuet JM, Sherwin WB (1998) Distortion of allele frequencies provides a test for recent population bottlenecks. Heredity, 89, 238-247. 113 Lynch M (1996) A quantitative-genetic perspective on conservation issues. In: Conservation Genetics: Case histories from nature, (ed. Avise JC, Hamrick JL), pp 471-501. Chapman and Hall, New York. Mantel N (1967) The detection of disease clustering and a generalized regression approach. Cancer Research, 27, 209-220. Marshall JT, Sage RD (1981) Taxonomy of the house mouse. Symposia of the Zoological Society of London, 47, 15-25. McAllister DE (1963) A revision of the smelt family, Osmeridae. National Museum of Canada Bulletin No. 191, Biological Series No. 71. McConnell SK, O'Reilly P, Hamilton L, Wright JM, Bentzen P (1995) Polymorphic microsatellite loci from Atlantic salmon {Salmo salar): genetic differentiation of North American and European populations. Canadian Journal of Fisheries and Aquatic Sciences, 52, 1863-1872. McElroy D, Moran P, BerminghamE, Kornfield I (1992) REAP: an integrated environment for the manipulation and phylogenetic analysis of restriction data. Journal of Heredity, 83, 157-158. McPhail JD, Lindsey CC (1970) Freshwater fishes of Northwestern Canada and Alaska. Fisheries Research Board of Canada Bulletin, 173, 1-381. McPhail JD, Lindsey CC (1986) Zoogeography of the freswaer fishes of Cascadia (the Columbia system and rivers north to the Stikine). In: The Zoogeography of North American Freshwater Fishes, (ed. Hocutt CH, Wiley EO) pp.615-637. Chichester: Wiley. Miller MP (1997) Tools for population genetic analyses (TFPGA) 1.3: A windows program for the analysis of allozyme and molecular population genetic data. Computer software distributed by author. Moritz C (1994) Defining evolutionarily significant units for conservation. Trends in Ecology and Evolution, 9, 373-375. Morris DB, Richard KR, Wright JM (1996) Microsatellites from rainbow trout and their use for genetic study of salmonids. Canadian Journal of Fisheries and Aquatic Sciences, 53, 120-126. Mourant A (1956) The distribution of the human blood groups. Charles C. Thomas, Springfield, Illinois. 114 Mulligan TJ, Chapman RW, Brown BL (1992) Mitochondrial DNA analysis of walleye pollock, Theragra chalcogramma, from the Eastern Bering Sea and Shelikof Strait, Gulf of Alaska. Canadian Journal of Fisheries and Aquatic Sciences, 49, 319-326. Nee S, Holmes EC, Rambaut A, Harvey PH (1995) Inferring population history from molecular phylogenies. Philosophical Transactions of the Royal Society of London B Biological Sciences, 349(1327), 25-31. Nei M (1987) Molecular evolutionary genetics. Columbia University Press, New York. Nei M, Li W-H (1979) Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of the National Academy of Sciences of the USA, 76, 5269-5273. Nei M, Tajima F (1981) DNA polymorphism detectable by restriction endonucleases. Genetics, 97, 145-167. Nei M, Tajima F (1983) Maximum likelihood estimation of the number of nucleotide substitutions from restriction sites data. Genetics, 105, 207-217. Newton AC, Allnutt TR, Gillies ACM, Lowe AJ, Ennos RA (1999) Molecular phylogeography, intraspecific variation and the conservation of tree species. Trends in Ecology and Evolution, 14(4), 140-145. O'Connell M, Danzmann RG, Cornuet JM, Wright JM, Ferguson MM (1997) Differentiation of rainbow trout populations in Lake Ontario and the evaluation of the stepwise mutation and infinite allele mutation models using microsatellite variability. Canadian Journal of Fisheries and Aquatic Sciences, 54, 1391-1399. O'Connell M, Dillon MC, Wright JM, Bentzen P, Merkouris S, Seeb J (1998) Genetic structuring among Alaskan Pacific herring populations identified using microsatellite variation. Journal of Fish Biology, 53, 150-163. O'Corry-Crowe GM, Suydam RS, Rosenberg A, Frost KJ, Dizon AE (1997) Phylogeography, population structure and dispersal patterns of the beluga whale Delphinapterus leucas in the western Nearctic revealed by mitochondrial DNA. Molecular Ecology, 6, 955-970. Ohta T, Kimura M (1973) A model of mutation appropriate to estimate the number of electrophorectically detectable alleles in a finite population. Genetical Research, 22, 201-204. OvendenJR (1990) Mitochondrial DNA and marine stock assessment: a review. Australian Journal of Marine and Freshwater Research, 41, 835-853. 