Effectiveness of selection in reducing the genetic load in populations of Peromyscus polionotus during generations of inbreeding.
Most of the data on reduction of the genetic load are from studies of self-compatible plants. Multigenerational studies of various species and traits have shown strong purging (McCall et al. 1994), weak purging (Barrett and Charlesworth 1991; Latta and Ritland 1994; Dudash et al. 1997), and no purging (McCall et al. 1994; Hauser and Loeschcke 1996). Among species, the trend has been for inbreeding depression to be weaker in species with higher rates of selfing (Husband and Schemske 1996), but exceptions are known (Barrett and Kohn 1991; Mayer et al. 1996).
Some animal species that regularly inbreed and that show little inbreeding depression are known (hermaphroditic snails: Doums et al. 1996). Among normally outcrossing animal species, there is some evidence for reductions in the genetic load during or following generations of inbreeding. Bowman and Falconer (1960) observed recovery of initially depressed litter size in inbred lines of Mus. Latter et al. (1995) obtained a decrease in the frequency of lethal second chromosomes and an increase in homozygote fitness after 210 generations of slow (1% per generation) inbreeding of D. melanogaster. Templeton and Read (1983, 1984) described a breeding program for the Speke's gazelle, in which inbred but healthy animals were selected as breeders to reduce the genetic load in the small captive population. However, their analysis suffered from several possible biases (Lacy 1997; Willis and Wiese 1997), and recent reanalyses of the studbook did not find that inbreeding depression had been reduced (Ballou 1997; Willis and Wiese 1997). Various other studies examining effects of multiple generations of inbreeding have not observed reductions in the effects of inbreeding (White 1972; Lynch 1977; Sharp 1984; Brewer et al. 1990; Lacy and Horner 1997), but often the tests were not rigorous.
The effectiveness of selection in reducing inbreeding depression depends critically on the nature of the genes causing inbreeding depression. Alternative hypotheses postulate that inbreeding depression is caused by the expression of deleterious recessive alleles (the dominance hypothesis) or by reduced heterozygosity at loci with heterozygote advantage (the overdominance hypothesis) (Crow 1948). Selection can remove recessive alleles, with substantial additive effects on fitness, but it cannot directly prevent the reduction in fitness as overdominant loci become homozygous (Charlesworth and Charlesworth 1987). Other factors also influence the effectiveness of selection against inbreeding depression, including the number of loci contributing to the genetic load (Ehiobu et al 1989; Charlesworth et al. 1992; Hedrick 1994), the population size (Charlesworth et al. 1992), the rate of inbreeding (Ehiobu et al. 1989), epistasis (Charlesworth et al. 1991), linkage disequilibrium (Charlesworth et al. 1992), and the mutation rate (Lande et al. 1994).
Multiple Components of Inbreeding Depression
In the simplest case, inbreeding depression might be reflected in a linear decline in fitness as a population becomes increasingly inbred. Exponential decline (Morton et al. 1956), a logistic relationship (Lacy et al. 1996), or other models (Ballou 1997) might be more appropriate for some fitness traits or some genetic mechanisms, but the differences between such models are often small except when inbreeding depression is severe. Although inbreeding depression is commonly attributed to the effects of homozygosity in the inbred offspring, in taxa with parental care the fitness of progeny might be determined also or instead by the genotypes of the parents. Inbreeding depression therefore, might be manifest either directly or via a maternal or paternal effect (Brewer et al. 1990; Lacy et al. 1996; Margulis and Altmann 1997; Margulis 1997, 1998a, 1998b). Inbreeding depression at these two levels can be partitioned in linear models of the form
w = [w.sub.0] + [Beta]f + [[Beta].sub.d][f.sub.d], (1)
in which w is the fitness of the progeny, [w.sub.0] is the mean fitness in a noninbred population, f and [f.sub.d] are the inbreeding coefficients of the offspring and dam, respectively, and [Beta] [less than] 0 and [[Beta].sub.d] [less than] 0 indicate inbreeding depression caused by inbred offspring and via a maternal effect.
