Recent developments in conservation genetics

Forest Ecology and Management 197 (2004) 3–19

Review

Recent developments in conservation genetics Philip W. Hedrick* Department of Biology, Arizona State University, Tempe, AZ 85287-1501, USA

Abstract Because of the new molecular genetics information from various genome projects, new applications and insights for genetics studies in endangered species are forthcoming. Neutral variants are generally used for conservation applications and estimating evolutionary parameters, and with highly variable loci, many more markers, or extensive sequence data, these approaches should become much more informative. Detrimental and adaptive variation is of importance in conservation genetics but identification and characterization of such variation is more difficult. Neutral variants might be used to identify adaptive variants but the overlay of different mutational processes and selective regimes suggests that great caution should be used in making such predictions. Examples in endangered species discussed below are (1) estimation of long-term effective population size using neutral markers in three fishes of the lower Colorado River; (2) genetic restoration of the Florida panther by the introduction of Texas cougars; (3) impact of pathogens and adaptive variation on the winter-run chinook salmon; (4) examination of the adaptive variation for a major histocompatibility complex (MHC) gene in red wolves. Although these examples are all from animals, the techniques and approaches used should prove equally useful in endangered plant species. # 2004 Elsevier B.V. All rights reserved. Keywords: Effective population size; Florida panthers; Inbreeding depression; Lethals; Neutrality; MHC; Red wolves; Winter-run chinook salmon

1. Introduction Although other forces are of primary concern for avoiding extinction of most endangered species, for long-term persistence, genetics has been a focus of conservation effort. Part of this emphasis on genetics is due to the extensive coverage in the early edited volumes on conservation biology (Soule´ and Wilcox, 1980; Soule´, 1986) and the detailed introduction of many aspects of evolutionary genetics to conservation (Frankel and Soule´, 1981; Schonewald-Cox et al., 1983). Various researchers obviously felt that the direct predictions about the history and status of endangered species appeared possible by measuring * Tel.: þ1-480-965-0799; fax: þ1-480-965-2519. E-mail address: [email protected] (P.W. Hedrick).

genetic variation and interpreting these data in a population genetics context. The framework of population genetics theory furnished an enticing and elegant approach to interpret the measured amounts of genetic variation and predict the future effects of evolutionary factors and management strategies. However, the recommendations from population genetics were general; avoid inbreeding and maintain genetic variation, with some caveats (Hedrick and Miller, 1992). In some cases, these recommendations were consistent with management ideas from other risk avoidance approaches, e.g. avoid a prolonged low population number, split the population into subpopulations, etc. However, the recommendations were generally vague and often focused on an expectation of what might occur for a typical, neutral polymorphic gene. Recently, the application of new molecular

0378-1127/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2004.05.002

4

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

techniques has made examination of genetics feasible for many more endangered species and the level of genetic analysis has become potentially much more sophisticated. For example, instead of a few polymorphic loci, many highly polymorphic loci may be available so that questions that at one time could not be resolved by molecular genetic surveys can now be definitively examined. These advances have generally applied to genetic variation that is thought to be neutral and consequently are used as markers of various evolutionary phenomena. Of great significance in conservation are variants that are not neutral, i.e. those that have either detrimental or adaptive effects. Here, I provide an overview and some examples of how these three major types of genetic variation—neutral, detrimental and adaptive—may contribute and be used in the conservation of endangered species (Hedrick, 2001b). Of course, whether a particular variant is neutral, detrimental or adaptive depends upon the environment, population size, genetic background and so on. For example, a particular allele that is adaptive, such as providing resistance to an infectious disease in one environment, could be detrimental when the pathogen is absent because of a pleiotropic cost associated with the allele. Or, genetic variants that are neutral in one situation may be adaptive in another. I also consider one of the great promises of neutral molecular variation, the use of the extent and pattern of neutral variation to predict the amount and significance of detrimental and adaptive variation. Although it has not proven compelling in some cases, with more informative loci, a larger number of genes or detailed sequence data, in the future we might be able to better utilize such observed associations.

2. Neutral variation The extent and pattern of molecular variation within a population is generally consistent with neutrality; that is, a balance predicted by a reduction in variation from genetic drift and an increase in variation from mutation (Kimura, 1983; Nei, 1987). When the population is small, even if selection is acting on the variation at a given gene, genetic drift could have a greater effect on allele frequencies than does selection. In general, neutrality of genetic variants can be

assumed when the selection coefficient s (either the selective disadvantage of a detrimental allele or the advantage of an adaptive allele) is <1/(2Ne) where Ne is the effective population size (Kimura, 1983). Therefore, because endangered species generally have low effective population sizes, genetic variants are more likely to be effectively neutral in endangered than in common species. For example, if s < 0:01 and the effective population size is 50, then genetic drift should play a more important role than selection in determining the fate of the variant, while in a population of size 500, selection should be more important than genetic drift. Most recent conservation genetics research has focused on the use of neutral molecular markers. Molecular genetic markers hold great promise for estimating fundamental parameters or characteristics important in conservation, such as past effective population size (Garrigan et al., 2002, see below), past bottlenecks (Luikart and Cornuet, 1998), population origin of individuals (Cornuet et al., 1999), individual inbreeding level (Ellegren, 1999; Lynch and Ritland, 1999) and sex-specific gene flow (Latta and Mitton, 1997) or founder contributions (Carvajal-Carmona et al., 2000). Overall, neutral genetic markers have been primarily used in conservation to identify species, evolutionarily significant units (ESUs) and management units (MUs) (e.g. Moritz, 1999). Highly variable genetic markers, such as microsatellite loci, have allowed the quantification of patterns that are not apparent when using genetic markers with less variation. The use of extensive sequence data or large numbers of single-nucleotide polymorphisms (SNPs) might also provide high genetic resolution. Molecular markers can also be used to infer the historical and geographical relationships between groups (Avise, 2000). Although the power to infer such relationships is substantial, data from ancient specimens can now provide additional insight into the relationship of contemporary groups (Barnes et al., 2002). However, application of highly variable loci (or very large numbers of markers) must be used with some caution (Hedrick, 1999) because statistically significant differences might not reflect biological important differences, or might give a different signal than do other markers (Balloux et al., 2000). For example, it is presumed that statistical significance between groups for neutral molecular markers indicates the presence of

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

biologically important differences. Or it is at least thought that it indicates that the groups have been separated long enough for biologically important difference to accumulate. Traditional molecular markers often provided inadequate statistical power to estimate the differences between groups in endangered species because there was generally little variation for these markers. However, with the use of highly variable loci, large numbers of independent markers, or extensive sequence data, the statistical power to differentiate between groups now is often very high. What are the possible associations of statistical significance based on molecular markers and biological meaningfulness of comparisons between groups? First, there may be both no significant statistical and no meaningful biological difference between groups or there may be both a significant statistical difference between groups and this reflects a meaningful biological difference. In both these cases, statistics based on molecular markers result in an appropriate evaluation of the real biological situation. However, problems result when statistical significance does not reflect biological meaningfulness, a conflict that can occur in two basic forms: there may be no statistical significance when there are actual biologically meaningful differences between groups and there may be statistical significance between groups when there is no meaningful biological difference. In the first instance, there may be no significant difference based on molecular genetic markers but other, adaptively important loci may be highly differentiated between populations. For example, in Scots pine (Pinus sylvestris) from Finland, a number of different molecular markers show very little differentiation between northern and southern populations Karhu et al. (1996). However, many important adaptive quantitative traits show high levels of genetic differentiation between these populations in common experimental environments. In this case, the molecular data appear to be adequately reflecting the high level of gene flow in Scots pine, however, the selective forces between populations are so strong that they overcome the effects of gene flow and result in large adaptive genetic differences between populations. In this case, the error is not a typical ‘‘false negative’’ because the result is correct for the neutral nuclear markers. The error results from not assaying, or being able to assay, directly the genes involved in adaptation.