115 Paetkau D, Strobeck C (1994) Microsatellite analysis of genetic variation in black bear populations. Molecular Ecology, 3, 489-495. Palumbi SR (1992) Marine Speciation on a small planet. Trends in Ecology and Evolution, 1, 114-118. Palumbi SR (1994) Genetic divergence, reproductive isolation, and marine speciation. Annual Review of Ecology and Systematics, 25, 547-572. Palumbi SR (1996) Nucleic acids II: The Polymerase Chain Reaction. In: Molecular Systematics (ed. Hillis DM, Moritz C, Mable BK) pp.205-248. Massachusetts, Sinauer Associates, Inc. Palumbi SR, Grabowsky G, Duda T, Geyer L, Tachino N (1997) Speciation and population genetic structure in tropical pacific sea urchins. Evolution, 51, 1506-1517. Park LK, Brainard MA, Dightman DA, Winans GA (1993) Low levels of intraspecific variation in the mitochondrial DNA of chum salmon (Oncorhynchus ketd). Molecular Marine Biology and Biotechnology, 2, 362-370. Phelps SR, LeClair LL, Young S, Blankenship, HL (1994) Genetic diversity patterns of chum salmon in the Pacific Northwest. Canadian Journal of Fisheries and Aquatic Sciences, 51 (Supplement 1), 65-83. Presa P, Gutomard R (1996) Conservation of microsatellites in three species of salmonids. Journal of Fish Biology, 49, 1326-1329. Queller DC, Strassman JE, Hughes CR (1993) Microsatellites and kinship, Trends in Ecology and Evolution, 8, 285-288. Quinn TP (1993) A review of homing and straying of wild and hatchery-produced salmon. Fisheries Research, 18, 29-44. Raymond M, Rousset F (1995) GENEPOP: a population genetic software for exact tests and ecumenicism. Journal of Heredity, 86, 248-249. Redenbach Z, Taylor EB (1999) Zoogeographical implications of variation in mitochondrial DNA of Arctic grayling (Thymallus arcticus). Molecular Ecology, 8, 23-35. Reynolds J, Weir BS, Cockerham CC (1983) Estimation of the coancestry coefficient: basis for a short-term genetic distance. Genetics, 105: 767-779. Rice WR (1989) Analyzing tables of statistical tests. Evolution, 43, 223-335. 116 Roff DA, Bentzen P (1989) The statistical analysis of mitochondrial DNA polymorphisms: chi-square and the problem of small samples. Molecular Biology and Evolution, 6, 539-545. Rogers AR, Harpending H (1992) Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution, 9, 552-569. Rogers AR (1995) Genetic evidence for a Pleistocene population explosion. Evolution, 49, 608-615. Rozen S, Skaletsky H (1998) Primer3. Code available at http://www-genome.wi.mit. edu/genome_software/other/primer3.html Ruzzante DE, Taggart CT, Cook D, Goddard SV (1996) Genetic differentiation between inshore and offshore Atlantic cod (Gadus morhua) off Newfoundland: microsatellite DNA variation and antifreeze level. Canadian Journal of Fisheries and Aquatic Sciences, 53, 534-645. Ruzzante DE, Taggart CT, Cook D, Goddard SV (1997) Genetic differentiation between inshore and offshore Atlantic cod {Gadus morhua) off Newfoundland: a test and evidence of temporal stability. Canadian Journal of Fisheries and Aquatic Sciences, 54, 2700-2708. Ryman N, Utter F, eds. (1987) Population genetics and fishery management. University of Washington Press, Seattle. 420 pp. Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution, 4, 406-425. Schlotterer C, Tautz D (1992) Slippage synthesis of simple sequence DNA. Nucleic Acids Research, 20, 211-215. Schneider S, Kueffer J-M, Roessli D, Excoffier L (1996) Arlequin: a software package for population genetics. Genetics and Biometry Lab, Dept. of Anthropology, University of Geneva. Schug MD, Mackay TFC, Aquadro CF (1999) Low mutation rates of microsatellite loci in Drosophila melanogaster. Nature Genetics, 15, 99-102. Schweigert JF, Withler RE (1990) Genetic differentiation of Pacific herring based on enzyme electrophoresis and mitochondrial DNA analysis. American Fisheries Society Symposium, 7, 459-469. Shaw PW, Pierce GJ, Boyle PR (1999) Subtle population structure within a highly vagile marine invertebrate, the veined squid Loligo forbesi, demonstrated with microsatellite DNA markers. Molecular Ecology, 8, 407-418. 