Purging of the genetic load during generations of inbreeding would result in an effect of the interaction between the inbreeding coefficient (f) and the amount of inbreeding in the ancestors of an individual. Ballou (1995, 1997) defined an "ancestral inbreeding coefficient," denoted [f.sub.a], as the cumulative proportion of an individual's genome that has previously been autozygous due to inbreeding in its ancestors. He gave the iterative formula for calculating [f.sub.a] from the inbreeding and the ancestral inbreeding of the parents
[f.sub.a] = [[f.sub.a(s)] + (1 - [f.sub.a(s)])[f.sub.s] + [f.sub.a(d)] + (1 - [f.sub.a(d)])[f.sub.d]]/2, (2)
in which the subscripts s and d indicate values for the sire and dam. If the deleterious alleles that contribute to the inbreeding depression are effectively removed by selection when exposed in homozygous ancestors, then the effects of inbreeding in the current generation would decline as [f.sub.a] increases. This interaction is most simply modeled as a linear effect on fitness
w = [w.sub.0] + [[Beta].sub.f]f + [[Beta].sub.d][f.sub.d] + [[Beta].sub.a]f[f.sub.a]
= [w.sub.0] + ([[Beta].sub.f] + [[Beta].sub.a][f.sub.a])f + [[Beta].sub.d][f.sub.d], (3)
in which [[Beta].sub.f] is the effect of current or incremental inbreeding and [[Beta].sub.a] is the effect of ancestral inbreeding. The effect of the f [f.sub.a] interaction on w may be more complex than the linear effect in eq. (3), and its precise form would be dependent on both the structure of the pedigree and the nature of selection on the genetic load.
The second form of eq. (3) shows that overall effect of cumulative inbreeding ([Beta] in Eq. 1) can be partitioned into an effect of incremental inbreeding and an effect of ancestral inbreeding ([Beta] = [[Beta].sub.f] + [[Beta].sub.a][f.sub.a]). If there is inbreeding depression that is purged (at least partly) during generations of inbreeding, then [[Beta].sub.f] [less than] 0 and [[Beta].sub.a] [greater than] 0. The combined effect, [[Beta].sub.f] + [[Beta].sub.a], can be considered a measure of the genetic load that would remain after selection has removed any component that would respond to selection during generations of inbreeding.
Likewise, any effect of maternal inbreeding ([[Beta].sub.d]) might also be purged by selection. This would be detected as an interaction effect between [f.sub.d] and [f.sub.a(d)], as in the following model
w = [w.sub.0] + [[Beta].sub.f]f + [[Beta].sub.d][f.sub.d] + [[Beta].sub.a(d)][f.sub.d][f.sub.a(d)]
= [w.sub.0] + [[Beta].sub.f]f + ([[Beta].sub.d] + [[Beta].sub.a(d)][f.sub.a(d)][f.sub.a(d)])[f.sub.d]. (4)
In theory, when there are both direct effects of inbreeding in the offspring and indirect maternal effects, then purging of inbreeding depression could occur at either or both levels. However, the ancestral inbreeding of the offspring ([f.sub.a]) contains maternal inbreeding ([f.sub.d]) and the ancestral inbreeding of the dam ([f.sub.a(d)]) (see eq. 2). This prevents robust statistical partitioning of the effects of ancestral inbreeding into its direct and indirect (maternal) components.
Here we use the above models to evaluate the effect of selection on reducing the genetic load on aspects of fitness from a large, multigenerational study of inbreeding in three subspecies of Peromyscus polionotus mice. The effects of inbreeding in these experimental populations were reported in Lacy et al. (1996). Significant variation in inbreeding depression was reported among replicate stocks of each taxon, but the variation among replicates could be wholly attributed to random variation among founder lineages. The existence of these large founder effects suggested that inbreeding depression is caused by relatively few genes in these populations and that selection within and between lineages might be effective at removing the deleterious alleles. Here we use these data to test for changes in inbreeding depression in each subspecies as a response to selection during 10 generations of inbreeding.
MATERIALS AND METHODS
Subjects comprised laboratory stocks of three subspecies of the beach mouse or old-field mouse, Peromyscus poliohorus. Peromyscus p. subgriseus were collected from the Ocala National Forest in Marion County, Florida. Peromyscus p. rhoadsi were collected near Lake Placid, Highlands County, Florida. Peromyscus p. leucocephalus were collected from a section of Eglin Air Force Base on Santa Rosa Island, Florida. The laboratory stocks of subgriseus, rhoadsi, and leucocephalus descended from 52, 52, and 50 wild-caught founders, respectively. The founders were randomly assigned to three replicate laboratory populations for each subspecies. The replicate stocks were kept separate through 10 generations of laboratory breeding.