5

The final classification portends to become a major concern in both evolutionary and conservation biology as large numbers of highly variable markers become available in many species. In general, statistical power for determining differentiation between groups is closely related to the number of independent alleles (Kalinowski, 2002), so that even for a few highly variable microsatellite loci, there may be extremely high statistical power. When there is such high statistical power, very small molecular genetic differences between groups become statistically significant. This is not a typical ‘‘false positive’’ because the differences detected are real but so small that they do not reflect biologically meaningful differences. To determine a biologically meaningful difference, we need to define some measure or effect related to the likelihood of the accumulation of significant biological differences. One potential way to examine the relationship between biological and statistical significance is to evaluate the statistical power to detect a known biological effect. For example, the statistical power to detect a one-generation genetic bottleneck of different sizes can be compared to the ancestral population for different numbers of loci (Hedrick, 1999). In general, we need to quantify the extent of the evolutionary effect that we are able to detect with highly variable molecular markers and evaluate whether this effect is likely to have important biological consequences. 2.1. Long-term effective population size in fishes from the lower Colorado River Conserving genetic variation has been a major focus of recovery efforts for many endangered species. Retaining variation for adaptation to environmental change is of great concern, particularly because many imperiled taxa are in recently altered habitats and exposed to new biological threats, including nonnative predators, competitors, and pathogens. In general, the amount of genetic variation within a population available for future adaptation results from a balance between mutation introducing new variation, and genetic drift, resulting from finite effective population size, reducing it. Efforts to measure the effective population size in wild populations have generally concentrated on the effective size in a given generation or over a few generations (Hedrick, 2005).

6

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

However, the amount of genetic variation potentially available for adaptation is determined by the long-term, or what could be called the evolutionary, effective population size. The long-term effective size may not be reflected in estimates of contemporary effective population size or as some percentage of the contemporary census number because the ancestral effective population size, particularly for endangered species, could be much larger, than contemporary estimates. Below, I will show how Garrigan et al. (2002) used molecular data to estimate the long-term effective population size in the endangered big fishes of the Colorado River and show that the contemporary effective size or census number are only a very small proportion of this estimated long-term effective size. Franklin (1980) presented the first effort to understand the balance of mutation and genetic drift in maintaining variation of endangered species. He suggested for neutral variants, if the effect of new mutations was about a thousandth of the environmental variance in fitness per generation, then loss of genetic variation in a finite population is balanced when the effective population size (Ne) is 500. This was the basis for his very general choice of Ne ¼ 500 for maintaining genetic variation. However, Ne is only equal to the adult breeding number if, from generation to generation, individuals at the same life stage are produced at random. For most organisms there is a higher variance in contribution than predicted from random because of unequal sex ratio, high variance in mating success, fecundity or progeny survival over individuals, and other factors. In addition, Ne over time (generations) depends on the harmonic mean for each generation, a value that may be far lower than the arithmetic mean over generations (Hedrick, 2005). Caution should be used in discussing Ne because many important parameters influencing it are not well understood and the actual Ne may be only a fraction of the total population of adults. For example, Frankham (1995), in a review of published estimates, suggested Ne is only about 10% that of adult population size. Within-generation estimates of the ratio of Ne to adult numbers often appear higher than 0.10 (Vucetich et al., 1997), but for maintaining genetic variation in the long-term, variance in Ne over time should be included (as in many estimates by Frankham, 1995; see also Kalinowski and Waples, 2002). In other words, to maintain genetic variation in a population with an

Ne of 500, even assuming that all variance generated by new mutations was potentially adaptive, would require a census population (Nc) size of 5000 adults per generation. In addition, Lande (1995) suggested as much as 90% of the increase in genetic variance by mutation over time may be caused by changes that unconditionally reduce fitness, so most new variation is unavailable for adaptive change. Based on this assumption, he thought Ne ¼ 5000 may be required to maintain potentially adaptive genetic variation. Lynch and Lande (1998) also pointed out the mutation rate for some traits, such as genes that may confer disease resistance, may be a 1000-fold lower than for quantitative traits, making numbers necessary to maintain variation for these a 1000-fold higher. The four big fishes of the Colorado River, humpback chub, bonytail chub, razorback sucker, and Colorado pikeminnow (Colorado squawfish), are all endangered and have been in deep decline in recent decades (Minckley et al., 2002). The last wild Colorado pikeminnow in the lower Colorado River was caught in 1975. Bonytail chub are critically imperiled, persisting only in Lake Mohave, Arizona–Nevada, and perhaps Lake Havasu, Arizona–California, as less than 100 wild fish augmented by hatchery reintroductions. Humpback chubs are represented by one viable population in the Little Colorado River-Grand Canyon complex. This population hovered near 10,000 adults into the early 1990s, but recently is thought to have declined substantially to around 2000 adults. Although annual spawning of razorback suckers occurs, the population consists mostly of large, very old adults and there is no evidence of recruitment success. A large population apparently formed when Lake Mohave filled in the early 1950s but annual estimates of adult fish have consistently declined to 9086 in 1999 (Minckley et al., 2002) and are thought to be around 4000 now. Indications are that historically there were large populations of all four of these endangered big fishes in the lower Colorado River as late as the mid-20th century (Minckley et al., 2002). To estimate the long-term effective population size in three of these fishes (no samples of Colorado pikeminnow were available), Garrigan et al. (2002) examined mitochondrial DNA (mtDNA) sequence data and used the maximum likelihood approach in the program FLUCTUATE (Kuhner et al., 1998).

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

Because the generation time is long for these fishes, and the decline is only over the past few generations, one would expect that estimates of the long-term effective population size from molecular data would not yet be overwhelmed by this recent decline. The maximum likelihood method assumes that new sequence variants appear by mutation and are eliminated by genetic drift. For a given mutation rate and Ne, a sample of mtDNA sequences should thus exhibit an appropriate pattern of pairwise differences. These long-term estimates are of the effective population size for a species throughout a substantial portion of its evolutionary history and do not necessarily reflect the historical or recent effective population sizes. Examination of mtDNA sequence variation in samples of bonytail, humpback, and razorback showed substantial variation (Garrigan et al., 2002). For humpback chub, bonytail chub, and razorback sucker, there were 5, 3, and 10, haplotypes found in samples of 18, 16, and 49, respectively (Table 1). Fig. 1 depicts the maximum likelihood genealogies for the three species for these samples. For example for bonytail chub, the three haplotypes, Zx, Zz, and Yy are represented by 4, 7, and 5, individuals in the sample of 16. The humpback chub and razorback genealogies show similarity in that the rare haplotypes are the most divergent and the most common haplotypes are closely related. In addition, the humpback chub and razorback sucker show similar divergence over all sequences, about 1.5 nucleotides between all pairwise comparisons, while the bonytail had an average of 2.8 nucleotide differences.