117 Shepherd BG, Vroom PR (1977) Biology of the Nass River eulachon. Technical Report: PAC/T-77-10. Department of fisheries and the environment, Pacific region. Shields GF, Gust JR (1995) Lack of geographic structure in mtDNA sequences of Bering Sea walleye pollock, Theragra chalcogramma. Molecular Marine Biology and Biotechnology, 4, 69-82. Shriver MD, Jin L, Chakraborty R, Boerwinkle E (1993) VNTR allele frequency distributions under the stepwise mutation model: a computer simulation approach. Genetics, 134, 983-993. Sinclair M, lies TD (1989) Population regulation and speciation in the oceans. Journal Conseil International Exploration Mer, 45, 165-175. Slatkin M (1993) Isolation by distance in equilibrium and non-equilibrium populations. Evolution, 47, 264-279. Slatkin M (1995) A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139, 457-462. Slatkin M, Hudson RH (1991) Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics, 129, 555-562. Small MP, Beacham TD, Withler RE, Nelson RJ (1998) Discriminating coho salmon {Oncorhynchus kisutch) populations within the Fraser River, British Columbia, using microsatellite DNA markers. Molecular Ecology, 1, 141-155. Sokal R (1979) Testing statistical significance of geographic variation patterns. Systematic Zoology, 28, 227-232. Stacey D (1998) Meeting of the BC Eulachon Research Council minutes, March 12. Strand M, Prolla TA, Liskey RM, Petes TD (1993) Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair. Nature, 365, 274-276. Taggart JB, Hynes RA, Ferguson A (1992) A simplified protocol for routine total DNA isolation from salmonid fishes. Journal of Fish Biology, 40, 963-965. Tautz D (1989) Hypervariability of simple sequences as a general source for polymorphic DNA markers. Nuleic Acids Research, 17, 6463-6471. Taylor EB (1991) A review of local adaptation in Salmonidae, with particular reference to Pacific and Atlantic salmon. Aquaculture, 98, 185-207. 118 Taylor EB, Bentzen P (1993) Evidence for multiple origins and sympatric divergence of trophic ecotypes of smelt (Osmerus) in northeastern North America. Evolution, 47, 813-832. Taylor EB, Dodson JJ (1994) A molecular analysis of relationships and biogeography within a species complex of Holarctic fish (genus Osmerus). Molecular Ecology, 3, 235-248. Taylor EB, Beacham TD, Kaeriyama M (1994) Population structure and identification of north Pacific Ocean chum salmon {Oncorhynchus keta) revealed by analysis of minisatellite DNA variation. Canadian Journal of Fisheries and Aquatic Sciences, 51, 1430-1442. Taylor EB, Foote CJ, Wood CC (1996) Molecular genetic evidence for parallel life-history evolution within a Pacific salmon (sockeye salmon and kokanee, Oncorhynchus nerka). Evolution, 50, 401-416. Taylor JS, Durkin JMH, Breden F (in press) The death of a microsatellite: a phylogenetic perspective on microsatellite interruptions. Thompson CE, Taylor EB, McPhail JD (1997) Parallel evolution of stream-lake pairs of threespine sticklebacks (Gasterosteus) inferred from mitochondrial DNA variation. Evolution, 51, 1955-1965. Upholdt WB (1977) Estimation of DNA sequence divergence from comparison of restriction endonuclease digests. Nucleic Acids Research, 4, 1257-1265. Valdes AM, Slatkin M, Freimer NB (1993) Allele frequencies at microsatellite loci: the stepwise mutation model revisited. Genetics, 133, 737-749. Varnavskaya NV, Beacham TD (1992) Biochemical genetic variation in odd-year pink salmon Oncorhynchus gorbuscha from Kamchatka. Canadian Journal of Zoology, 70, 2115-2120. Waples RS (1998) Separating the wheat from the chaff: patterns of genetic differentiation in high gene flow species. Heredity, 89, 438-450. Ward RD, Woodward M, Skibinski DOF (1994) A comparison of genetic diversity levels in marine, freshwater, and anadromous fishes. Journal of Fish Biology, 44, 213-232. Weber JL, Wong C (1993) Mutation of human short tandem repeats. Human Molecular Genetics, 2, 1123-1128. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis of population structure. Evolution, 38, 1358-1370. 119 Wenburg JK, Bentzen P, Foote CJ (1998) Microsatellite analysis of genetic population structure in an endangered salmonid: the coastal cutthroat trout (Oncorhynchus clarki clarki). Molecular Ecology, 7, 733-749. Wilson AC, Cann RL, Carr SM, George M, Gyllensten JB, Helm-Bychowski KM, Higuchi RG, Palumbi SR, Prager EM, Sage RD, Stoneking M (1985) Mitochondrial DNA and two perspectives on evolutionary genetics. Biological Journal of the Linnean Society, 26, 375-400. Winans GA, Aebersold PB, Ishida Y, Urawa S (1998) Genetic stock identification of chum salmon in highseas test fisheries in the Western North Pacific Ocean and Bering Sea. North Pacific Anadromous Fish Commission Bulletin, 1, 220-226. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089038/manifest

Comment

Related Items