In each generation, some mice of low inbreeding coefficients (f [less than] 0.10) were paired to produce progeny with similarly low inbreeding; some noninbred but related mice were paired to produce inbred litters; some inbred but unrelated mice were paired to produce noninbred litters; and some inbred and related mice were paired to produce more highly inbred progeny. Inbred pairings often consisted of full-sib or half-sib matings, but littermates were never paired for breeding. The breeding design was not regular. As a result, the highest inbreeding coefficients achieved (f = 0.594) were lower than would have been produced from line-breeding of first-order relatives. Details on stock maintenance and breeding protocols are provided in Lacy et al. (1996).
Six components of reproductive performance were recorded: the proportion of pairs producing litters conceived within the up to 63 days that each pair was kept together (P[litter]); the proportion of breeding pairs (those that produced at least one litter and were housed together through the rearing of the first litter) that rebred during the postpartum estrous and produced a second litter (P[2nd litter]); the number of pups born in each litter (litter size); the survival of pups from birth to weaning (viability); the mass of pups at weaning (mass); and the within-litter variance in mass at weaning (for those litters with two or more offspring weaned), expressed as a standard deviation (SD[mass]). The fates of pups within litters are not independent. Therefore, each litter rather than each offspring was considered as an independent data point for analysis. Viability was analyzed as a categorical trait, and was assigned a value of one if more than half of the progeny of a litter survived and zero otherwise. Mass at weaning was averaged across the surviving progeny per litter. To evaluate the combined effects of fertility, fecundity, infant survival, and growth, overall reproductive success (RS) was assessed as the total mass of offspring weaned for the zero, one, or two litters produced by each pair.
The probability of pairs producing at least one litter (P[litter]), the probability of those breeding pairs producing a second litter (P[2nd litter]), and viability were scored as 0/1 categorical responses and are presumed to result from binomial processes. (When viability is assessed as the proportion of pups surviving, the distribution is strongly bimodal. In the majority of litters either all pups survived or all died.) Responses of these variables to inbreeding were examined with logistic multiple regression models. The regression coefficients were fitted by maximum-likelihood estimation, with significance determined by likelihood-ratio tests (Hosmer and Lemeshow 1989).
The distributions of litter size, mass, and SD[mass] were approximately normal. The distribution of the total mass of offspring weaned per pair (RS) deviated considerably from normality. Analyses of this composite measure of fitness are presented in the results, but the tests of statistical significance of effects on RS should not be considered accurate, and they would not be independent of the tests of effects on the fitness components. The responses of the continuous variables to inbreeding were assessed by least-squares linear regressions.
For each fitness measure, we examined the effects of inbreeding, maternal inbreeding, and the interaction of inbreeding depression at one or the other of these levels with ancestral inbreeding. When the direct inbreeding depression was greater than the indirect maternal effect (-[[Beta].sub.f] [greater than] -[[Beta].sub.d]) for a trait, a regression model of the form of equation (3) was used to examine the effects of ancestral inbreeding, while recognizing that the maternal effect ([[Beta].sub.d]) would encompass the deleterious effects of maternal inbreeding as moderated by any purging of the genetic load in the dam. When the effect of maternal inbreeding was greater than the effect of inbreeding of the offspring (- [[Beta].sub.d] [greater than] - [[Beta].sub.f]), a model of the form of equation (4) was used to examine purging of the maternal inbreeding depression.
Primiparous female mice typically produce smaller litters with slower growth and lower viability (Brewer et al. 1990; Lacy et al. 1996). Therefore, the effect of parity was included in all analyses of litter size, viability, mass, and SD[mass], by adding a term ([[Beta].sub.p]p, in which p is either zero or one for first vs. second litters produced by each breeding pair) to the regression models.
Statistical analyses were conducted with SYSTAT (Wilkinson 1994). The multiple regressions used Type III sums of squares (SS), in which the effect attributed to each factor is partitioned from the residuals after all other factors have been included in the model. Inbreeding of the dam ([f.sub.d]) can contribute to inbreeding depression (a negative [[Beta].sub.d] effect) and is also a component of ancestral inbreeding of the litter ([f.sub.a]), so the net effect of maternal inbreeding can be either positive or negative. Maternal effects ([[Beta].sub.d]) were removed prior to estimating ancestral effects ([[Beta].sub.a]), and ancestral effects were removed prior to estimating maternal effects. Therefore, tests of both effects are conservative. For each of the levels of inbreeding depression ([[Beta].sub.f] and [[Beta].sub.d]) and ancestral effects ([[Beta].sub.a] or [[Beta].sub.a(d)]), significance values for the seven fitness components examined were adjusted with sequential Bonferroni inequalities (Rice 1989) to account for the multiple tests of the hypothesis that inbreeding at that level impacts fitness in the population.