7

Let us conservatively assume that the mutation rate in mammals is 2  108 per nucleotide and that the mutation rate in fishes is about five times lower (Hedrick, 2005). Therefore, we will assume that the mutation rate per year is 4  109 and that the generation length for the chubs is 10 years and for razorback suckers, it is 15 years. Using the program FLUCTUATE, we can estimate the long-term effective population size from these data (Table 1). If the population size is assumed to be constant over evolutionary time, then estimates are 48,800, 44,800, and 223,000 for humpback, bonytail, and razorback respectively. When population growth is taken into account, the estimates suggest that bonytail chub has been declining and razorback sucker expanding in numbers over evolutionary time (Table 1). Overall, this analysis of genetic data suggests the three species existed in large numbers until only recently. Although there is less genetic variation for bonytail chub, and the estimates of effective population size are the smallest for it, the three remaining haplotypes are quite divergent suggesting that present genetic variation still reflects a high degree of the variation present ancestrally. Variation in the other two species remains even more intact. Overall, the present census number for all three species (remember the effective population size may be only a small proportion of this value) is orders of magnitude less that the estimated longterm effective population size. Many endangered species, such as these big fishes, may have evolved when their population size was much larger than presently or even known historically.

Table 1 Estimates of mtDNA variation in three endangered big fishes from the Colorado River Species Humpback chub

Bonytail chub

Razorback sucker

Data mtDNA gene Number of nucleotides Sample size Number of haplotypes

ND2 790 18 5

ND2 763 16 3

cytb 311 49 10

Estimates of Ne Ne (constant size) Ne (variable) and population trend

48,800 74,400 stable

44,800 31,000 declining

223,000 313,600 expanding

Using these data, the maximum likelihood estimates are given for the long-term effective population size, Ne, if the population is assumed constant over evolutionary time and if the population is allowed to grow or contract and the direction of that change (Garrigan et al., 2002; Hedrick, 2005).

8

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

Fig. 1. Coalescent genealogies for (A) humpback chub; (B) bonytail chub; (C) razorback sucker that maximize the likelihood of the mtDNA data (from Garrigan et al., 2002). Branch lengths are scaled in terms of the number of substitutions per site. The letters of the tree branches represent the names of the haplotype and numbers represent individuals with those sequences. Unlike a phylogenetic tree, the distance between identical sequences represents the time in generations since a common ancestor.

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

As a result they may have substantial genetic variation remaining for potential adaptation but this variation could be quickly lost if the contemporary population size is small. Further, if the current effective population size is reduced, either naturally or through inappropriate management, genetic variation will be diminished and less new variation generated, increasing potentials for reduced fitness due to fixation of detrimental alleles. Such reductions in fitness when the effective population size declines appear a particularly severe problem in species with large ancestral populations and consequently high historical genetic loads (Hedrick and Kalinowski, 2000).

3. Detrimental variation Perhaps the most important early contribution of genetics to conservation was the recognition of the importance of inbreeding depression (for a recent review, see Hedrick and Kalinowski (2000); see also Keller and Waller (2002). Inbreeding depression is generally defined as a reduction in fitness (or some component of fitness) with an increase in inbreeding, an effect thought to be due to increasing the homozygosity of detrimental alleles (Charlesworth and Charlesworth, 1999). The mean population fitness can also decline over time because detrimental mutations with a small selective disadvantage in a small population will become fixed, much as if they were neutral (Wang et al., 1999). It is useful to distinguish between these effects (Kirkpatrick and Jarne, 2000) and define the genetic load as the reduction in mean population fitness compared with a population without lowered fitness resulting from detrimental variation. Inbreeding depression and genetic load have been of major concern for endangered species and inbreeding avoidance has become a priority in captive breeding. In a large population at equilibrium, substantial standing detrimental genetic variation is expected and, consequently, a large reduction of fitness is expected if inbreeding occurs. There is also little genetic load, because, due to the efficacy of selection in large populations, most of the detrimental variants are in low frequency and are recessive. However, if the population declines in number, purging of detrimental variation should take place, especially for alleles of large detrimental effect, thereby reducing inbreeding

9

depression, but some detrimental variants might become fixed, particularly those of smaller effect, causing an increase in genetic load (Wang et al., 1999). If the population remains small for an extended period, more detrimental variation could be purged, further reducing inbreeding depression, but more detrimental variants could be fixed, causing higher genetic load. Such a population might show no lowered fitness upon inbreeding but owing to fixation of detrimentals, all individuals in the population might have a low fitness and the population might have a high genetic load. Several caveats should be mentioned. First, during this process, some populations (or even species) might become extinct and the ones going extinct could be the ones with higher genetic load. As a result, the remaining populations might not have as high a genetic load as would be expected from the standing amount of detrimental genetic variation. Second, genetic load might be documented as a low estimate of fitness compared with other populations, or by crossing with individuals from another population and observing the fitness of their progeny compared with progeny of within-population crosses. However, making such crosses might not be possible or the groups might differ in other characteristics (Wang, 2000). In Drosophila melanogaster approximately half the effect of inbreeding depression is thought to be from nearly recessive lethals and half from detrimentals of small effect but which have higher dominance (Wang et al., 1999; Lynch et al., 1999). However, D. melanogaster generally has a very large effective population size and the genetic architecture of their detrimental genetic variation probably reflects that of a large population near equilibrium. Alternatively, for many endangered species, genetic drift has been important, either because of a current small population size or a history of bottlenecks. As a result, endangered species might have a different genetic architecture with fewer segregating variants of large detrimental effect (Hedrick, 2001a), lower inbreeding depression, and perhaps higher genetic load, than do species with histories of larger population size. The magnitude and specific detrimental effects of alleles on fitness are highly variable because they might greatly depend on how these genotypes interact with the environment. Recent natural experiments are generally consistent with greater inbreeding depression in more stressful environments (Hedrick and

10

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

Kalinowski, 2000). Estimates of inbreeding depression from captivity or laboratory environments are generally thought to be an underestimate compared with that in a natural environment. For example, the effect of inbreeding on male reproductive success in wild mice under laboratory conditions was minor, whereas in semi-natural conditions, inbred males had only about 20% of the success of outbred males (Meagher et al., 2000). In a comprehensive examination in Drosophila, Bijlsma et al. (1999) found that the extent of inbreeding depression was greatly increased in stressful laboratory conditions and also found a low correlation between the fitness of genetic variants in different stressful environments. This suggests that generalizations about the detrimental effects of a variant over different stressful environments may not be possible. Inbreeding may potentially be reduced, or purged, by breeding related individuals (Byers and Waller, 1999). However, the effects observed in the Speke’s gazelle example (Templeton and Read, 1983), often cited as a demonstration of reduction of inbreeding depression are consistent with a temporal change in fitness in inbred individuals and are not necessarily the result of a reduction in inbreeding depression (Kalinowski et al., 2000). Significantly, in an examination of the potential effects of purging in 17 mammalian species (Ballou, 1997), a non-significant reduction in inbreeding depression in the updated Speke’s gazelle captive population was found and the inbreeding depression in the Speke’s gazelle was the highest of any of the species analyzed. After a thorough theoretical examination of the factors that may influence purging of inbreeding depression, Wang (2000) concluded that ‘‘it is not justified to apply a breeding program aimed at purging inbreeding depression by inbreeding and selection to a population of conservation concern’’. Little progress has been made on the genetic characterization of genes with detrimental effects segregating in endangered species. However, the effort in the human and other genome projects to identify genes causing inherited disorders should provide information about homologous genes in endangered species. In addition, the ability to map genes affecting fitnessrelated traits portends imminent knowledge of the detailed architecture of genes affecting inbreeding depression, that is, the number and location of the