The laboratory stocks of each subspecies were maintained as three independent replicates; significant variation among replicates in the magnitude of overall inbreeding depression in mass was reported previously (Lacy et al. 1996). Although the purpose of this study was to examine the effectiveness of selection in reducing the average genetic load in each taxon, large variation between replicates in the extent of purging would suggest that any observed between-taxa trends could have arisen due to the random sampling of founders for each laboratory population. Therefore, variation between replicates was tested by comparing models in which separate effects of inbreeding ([[Beta].sub.f]) and ancestral inbreeding ([[Beta].sub.a]) were fitted for each replicate to models in which the effects of ancestral inbreeding were assumed to be the same across the three replicates of each taxon. Across the three subspecies and seven fitness measures, in only one of 21 cases was the variation among replicates in the effect of ancestral inbreeding nominally significant (P = 0.03, without adjustment for multiple tests), as could be expected to occur by chance. In addition to the relative consistency among replicates in the effects of ancestral inbreeding, the different patterns observed in the three subspecies were largely consistent across the seven measures of fitness. Consequently, in the results reported here, data are pooled across replicates to minimize standard errors and to help identify the taxon-level trends. Variation among subspecies in the effects of ancestral inbreeding was tested by comparing models in which the effects were allowed to vary among subspecies to models that estimated a common effect.
[TABULAR DATA FOR TABLE 1 OMITTED]
The presence or absence of purging of the genetic load was largely consistent across replicates of each subspecies, although taxon-level effects that were significant when data were pooled across replicates were often not significant within some individual replicates. The only nominally significant variation among replicates in the extent of purging was observed in P. p. leucocephalus. One replicate showed significant (P = 0.004) negative effects of ancestral inbreeding on RS, but the other two replicates showed no effect and the pooled effect was not significant for this subspecies. Variation among subspecies in the effect of ancestral inbreeding was significant for three of seven fitness measures (litter size: P = 0.032; mass: P = 0.012; RS: P = 0.012), and the ranking of effects across the three subspecies (rhoadsi [greater than] subgriseus [greater than] leucocephalus) was consistent for these three and for two other fitness measures.
Tables 1-3 show the regression coefficients from ancestral inbreeding models for the three subspecies. The effects of parity were removed from all regression models for litter size, viability, mass, and SD[mass]. Inbreeding depression ([[Beta].sub.f] and/or [[Beta].sub.d] significant) was observed in every fitness component measured for subgriseus and in six of seven measures in rhoadsi. Production of litters was depressed if the dam was inbred. Litter size and RS were depressed by both inbreeding of the litter and of the dam; viability and mass were depressed by litter inbreeding; while within-litter variance in mass (SD[mass]) was elevated in inbred litters, as would be expected [TABULAR DATA FOR TABLE 2 OMITTED] if inbred litters are more susceptible to environmental variation. The same trends were generally observed in leucocephalus, but the effects were often weaker and nonsignificant. This difference between leucocephalus and the other two subspecies is much less pronounced, however, in regression models that do not include ancestral inbreeding as a factor (Lacy et al. 1996).
Ancestral inbreeding of the litter or the dam, as appropriate depending on at what level inbreeding depression was expressed, was not a significant contributor to the regression model for any of the fitness measures in subgriseus (Table 1). In contrast, ancestral inbreeding had strong positive effects (countering the effects of current inbreeding) for four fitness traits in rhoadsi (Table 2). The interaction between [f.sub.a] and f was significant for SD[mass] in leucocephalus, but the effect of ancestral inbreeding was to worsen the effect off on this trait, as it was (but not significantly so) for five other fitness measures.
Figure 1 illustrates the effects of inbreeding of the litter on the composite measure of reproductive success (RS). Solid lines show the linear regressions estimated from the coefficients given in Tables 1-3 for the three subspecies, when adjusted for [f.sub.d] = 0 and [f.sub.a] = 0. Dashed lines show the inbreeding depression estimated when [f.sub.d] = 0 and [f.sub.a] = 0.25 (approximately the mean [f.sub.a] in the study). The changes in slope between pairs of lines indicate purging (in rhoadsi), no change (in subgriseus), or exacerbation (in leucocephalus) of the genetic load. Comparable figures for the six components [TABULAR DATA FOR TABLE 3 OMITTED] of overall RS are not given, because such linear interpolations do not always show well the effects of multiple generations of inbreeding. Litter inbreeding (f), maternal inbreeding ([f.sub.d]), ancestral inbreeding ([f.sub.a]), and ancestral inbreeding of the dam ([f.sub.a(d)]) all change during generations of inbreeding, and can impact fitness as complex interactions (eq. 3, 4). Full representation of their interacting effects on fitness would require a five-dimensional surface. However, in a pedigree with regular inbreeding, as in line-breeding experiments, the changes in the components of inbreeding are determined and the responses to inbreeding are easily described.
The first panel of Figure 2 shows the change in fitness under three models of selection on inbreeding depression for the levels of inbreeding, maternal inbreeding, and ancestral inbreeding that would result from four generations of fullsib matings (f = 0.000, 0.250, 0.375, 0.500, 0.594; [f.sub.d] = 0.000, 0.000, 0.250, 0.375, 0.500; [f.sub.a] = 0.000, 0.000, 0.250, 0.5313, 0.7656). These levels are comparable to those reached in the 10 generations of mixed inbreeding and outcrossing in this study; the maximum level of inbreeding attempted in our pairings was f = 0.594, 0.637, and 0.547 for subgriseus, rhoadsi, and leucocephalus, respectively. Ancestral inbreeding ranged up to [f.sub.a] = 0.8172, 0.9095, and 0.6616, respectively. In the case of no purging (solid line), fitness declines linearly with homozygosity (w/[w.sub.0] = 1 - [Beta]f). If there is 100% purging (short dashes), fitness is proportionate to the homozygosity of alleles which had not been previously homozygous in any ancestors (w/[w.sub.0] = 1 - [Beta]f + [Beta]f[f.sub.a] = 1 [Beta]f[1 - [f.sub.a]]). If 50% of the genetic load is purged when exposed to selection (long dashes), homozygosity of alleles in ancestors reduces by half the deleterious effect of future inbreeding (w/[w.sub.0] = 1 - [Beta]f + 0.5 [Beta]f[f.sub.a] = 1 - [Beta]f[1 - 0.5 [f.sub.a]]). Similarly, using the regression coefficients estimated in the complex pedigree in our study, the effects of a regular system of inbreeding can be predicted, thereby illustrating the interacting effects of generations of inbreeding expected for a simple pedigree. The seven subsequent panels of Figure 2 show the predicted responses for the three subspecies for each of the fitness measures. Error bars show the SE of each predicted value, based on the SEs of the regression coefficients.
The changes in fitness from generation 0 to 1 in Figure 2 indicate the direct inbreeding depression that occurs prior to any maternal effects or any opportunity for selection to reduce the genetic load in the stocks. From generation 1 to 2, maternal inbreeding often causes further decline in fitness, but ancestral inbreeding begins to ameliorate some of the impacts of the litter inbreeding. A cessation of further decline or even some recovery of fitness after initial inbreeding depression is apparent in rhoadsi for almost every measures of fitness, including measures that would be largely independent or even inversely related due to energetic or allometric constraints (e.g, litter size, weaning mass, within-litter variance in mass). The inbreeding depression in subgriseus is often more nearly linear. Peromyscus p. leucocephalus shows linear inbreeding depression in litter size, but for several other traits (viability, mass, SD[mass], P[2nd litter]) shows threshold effects in which the impacts are initially small but become increasingly severe at higher levels of inbreeding.