genes, the distribution of their effects and their dominance, and the interaction (epistasis) of different genes. There are documented examples of inherited disorders in captive populations of endangered species (Laikre, 1999; Ralls et al., 2000), which would have important negative consequences for animals reintroduced to the wild. Although these recessive alleles are in low frequency, many individuals in the population may be heterozygous for them. Management with the aim of reducing the frequency of these alleles could be complicated and must be undertaken carefully to avoid jeopardizing the remaining genetic variation in these species (Lacy, 2000). For example, to eliminate the recessive allele for chondrodystrophy, a form of dwarfness, in the California condor (Gymnogyps californianus), would require that over half the population would be prevented from breeding (Ralls et al., 2000). On the other hand, it appears that hereditary blindness in a captive population of gray wolves (Canis lupus) could be reduced without greatly influencing the genetic variation in the rest of the genome (Laikre et al., 1993). Selection will also act to reduce the frequency of detrimental alleles, however, if they are recessive the effect will be quite slow. If the population size is small, then some detrimentals will be effectively neutral and purifying selection will not be effective. On the other hand, in small populations the expected frequency of recessive detrimentals of large effect is much lower than in an infinite population (Wright, 1937; Hedrick, 2001a). 3.1. Genetic restoration in the Florida panther Populations of some endangered species have become so small that they have lost genetic variation and appear to have deleterious genetic variants at high frequency (or fixed) (Land et al., 2001; Westemeier et al., 1998; Madsen et al., 1999; Vila´ et al., 2003). To avoid extinction from this genetic deterioration, some populations may benefit from the introduction of individuals from related populations or even subspecies for genetic restoration, i.e. elimination of deleterious variants and recovery to normal levels of genetic variation. Recent research (Ebert et al., 2002; Saccheri and Brakefield, 2002) have shown that immigration can result in genetic rescue or restoration (Richards, 2000) in experimental populations.

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

11

Table 2 The proportion of Florida panthers (sample size in parentheses) with a kinked tail, cowlick, or chryptochordism with no Texas cougar ancestry and those with Texas cougar ancestry (F1, F2, backcross to Texas, and backcross to Florida) (from Land et al., 2001; Roelke et al., 1993) Texas ancestry

Kinked tail Cowlick Cryptochordism

No Texas ancestry

F1

F2

BC-TX

BC-FL

Total

0.88 (48) 0.93 (46) 0.68 (22)

0.00 (17) 0.20 (10) 0.00 (2)

0.00 (7) 0.00 (5) 0.00 (2)

0.00 (3) 0.00 (1) –

0.20 (15) 0.60 (5) 0.00 (1)

0.07 (42) 0.24 (21) 0.00 (5)

The last remaining population of the Florida panther (Puma concolor coryi) provides an extreme example of this phenomenon in an endangered species. This population has a suite of traits that suggests genetic drift has fixed (or nearly fixed) the population for previously rare and potentially deleterious traits. These traits, which are found in high frequency only in the Florida panther and are unusual in other puma subspecies, include high frequencies of cryptochordism (unilateral undescended testicles), kinked tail for the last five vertebra, cowlick on the back, and the poorest semen quality recorded in any felid (Roelke et al., 1993, Table 2). In addition, a large survey of microsatellite loci have shown that Florida panthers have much lower molecular variation than other North American populations of pumas (Culver et al., 2000). The potential positive and negative genetic effects of introducing individuals from genetically diverse but geographically isolated populations into apparently inbred population was theoretically evaluated before the introduction of Texas cougars into Florida (Hedrick, 1995). Assuming 20% gene flow from outside in the first generation (and 2.5% every generation thereafter), the fitness generally quickly improves. One concern about this approach is that any locally adapted alleles may be swamped by outside gene flow. However, with this level of gene flow, fitness from advantageous alleles is only slightly reduced in spite of gene flow. A program to release females from the closest natural population from Texas was initiated in 1995 to genetically restore fitness in this population. Five of the eight introduced Texas females produced offspring with resident Florida panther males and a number of F1 and F2 offspring have been reproduced (Land and Lacy, 2000). At this point approximately 20% of the overall ancestry is from the introduced Texas cougars. Of the animals with Texas ancestry, only 7% have a

kinked tail (compared to 88% before) and the ones with a kinked tail are progeny from backcrosses to Florida cats (Table 2). Similarly but not as dramatic, only 24% have a cowlick (compared to 93% before), two F1s and three backcrosses to Florida cats. Only five males with Texas ancestry have been evaluated for chrytochordism and all have two descended testicles, in other words, a reduction from 68 to 0% chryptorchidism. Semen characteristics have been evaluated in only one F1 male and it appears as good or better than the average of Texas cougars, much better than Florida panthers. In other words, the introduction of Texas cougars has already resulted in a substantial reduction of the frequency of the detrimental traits that have accumulated in the Florida panther.

4. Adaptive variation The extent and pattern of adaptive (advantageous) variation is crucial to the long-term survival of endangered species. In particular, if there is no adaptive variation in a population and it faces a new environmental challenge, such as a new disease or introduced species, it has no potential for adaptive response except from new mutations or gene flow from other taxa. However, determining the extent and pattern of adaptive variation in the present or presumed future environments is quite difficult. Potentially, experimental tests of fitness and adaptiveness in a variety of environments could be carried out but this is difficult, even in a model organism, such as D. melanogaster, and virtually impossible in an endangered species. The extensive molecular data now available today potentially provides new ways to determine whether adaptive selection has operated in the past on a given gene and, therefore, whether it might operate in the future. For example, rather than being able to measure

12

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

the impact of selection in a single or a few generations by determining differential viability or reproduction, the cumulative effect over many generations, or the results of a selection episode some time in the past, could be observed in analysis of DNA variation. Recent surveys estimate the number of genes that can be classified as having experienced bouts of adaptive evolution in vertebrates to be 280 of 5305 genes examined (5.3%), while in plants 123 of 3385 (3.6%) genes show a signal of adaptive evolution (Liberles et al., 2001). Two approaches used to detect adaptive selection are comparisons of rate of non-synonymous (amino acid changing) to synonymous substitutions and the pattern of sequence variation within and between species (Richman, 2000; Yang and Bielawski, 2000; Hughes, 1999). However, Garrigan and Hedrick (2003) have found that the signal of adaptive selection observed by these tests is generally generated much more quickly than it is lost. For example, an observation that the rate of non-synonymous substitutions is greater than the synonymous rate only indicates that some time in the distant past, maybe even before the origin of the current species, that adaptive selection occurred. To determine whether adaptive selection is occurring presently, then direct measures of selective difference should be estimated (Hedrick, 2005; Garrigan and Hedrick, 2003). In general, there is no obvious molecular way to differentiate alleles that are adaptive, neutral or detrimental, except for detrimental variants that have stop codons or those for which there is detailed structural information on the molecule (Gao et al., 2001) or alleles that have quantifiable adaptive characteristics (Yokoyama et al., 1999). However, favorable alleles that have become fixed by a selective sweep are expected to show low variation at, and high linkage disequilibrium with, tightly lined markers (Hudson et al., 1997). This pattern has now proven useful in estimating the age of adaptive, polymorphic alleles at the G6PD locus in humans (Tishkoff et al., 2001) and could be used to identify alleles at other loci that have recently increased in frequency because of their adaptive importance. The foremost example of adaptive polymorphism at the DNA level is the genes of the major histocompatibility complex (MHC). These genes encode proteins with a pivotal role in immunorecognition and MHC