Inbreeding depression is a selective force, with inbreeding depression constituting selection against the genetic load of deleterious alleles that causes inbreeding depression. However, it is not yet known whether the genetic load revealed by inbreeding is composed primarily of genes that would respond readily to that selection (such as recessive lethal alleles), primarily of genes that would not respond readily to selection (such as a genes showing overdominant fitness effects or genes with individually small effects), or different genetic bases in different populations. Across the seven fitness measures we examined in P. p. subgriseus, there was almost no average effect of ancestral inbreeding (Table 1; [ILLUSTRATION FOR FIGURE 1 OMITTED]). In P. p. rhoadsi, however, the average effect of ancestral inbreeding was as expected if selection removes almost all deleterious alleles exposed in homozygous inbred individuals in the pedigree. When extrapolated to fully inbred animals, the effect [[Beta].sub.a] completely counteracted the effect [[Beta].sub.f] across most measures of fitness for this subspecies (Table 2). For P. p. leucocephalus, the mean effect of ancestral inbreeding on current inbreeding depression was large and negative (Table 3). Inbreeding depression in several components of fitness and in overall reproductive success became steeper when ancestors were inbred [ILLUSTRATION FOR FIGURE 1 OMITTED]. These contrasting patterns among the subspecies were consistent across most of the measures of fitness. The effects of ancestral inbreeding were ordered rhoadsi [greater than] subgriseus [greater than] leucocephalus for litter size, viability mass, SD[mass], and RS (Tables 1-3).
Ballou (1997) detected little purging of the genetic load in a study of pedigrees of 25 captive populations of mammals in zoos and research colonies. He suggested six hypotheses to explain the lack of purging in these pedigrees: (1) inbreeding depression could be due primarily to overdominant loci; (2) inbreeding depression could be due to associative overdominance, in which linkage among loci with dominance causes linked chromosome segments to act as overdominant gene complexes; (3) inbreeding depression could be due primarily to mildly deleterious alleles, rather than lethals; (4) the opportunity for selection may have been weak, because inbreeding levels were not high in many of the populations examined by Ballou; (5) small sample sizes may have precluded detecting some effects; and (6) a prior history of inbreeding in some of the populations may have reduced the genetic load before the captive populations were established. In this study, sample sizes and average levels of inbreeding were comparable to the best datasets in Ballou (1997), and much higher than the average levels in that study. The levels of inbreeding we achieved were high enough so that substantial purging of highly deleterious recessive alleles is expected (Hedrick 1994). We observed significant purging of the genetic load for four of seven fitness measures in P. p. rhoadsi. The extent of reduction of the genetic load was typically greater (but not significantly so) than would be expected with completely efficient selection (i.e., [[Beta].sub.a] [greater than or equal to] -[[Beta].sub.f] or [[Beta].sub.a(d)] [greater than or equal to] -[[Beta].sub.d]). Hence, most of the inbreeding depression in this subspecies may be caused by recessive lethal or highly detrimental alleles which were effectively purged by selection.
In P. p. subgriseus, inbreeding depression was observed in every fitness component, but ancestral inbreeding did not significantly reduce inbreeding depression in any of the seven fitness measures. We cannot determine whether the lack of reduction of the genetic load in subgriseus was due to weakly deleterious alleles, overdominant loci, or associative overdominance, because selection would be ineffective at reducing inbreeding depression caused by any of these genetic bases. However, large variation was observed among founder lineages in the depression of fitness resulting from homozygosity of their alleles (Lacy et al. 1996), suggesting that the observed inbreeding depression was caused by relatively few segregating alleles (or linked blocks of alleles). Overdominance or associative overdominance remain as possible explanations for the consistency of inbreeding depression across generations in this taxon.
The data for P. p. leucocephalus are harder to interpret. Neither the effect of inbreeding nor purging was significant for most of the traits, in part due to poor performance of even the noninbred mice (Lacy et al. 1996). In six of the seven fitness measures, however, the effect of ancestral inbreeding was to exacerbate inbreeding depression, and for one trait (SD[mass]) this effect was significant. Additionally, the effect on composite RS was significant in one of the replicate stocks of the subspecies. These negative statistical effects of ancestral inbreeding in leucocephalus were caused by a non-linear response to cumulative inbreeding, in which early generations of inbreeding (when ancestors were generally not inbred) showed little inbreeding depression, while later generations (when ancestors often were inbred) showed severe depression at the higher levels of accumulated inbreeding. McCall et al. (1994) noted that maternal effects could cause a delayed effect of inbreeding, mimicking a threshold effect in which inbreeding depression occurs only at later generations (and therefore higher levels) of inbreeding. However, we factored out maternal effects prior to testing for effects of ancestral inbreeding, so maternal effects would not have been mistakenly attributed to direct effects of the genetic load expressed in the progeny.