variation appears to be an important component of pathogen resistance (Edwards and Hedrick, 1998; Hedrick and Kim, 2000; Hill, 2001). MHC genes have been characterized at the molecular level for over two decades and nearly every approach has been taken to examine the effects of selection on these genes. Because extensive and ongoing research on MHC evolution is carried out in a wide range of model vertebrate taxa, molecular studies of many of organisms of evolutionary and conservation significance are now possible. 4.1. Pathogen resistance in the winter-run chinook salmon In recent years, it has become widely recognized that endangered species often may be threatened by exposure to pathogens (e.g. Lafferty and Gerber, 2002), many of them exotic and novel to the endangered species. Pathogens can be defined broadly to include viruses, bacteria, protozoa, etc., which can cause a reduction in fitness to the host. In endangered species, a pathogen introduced from another more common species, particularly if the endangered species has low resistance, may result in final decline in numbers to extinction. Resistance of a host to pathogens often has an important genetic component. In particular, genes in the major histocompatibility complex (Edwards and Hedrick, 1998) play an important role in disease resistance to pathogens in vertebrates (Hedrick and Kim, 2000). For example, the MHC in humans has been shown to be important in resistance to HIV, hepatitis, and malaria. Resistance to HIV and hepatitis in humans appears to be higher for MHC heterozygotes than for homozygotes (Thurz et al., 1997; Carrington et al., 1999). Further, susceptibility to HIV may be the result of a single amino acid change in a MHC molecule (Gao et al., 2001), suggesting that the role of small changes in the MHC may be important. In endangered species, the level of genetic variation in the MHC, and other genes that may influence host resistance, may be lower because of past or present small population size, than in more common species. This may make endangered species even more susceptible to the effects of pathogens than common species that have large, stable population sizes, which generally have greater genetic variation. Below I will

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

13

To examine the impact of MHC genetic variation, parents of families were chosen so that they would be segregating for MHC genotypes with the expectation that 50% of the progeny would be heterozygotes and 50% homozygotes. An easy way to visualize the influence of genetic variation at a MHC locus is to examine the difference in proportion of survival rates for the MHC heterozygotes and homozygotes. For the bacterium, there was no difference in overall survival of different MHC genotypes but there was a significant overall difference in survival of MHC heterozygotes and homozygotes upon exposure to IHNV with higher survival for heterozygotes than homozygotes in 7 of the 10 families (Fig. 2). The overall survival upon exposure to IHNV of MHC heterozygotes and homozygotes were 0.82 and 0.75, respectively. The extent of differential selection of viability for IHNV can be estimated if these values are standardized so the relative survival of heterozygotes is unity and the relative survival of the homozygotes is 1s, where s is the selection coefficient (Hedrick, 2005). In this case, the estimate of s is 0.085. These results (and the results on inbreeding and pathogen resistance not discussed here) demonstrate that loss of genetic selection can result in increased susceptibility to disease and subsequent mortality. If there are further losses of genetic variation because of continued small population size in winter-run, then one would predict both a higher level of homozygosity

summarize some of the results of an experiment to investigate both genetic resistance from an MHC gene (and from inbreeding, not discussed here) in the endangered winter-run chinook salmon, Oncorhynchus tshawytscha, to three very different pathogens (Arkush et al., 2002). To my knowledge, this is the largest and most comprehensive investigation of genetic resistance to disease in an endangered species ever carried out. The annual number of spawning adults of winter-run chinook salmon from the Sacramento River, California drainage has declined from over 100,000 in 1969 to about 200 in 1991. One response to this situation was to establish a refuge population at Bodega Marine Laboratory (BML) in case the natural run became extinct. As a result, it was possible to investigate the genetic resistance of winter-run chinook salmon produced at BML to three different pathogens; a bacteria, a virus, and a parasite (Arkush et al., 2002). The bacterium used, Listonella (Vibrio) anguillarum, is found worldwide and is a significant pathogen of salmonids. The virus used, infectious hematopoietic necrosis virus (IHNV), is considered the most important viral pathogen affecting salmonids in North America. The other pathogen, Myxobolus cerebralis, is a myxozoan parasite that causes whirling disease, which has been detected in salmonid populations throughout the US (the impact of M. cerebralis was examined in an inbred–outbred comparison and not for MHC genotypes). 0.25

Difference in proportion

0.20 0.15 0.10

* 0.05 0.00 -0.05 -0.10 -0.15 1

3

7

8

9

10

11

12

13

14

Total

Family

Fig. 2. Difference in survival between major histocompatibility complex heterozygotes and homozygotes after exposure to IHNV (infectious hematopoietic necrosis virus) for the 10 families segregating for the MHC locus where * indicates P < 0:05.

14

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

at MHC genes and higher levels of inbreeding. As a result, these known pathogens, and potentially others that may infect winter-run, may cause in a decline in the number of winter-run chinook salmon and an additional extinction threat to this critically endangered species. For species with a history of small population size, such as many endangered species, there also may be both low MHC variation and high inbreeding, suggesting that susceptibility to pathogens may result in a further reduction of fitness. In addition, endangered species may exist in stressful or marginal environments, potentially causing greater susceptibility to infectious disease. Finally, a number of pathogens appear to have host reservoirs in more common species, game species, livestock, or pets, and they may transmit pathogens to highly susceptible endangered species when they come in contact. For all these reasons, the high susceptibility of endangered species to pathogens may constitute a grave threat in coming decades. 4.2. MHC variation in red wolves The red wolf, Canis rufus, once had a distribution throughout much of the eastern part of the US (Nowak et al., 1995). However, in the early twentieth century, the numbers of red wolves declined dramatically because of eradication programs, habitat destruction, hybridization with coyotes, and parasite infestation (McCarley, 1962; Nowak, 1979). In 1967, red wolves were listed as endangered and by 1970, they remained in only a small area of Texas and Louisiana. A captive breeding program was initiated in 1974, which now contains contributions from 14 individuals from this remnant population, while the natural population became extinct in 1975. The captive population was used to start a reintroduced wild population in eastern North Carolina in 1987 (Phillips et al., 1995), which has now grown to approximately 100 individuals. Fig. 3 presents a neighbor-joining tree with the 28 alleles that we have found at a MHC gene called DRB in red wolves, gray wolves, and coyotes. As has been found for MHC genes in other taxa, the alleles for a given taxon are dispersed throughout the phylogenetic tree. In particular, the four sequences from red wolves are widely dispersed in the tree with high bootstrap numbers separating them. The average number of

Fig. 3. A neighbor-joining tree (with bootstrap values) showing the four red wolf MHC alleles, Caru-1 to Caru-4 (indicated by closed circles). Also given are 14 coyote alleles (open squares) and seven gray wolf and six Mexican wolf alleles found (open triangles). The vertical lines indicate identical sequences found in different taxa and the scale bar indicates the number of substitutions per site.

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

15

Fig. 4. The heterozygosity at individual amino-acid positions in the red wolf where * indicates sites that are putative antigen binding site positions.

amino acid differences between the four red wolf alleles is 25 nucleotides (15.5 amino acids) and the range in difference between pairs for nucleotides (19– 31) and amino acids (13–19) is not large. To determine the probability that four such divergent alleles in red wolves could persist by chance, Hedrick et al. (2002) used Monte Carlo simulation. The simulation starting point was the 28 different alleles given in Fig. 3, all with equal frequencies, and then genetic drift was allowed to reduce the allele number to four. The process was replicated 1000 times and the distribution of the number of pairwise amino acid differences between the remaining four alleles was determined. The mean expected from the simulation results was only 51% of the observed mean difference for the four red wolf alleles and none of the simulations resulted in a value as high as that observed. In other words, it appears very unlikely that the four alleles remaining in the red wolves would be as divergent as those observed by chance, suggesting that balancing selection has favored retention of divergent MHC alleles.

Using the frequencies observed for the four alleles, we can calculate the average heterozygosity for each of the amino acid positions in the sequenced gene (Fig. 4). The highest heterozygosity was for position 28 (0.714) and 18 amino acid positions have heterozygosities greater than 0.4. Most of the variation was concentrated in the functionally important antigenbinding site (ABS) positions (indicated by asterisks) with an average heterozygosity of 0.349 while the non-antigen binding sites had a significantly lower average heterozygosity of 0.043 (12% of that found for ABS positions). In addition, we found that there was an excess of heterozygotes compared to expectations, a higher rate of non-synonymous than synonymous substitution for the functionally important antigen binding site positions, and the distribution of alleles, and the distributions of amino-acids at many positions were more evenly distributed than expected from neutrality. In other words, there are five different types of evidence that support the importance of adaptive selection for this MHC gene in red wolves.

16

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

5. Neutral variation as an indicator of detrimental and adaptive variation The extent and pattern of neutral genetic variation has been used as a guide to the amount and pattern of detrimental and adaptive variation. If this association were always positive and strong, it could be a good predictor of how much detrimental variation might be present that could lower fitness and how much adaptive variation might be present to deal with future challenges. With more loci and more variable loci, estimates of the extent and pattern of neutral genetic variation will presumably become more accurate. However, different evolutionary scenarios might be responsible for different amounts of neutral variation (Table 3) and these scenarios might in turn result in different amounts or patterns of detrimental (Kirkpatrick and Jarne, 2000) or adaptive variation. Positive associations might occur when the population size has been large or small for a long time or where stochastic effects dominate the extent of genetic variation. For example, there is a high correlation of (neutral) microsatellite and (adaptive) MHC loci variation over Gila topminnows (Poeciliposis occidentalis) (Hedrick et al.,

2001) and desert bighorn sheep (Ovis canadensis) (Gutierrez-Espeleta et al., 2001a,b) populations, two species in which stochastic factors appear to be important in the present spatial pattern of variation. However, there might be situations in which such an inference based on a positive association is unfounded. For example, some time after a bottleneck there might be a negative association between neutral and detrimental variation because the amount of detrimental variation could recover faster as a result of a higher mutation rate. Although this might not be the case for all neutral variation, because some genes, such as microsatellite loci, have mutation rates as high as those for detrimental variation. A negative association may also occur when separately bottlenecked populations become mixed, resulting in high neutral variation but low detrimental variation. Finally, in a metapopulation, there might be a positive association between neutral and detrimental variation, but, depending upon the level of extinction, recolonization, and gene flow, the levels of variation might be very low to high. In some specific cases, it appears that there might be little genetic variation for some markers and substan-

Table 3 The amount of observed neutral variation and general predictions of the amounts of detrimental and adaptive variation under some different evolutionary scenarios (after Hedrick, 2001b) Scenario

(1) Equilibrium (a)Large population (b) Small population (2) Bottleneck (a) Shortly after (b) Some time after

(c) Mixture of separately bottlenecked populations (3) Metapopulation (a) With extinction, recolonization, and low gene flow (b) No extinction and more gene flow

Observed

Predicted

Neutral variation

Detrimental variation

Adaptive variation

High Low

High Low

High Low to medium (retention more than neutrality)

Low (loss of alleles more than heterozygosity) Low to high (depending upon mutation rate of variants)

Low (loss of lethals more than detrimentals) High (assuming high mutation rate)

High (due to fixation of different alleles in different populations)

Low (due to purging of variants within each population)

Low to medium (retention more than neutrality) Low to medium (retention more than neutrality, function of mutation rate) Low to medium (retention more than neutrality)

Low (due to low effective population size)

Low (due to purging)

Low to medium (depending upon selection model)

High (as if one large population)

High (as if one large population)

High (as if one large population)

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

tial variation for others. For example, the cheetah (Acinonyx jubatus) was documented to have low genetic variation for allozymes and other genetic markers (O’Brien et al., 1985) and great concern was raised about its long-term survival. However, further studies showed that cheetahs have substantial genetic variation for microsatellites (Menotti-Raymond and O’Brien, 1995) and captive cheetahs appear to exhibit inbreeding depression for juvenile survival (Wielebnowski, 1996), indicating variation for detrimental variants. These differences were explained with scenarios that include a bottleneck or a metapopulation with a small effective population size (Hedrick, 1996) and different mutation rates for different markers. Overall, the positive correlation between neutral and adaptive variation might not be very high. However, high neutral variation may indicate the potential for significant adaptive variation. Low neutral variation could indicate low adaptive variation but the present population might either be well adapted or poorly adapted to its environment. Furthermore, gene flow between populations could result in low differentiation for neutral markers between populations, but there might still be strong adaptive differences in the populations. The extent and pattern of neutral variation is the result of non-selective forces and can potentially be used to identify the past importance of finite size, bottlenecks and population structure. However, adaptive variants might differ in mutation rate from neutral markers and the selection might result in quite different patterns of variation within and between populations. In other words, neutral variants might be used as a guide to understanding non-selective effects, but the overlay of different mutational processes and selective regimes suggests that great caution should be used in making such predictions.

6. Overview Genetics will probably play an even more important role in conservation in the future. The utilization of information from the human and other genome projects will provide substantially more background understanding and the appropriate use of these data should be of great benefit to conservation. In parti-

17

cular, techniques to screen and analyze large amounts of data, whether it is variation at marker loci or DNA sequence data will be used to determine specific groups and individuals. However, above the high statistical power from these data will require an evolutionary perspective to evaluate the biological importance of these differences. In addition, these data will allow an understanding past evolutionary events in endangered species, whether they are non-selective, such as bottlenecks or gene flow, or selective such as detrimental mutants or adaptive variants.

References Arkush, K.D., Giese, A.R., Mendonca, H.L., McBride, A.M., Marty, G.D., Hedrick, P.W., 2002. Resistance to three pathogens in the endangered winter-run chinook salmon: effects of inbreeding and MHC genotypes. Can. J. Fish. Aquat. Sci. 59, 966–975. Avise, J., 2000. Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, MA, USA. Ballou, J.D., 1997. Ancestral inbreeding only minimally affects inbreeding depression in mammalian populations. J. Hered. 88, 169–178. Balloux, F., Brunner, H., Lugon-Moulin, N., Hausser, J., Goudet, J., 2000. Microsatellites can be misleading: an empirical and simulation study. Evolution 54, 1414–1422. Barnes, I., Matheus, P., Shapiro, B., Jensen, D., Cooper, A., 2002. Dynamics of Pleistocene population extinctions in Beringian brown bears. Science 295, 2267–2270. Bijlsma, R., Bundgaard, J., Van Putten, W.F., 1999. Environmental dependence of inbreeding depression and purging in Drosophila melanogaster. J. Evol. Biol. 12, 1125–1137. Byers, D.L., Waller, D.M., 1999. Do plant populations purge their genetic load? Effects of population size and mating history on inbreeding depression. Ann. Rev. Ecol. Syst. 30, 479–513. Carrington, M., Nelson, G.W., Martin, M.P., Kissner, T., Vlahov, D., Goedert, J.J., Kaslow, R., Buchbinder, S., Hoots, K., O’Brien, S.J., 1999. HLA and HIV-1: heterozygote advantage and B35-Cw04 disadvantage. Science 238, 1748–1752. Carvajal-Carmona, L.G., Soto, I.D., Pineda, N., Ortiz-Barrientos, D., Duque, C., Ospina-Duque, J., McCarthy, M., Montoya, P., Alvarez, V.M., Bedoya, G., Ruiz-Linares, A., 2000. Strong Amerind/white sex bias and a possible Sephardic contribution among the founders of a population in northwest Colombia. Am. J. Human Genet. 67, 1287–1295. Charlesworth, D., Charlesworth, B., 1999. The genetic basis of inbreeding depression. Genet. Res. 74, 329–340. Cornuet, J.-M., Piry, S., Luikart, G., Estoup, A., Sloignac, M., 1999. New methods employing multilocus genotypes to select or exclude populations as origins of individuals. Genetics 153, 1989–2000.