Although earlier studies (Brewer et al. 1990; Lacy et al. 1996) could not demonstrate that inbreeding depression is significantly lower in leucocephalus than in other populations of Peromyscus, the further analyses presented in this paper suggest that this island population may be initially less impacted by inbreeding than are mainland populations. However, multiple generations of inbreeding, rather than purging the genetic load further, result in increasingly severe effects. This suggests that there could be a residual component of the genetic load that cannot be removed by selection and that there may be threshold of inbreeding beyond which fitness declines even more rapidly. Such a threshold effect could result from synergistic epistasis, if fitness is reduced in the multiple homozygote more than expected based on independent effects of loci (Charlesworth et al. 1991). Kosuda (1972) reported a quadratic relationship between inbreeding and log(viability) in D. virilis, indicative of synergistic epistasis. Griffin and Lindgren (1985) found that a model of synergistic epistasis fit an observed threshold effect of inbreeding on seed yield in Pinus radiata. Willis (1993) found evidence for synergistic epistasis for pollen viability, but not for other fitness components, in Mimulus guttatus. In Mimulus, Dudash et al. (1997) found evidence for purging of inbreeding depression in flower production in some families, synergistic epistasis in other families, and linear inbreeding depression in yet other families.
Templeton suggested that selection acting on epistatic genetic systems could achieve even more rapid reduction in inbreeding depression than would occur from simple removal of deleterious recessive alleles (Templeton et al. 1976; Templeton 1979), and Templeton and Read (1994) proposed such a mechanism for the apparently rapid elimination of inbreeding depression in Speke's gazelle (but see Willis and Wiese 1997). However, our results with P. p. leucocephalus and the findings of the studies cited above suggest that the effects of epistasis might be at least as likely to exacerbate inbreeding depression.
Epistasis is not the only mechanism that could lead to threshold inbreeding depression. Connor and Bellucci (1979) suggested that selection for heterozygosity (due to balanced polymorphisms) could have delayed inbreeding depression in lines of Mus. However, Garcia et al. (1994) suggested that selection opposing homozygosity in Drosophila could lead to the opposite effect: a plateau in fitness depression after inbreeding reaches high levels. Frankham (1995) reported that experimental data on Drosophila and Mus often show threshold relationships between inbreeding and population extinction. Rapid extinction could result from cumulative inbreeding depressing fitness below a level necessary for population persistence, but this extinction threshold could be even more abrupt if there is also threshold relationship between inbreeding and fitness.
The effectiveness of selection can be dependent on the rate of inbreeding, with slower inbreeding allowing selection to more effectively reduce the genetic load (Ehiobu et al. 1989). The rate of inbreeding in this study, accumulating to about f = 0.6 over 10 generations, was slower than is often imposed in experiments but faster than would be expected in almost any natural population of mammals. Thus, selection may have been more effective in this study than would often be observed in short-term experiments, but perhaps less effective than might occur more gradually over many more generations in long-term studies or in natural populations.
The Effect of Population History on the Genetic Load
Among three subspecies of a single species, we found three different patterns for the effects of multiple generations of inbreeding. Notably, the subspecies differences are in accord with what is known about the structure and history of each population. Peromyscus p. rhoadsi was collected from an area of relatively continuous sand-scrub habitat, and the subspecies has been found to have the high levels of allozyme heterozygosity typical of the inland populations of the species (Selander et al. 1971; Brewer et al. 1990). It is likely that little inbreeding has occurred in the past in this population. The effects of ancestral inbreeding were to reduce inbreeding depression to an extent expected if recessive deleterious alleles were efficiently removed by selection.
Peromyscus p. subgriseus has a similar level of allozyme variation, suggesting that there has not been a recent populationwide bottleneck. However, the founders in this study were collected from fragmented patches of habitat, created by clear-cutting timber harvest within a forest matrix (where P. polionotus are not found). These patches of open habitat are suitable for P. polionotus for only a few years. Local extinction-recolonization dynamics could cause occasional localized inbreeding, even while dispersal between habitat patches would preserve genetic variation at a wider scale (Lacy 1987). Inbreeding depression in this population was unaffected by selection during generations of inbreeding in the lab, suggesting that highly deleterious recessive alleles do not contribute substantially to its genetic load. Nason and Ellstrand (1995) similarly suggested that the component of the genetic load due to recessive alleles may have been removed by selection during local bottlenecks associated with the introductions and range expansion of the wild radish Raphanus sativus.