18

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19

Culver, M., Johnson, W.E., Pecon-Slattery, J., O’Brien, S.J., 2000. Genomic ancestry of the American puma (Puma conclor). J. Hered. 91, 186–197. Ebert, D., Haag, C., Kirkpatrick, M., Riek, M., Hottinger, J.W., Pajunen, V.I., 2002. A selective advantage to immigrant genes in a Daphnia metapopulation. Science 295, 485–488. Edwards, S.V., Hedrick, P.W., 1998. Evolution and ecology of MHC molecules: from genomics to sexual selection. Trends Ecol. Evol. 13, 305–311. Ellegren, H., 1999. Inbreeding and relatedness in Scandinavian grey wolves Canis lupus. Hereditas 103, 239–244. Frankel, O.H., Soule´ , M.E., 1981. Conservation and Evolution. Cambridge University Press, Cambridge, UK. Frankham, R., 1995. Effective population size/adult population size ratio in wildlife: a review. Genet. Res. 66, 95–107. Franklin, I.R., 1980. Evolutionary change in small populations. In: Soule´ , M.E., Wilcox, B.A. (Eds.), Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer Associates, Sunderland, Massachusetts, pp. 135–149. Gao, X., Nelson, G.W., Karacki, P., Martin, M.P., Phair, J., Kaslow, R., Goedert, J.J., Buchbinder, S., Hoots, K., Vlahov, C., O’Brien, S.J., Carrington, M., 2001. Effect of a single amino acid substitution in the MHC class I molecule on the rate of progression to AIDS. New Engl. J. Med. 344, 1668– 1675. Garrigan, D., Marsh, P.C., Dowling, T.E., 2002. Long-term effective population size of three endangered Colorado River fishes. Anim. Conserv. 5, 95–102. Garrigan, D., Hedrick, P.W., 2003. Perspective: detecting adaptive molecular evolution, lessons for the MHC. Evolution, submitted for publication. Gutierrez-Espeleta, G.A., Kalinowski, S.T., Boyce, W.M., Hedrick, P.W., 2001a. Genetic variation and population structure in desert bighorn sheep: implications for conservation. Conserv. Genet. 1, 3–15. Gutierrez-Espeleta, G.A., Hedrick, P.W., Kalinowski, S.T., Garrigan, D., Boyce, W.M., 2001b. Is the decline of desert bighorn sheep from infectious disease the result of low MHC variation? Heredity 86, 439–450. Hedrick, P.W., 1995. Gene flow and genetic restoration: the Florida panther as a case study. Conserv. Biol. 9, 996–1007. Hedrick, P.W., 1996. Bottleneckor(s) metapopulation in cheetahs. Conserv. Biol. 10, 897–899. Hedrick, P.W., 1999. Perspective: highly variable loci and their interpretation in evolution and conservation. Evolution 53, 313–318. Hedrick, P.W., 2005. Genetics of Populations, 3rd ed. Jones and Bartlett, Boston. Hedrick, P.W., 2001a. Lethals in finite populations. Evolution 56, 654–657. Hedrick, P.W., 2001b. Conservation genetics: where are we now? Trends Ecol. Evol. 16, 629–636. Hedrick, P.W., Kalinowski, S.T., 2000. Inbreeding depression in conservation biology. Ann. Rev. Ecol. Syst. 31, 139–216. Hedrick, P.W., Kim, T.J., 2000. Genetics of complex polymorphisms: parasites and maintenance of the major histocompatibility complex variation. In: Singh, R.S., Krimbas, C.B. (Eds.),

Evolutionary Genetics: From Molecules to Morphology. Cambridge University Press, pp. 204–234. Hedrick, P.W., Lee, R.N., Garrigan, D., 2002. MHC variation in red wolves: evidence for common ancestry from coyotes and balancing selection. Mol. Ecol. 11, 1905–1913. Hedrick, P.W., Miller, P.S., 1992. Conservation genetics: techniques and fundamentals. Ecol. Appl. 2, 30–46. Hedrick, P.W., Parker, K.M., Lee, R., 2001. Genetic variation in the endangered Gila and Yaqui topminnows: microsatellite and MHC variation. Mol. Ecol. 10, 1399–1412. Hill, A.V.S., 2001. The genomics and genetics of human infectious disease susceptibility. Ann. Rev. Genom. Human Genet. 2, 373–400. Hudson, R.R., Saez, A.G., Ayala, F.J., 1997. DNA variation at the Sod locus of Drosophila melanogaster: an unfolding story of natural selection. Proc. Natl. Acad. Sci. U.S.A. 94, 7725–7729. Hughes, A.L., 1999. Adaptive Evolution of Genes and Genomes. Oxford University Press, Oxford, UK. Kalinowski, S.T., 2002. How many alleles should be used to estimate genetic distances? Heredity 88, 62–65. Kalinowski, S.T., Hedrick, P.W., Miller, P.S., 2000. A close look at inbreeding depression in the Speke’s gazelle captive breeding program. Conserv. Biol. 14, 1375–1384. Kalinowski, S.T., Waples, R.S., 2002. Relationship of effective to census size in fluctuating populations. Conserv. Biol. 16, 129–136. Karhu, A., Hurme, P., Karjalainen, M., Karvonen, P., Karkkainen, K., Neale, D., Savolainen, O., 1996. Do molecular markers reflect patterns of differentiation in adaptive traits of conifers? Theoret. Appl. Genet. 93, 215–221. Keller, L.F., Waller, D.M., 2002. Inbreeding effects in wild populations. Trends Ecol. Evol. 17, 230–241. Kimura, M., 1983. The Neutral Theory of Molecular Evolution. Cambridge University Press, Cambridge, UK. Kirkpatrick, M., Jarne, P., 2000. The effects of a bottleneck on inbreeding depression and the genetic load. Am. Nat. 155, 154–167. Kuhner, M.K., Felsenstein, J., Yamato, J., 1998. Maximum likelihood estimation of population growth rates based on the coalescent. Genetics 149, 429–434. Lacy, R.C., 2000. Should we select genetic alleles in our conservation breeding programs? Zoo Biol. 19, 279–282. Lafferty, K.D., Gerber, L., 2002. Good medicine for conservation biology: the intersection of epidemiology and conservation theory. Conserv. Biol. 16, 593–604. Laikre, L., 1999. Hereditary defects and conservation genetic management of captive populations. Zoo Biol. 18, 91–99. Laikre, L., Ryman, N., Thompson, E.S., 1993. Hereditary blindness in a captive wolf (Canis lupus) population: frequency reduction of a deleterious alleles in relation to gene conservation. Conserv. Biol. 7, 592–601. Land, D., Cunningham, M., Lotz, M., Shindle, D., 2001. Florida panther genetic restoration and management. Annual report 2000–2001. Florida Fish and Wildlife Conservation Commission. Land, E.D., Lacy, R.C., 2000. Introgression level achieved through Florida panther genetic restoration. Endangered Species Update 17, 100–105. Lande, R., 1995. Mutation and conservation. Conserv. Biol. 9, 782– 791.

P.W. Hedrick / Forest Ecology and Management 197 (2004) 3–19 Latta, R.G., Mitton, J.B., 1997. A comparison of population differentiation across four classes of gene markers in limber pine (Pinus flexilis James). Genetics 146, 1153–1163. Liberles, D.A., Schreiber, D.R., Govindarajan, S., Chamberlin, S.G., Benner, S.A., 2001. The adaptive evolution database (TAED). Genome Biol. 2, 20–28. Luikart, G., Cornuet, J.-M., 1998. Empirical evaluation of a test for identifying recently bottlenecked populations from allele frequency data. Conserv. Biol. 12, 228–237. Lynch, M., Blanchard, J., Houle, D., Kibota, T., Schultz, S., Vassilieva, L., Willis, J., 1999. Perspective: spontaneous deleterious mutation. Evolution 53, 645–663. Lynch, M., Lande, R., 1998. The critical effective size for a genetically secure population. Anim. Conserv. 1, 70–72. Lynch, M., Ritland, K., 1999. Estimation of pairwise relatedness with molecular markers. Genetics 152, 1753–1766. Madsen, T., Shine, R., Olsson, M., Wittsell, H., 1999. Restoration of an inbred adder population. Nature 402, 34–35. McCarley, H., 1962. The taxonomic status of wild Canis (Canidae) in south central United States. Southwest. Nat. 7, 227–235. Meagher, S., Penn, D.J., Potts, W.K., 2000. Male–male competition magnifies inbreeding depression in wild house mice. Proc. Natl. Acad. Sci. U.S.A. 97, 3324–3329. Menotti-Raymond, M., O’Brien, S.J., 1995. Evolutionary conservation of ten microsatellite loci in four species of Felidae. J. Hered. 86, 319–322. Moritz, C., 1999. Conservation units and translocation: strategies for conserving evolutionary processes. Hereditas 130, 217–228. Nei, M., 1987. Molecular Evolutionary Genetics. Columbia University Press, New York. Nowak, R.M., 1979. North American Quaternary Canis. University Kansas Museum Natural History Monograph No. 6. Nowak, R.M., Phillips, M.K., Henry, V.G., Hunter, W.C., Smith, R., 1995. The origin and fate of the red wolf. In: Carybn, L.N., Fritts, S.H., Seip, D.R. (Eds.), Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, Occasional Publication No. 35, pp. 409–415. O’Brien, S.J., Roelke, M.E., Marker, L., Newman, A., Winkler, C.A., Meltzer, D., Colly, L., Evermann, J.F., Bush, M., Wildt, D.E., 1985. Genetic basis for species vulnerability in the cheetah. Science 227, 1428–1434. Phillips, M.K., Smith, R., Henry, V.G., Lucash, C., 1995. Red wolf reintroduction program. In: Carybn, L.N., Fritts, S.H., Seip, D.R. (Eds.), Ecology and Conservation of Wolves in a Changing World. Canadian Circumpolar Institute, Occasional Publication No. 35, pp. 157–168. Ralls, K., Ballou, J.D., Rideout, B.A., Frankham, R., 2000. Genetic management of chondrodystrophy in California condors. Anim. Conserv. 3, 145–153. Richards, C.M., 2000. Inbreeding depression and genetic rescue in a plant metapopulation. Am. Nat. 155, 383–394. Richman, A., 2000. Evolution of balanced genetic polymorphism. Mol. Ecol. 9, 1953–1963. Roelke, M.E., Martenson, J.S., O’Brien, S.J., 1993. The consequences of demographic reduction and genetic depletion in the endangered Florida panther. Curr. Biol. 3, 340–350.

19

Saccheri, I.J., Brakefield, P.M., 2002. Rapid spread of immigrant genomes into inbred populations. Proc. Roy. Soc. London B 269, 1073–1078. Schonewald-Cox, C.M., Chambers, S.M., MacBryde, B., Thomas, W.L. (Eds.), 1983. Genetics and Conservation. Benjamin/ Cummings, Menlo Park, California. Soule´ , M.E. (Ed.), 1986. Conservation Biology: The Science of Scarcity and Diversity. Sinauer Associates, Sunderland, Massachusetts. Soule´ , M.E., Wilcox, B.A. (Eds.), 1980. Conservation Biology: An Ecological-Evolutionary Perspective. Sinauer Associates, Sunderland, Massachusetts. Templeton, A.R., Read, B., 1983. The elimination of inbreeding depression in a captive herd of Speke’s gazelle. In: Schonewald-Cox, C.M., Chambers, S.M., MacBryde, B., Thomas, L. (Eds.), Genetics and Conservation. Benjamin/Cummings, Menlo Park, CA, pp. 241–261. Thurz, M.R., Thomas, H.C., Greenwood, B.M., Hill, A.V.S., 1997. Heterozygote advantage for HLA class-II type in hepatitis B virus infection. Nat. Genet. 17, 11–12. Tishkoff, S.A., Varkonyi, R., Cahinhinan, N., Abbes, S., Argyropoulos, G., Destro-Bisol, G., Drousiotou, A., Dangerfield, B., Lefranc, G., Loiselet, J., Piro, A., Stoneking, M., Tagarelli, A., Tagarelli, G., Touma, E.H., Williams, S.M., Clark, A.G., 2001. Haplotype diversity and linkage disequilibrium at human G6PD: recent origin of alleles that confer malarial resistance. Science 293, 455–462. Vila´ , C., Sundqvist, A.K., Flagstad, O., Seddon, J.M., Bjornerfeldt, S., Kojola, I., Sand, H., Wabakken, P., Ellegren, H., 2003. Rescue of a severely bottlenecked wolf (Canis lupus) population by a single immigrant. Proc. Roy. Soc. London B 270, 91–97. Vucetich, J.A., Waite, T.A., Nunney, L., 1997. Fluctuation in population size and the ratio of effective to census population size. Evolution 51, 2017–2021. Wang, J., Hill, W.G., Charlesworth, D., Charlesworth, B., 1999. Dynamics of inbreeding depression due to deleterious mutations in small populations: mutation parameters and inbreeding rate. Genet. Res. 74, 165–178. Wang, J., 2000. Effects of population structure and selection strategies on the purging of inbreeding depression due to deleterious mutations. Genet. Res. 76, 75–86. Westemeier, R.L., Brown, J.D., Simpson, S.A., Esker, T.L., Jansen, R.W., Walk, J.W., Kershner, E.L., Bouzat, J.L., Paige, K.N., 1998. Tracking the long-term decline and recovery of an isolated population. Science 282, 1695–1698. Wielebnowski, N., 1996. Reassessing the relationship between juvenile mortality and genetic monomorphism in captive cheetahs. Zoo Biol. 15, 353–369. Wright, S., 1937. The distributions of gene frequencies in populations. Proc. Natl. Acad. Sci. U.S.A. 23, 307–320. Yang, Z., Bielawski, J.P., 2000. Statistical methods for detecting molecular adaptation. Trends Ecol. Evol. 15, 496–503. Yokoyama, S., Zhang, H., Radlwimmer, F.B., Blow, N.S., 1999. Adaptive evolution of color vision of the Comoran coelacanth (Latimeria chalumnae). Proc. Natl. Acad. Sci. U.S.A. 96, 6279– 6284.