Peromyscus p. leucocephalus is endemic to a coastal barrier island that is periodically devastated by hurricanes. The subspecies has low genetic variation relative to inland populations, as would be expected for an isolated population that experiences taxon-wide bottlenecks. Low levels of experimental inbreeding caused little or no reductions in fitness in leucocephalus. However, the population had relatively low productivity even prior to experimental inbreeding, perhaps as a consequence of losses of genetic variation that occurred earlier in the wild. (Preliminary data from ongoing studies on the effects of outcrossing indicate that this subspecies, more so than the inland subspecies, shows very large heterosis when intercrossed with other populations.) Moreover, inbreeding accumulated across generations in the lab resulted in a virtual collapse of reproductive fitness. Indeed, in both this study and a prior study (Brewer et al. 1990), experimental inbreeding of leucocephalus was restricted to levels lower than those achieved in other populations because of poor productivity. The inbreeding depression in leucocephalus was not consistent through the generations of inbreeding, but it did not show the reduction expected if it were caused by recessive lethals. The trends may result from a threshold effect of heterozygosity.
Overall, the patterns observed in these populations (Brewer et al. 1990; Lacy et al. 1996; this study) suggest that inbreeding depression may be caused by a combination of highly deleterious recessives alleles causing transitory inbreeding depression, weakly deleterious alleles that may become fixed by drift in small populations rather than removed by selection, overdominant or associative overdominant loci causing linear declines, and epistatic interactions causing threshold effects as heterozygosity is depleted by inbreeding. The balance between these components of the genetic load might reflect the past history of selection during episodes of inbreeding. There are important implications of these conclusions for natural extinction-recolonization dynamics and for conservation. Small, isolated populations with reduced genetic variability might harbor a genetic load that is not susceptible to further reduction by selection, and may even be close to dangerous thresholds of low heterozygosity. Thus, they may be vulnerable to further bottlenecks and may be poor colonizers. Larger, continuous, and heterogeneous populations may initially suffer inbreeding depression if heterozygosity is reduced below the historically high levels, but might subsequently recover fitness. Such populations may have the resilience to recover from bottlenecks and may be the source of more successful colonizers.
Measurement of Inbreeding Depression in Multigeneration Pedigrees
Assessments of inbreeding depression can be confounded by the unrelated effects of selection and drift on the experimental populations (Lynch 1988). Serial generations of sib-mating, as are often used for inbreeding studies, can be especially problematic, because the level of inbreeding, the effects of genetic drift, and response to selection all increase in parallel through the generations. In the complex breeding design we used, a wide range of inbreeding levels (including noninbred controls) were produced within each replicate stock at each generation. Thus, multigenerational trends caused by selection could be partitioned from effects of accumulated inbreeding, and these would not be confounded with divergence between lines caused by random genetic drift. If inbreeding levels of experimental stocks accumulate regularly across generations in captivity, then adaptation to captivity (domestication) could be confounded also with adaptation to inbreeding (Lacy and Horner 1996). Our breeding design used about equal numbers of inbred and noninbred pairs each generation (after the first). Therefore, any changes due to adaptation to captivity would not have been confounded with effects of inbreeding.
Our finding of complex interactions between the effects of inbreeding at the current generation and inbreeding in prior generations in the pedigree demonstrates that it is useful to partition the effects of ancestral inbreeding from the effects of newly autozygous genes. If an interaction between past and present inbreeding is ignored, then the inbreeding depression in the base populations can be underestimated or overestimated, and the dynamic nature of the genetic load will be overlooked. Estimates of inbreeding depression in subgriseus were changed little when ancestral inbreeding was added to the regression models, as we were unable to detect any effect of [f.sub.a] on this subspecies. For rhoadsi, however, omission of ancestral inbreeding from the models resulted in underestimation of the initial genetic load affecting the probability of breeding, litter size, and total reproductive success. Moreover, mass at weaning had been reported to respond positively to maternal inbreeding (Lacy et al. 1996), but this effect is now shown to be due to the contribution that maternal inbreeding makes to ancestral inbreeding (Table 2), not an independent effect of inbreeding on maternal care. For leucocephalus, models without [f.sub.a] overestimated effects of early generations of inbreeding on the probability of producing a second litter, litter viability, mass at weaning, within-litter variance in mass, and reproductive success. These important differences among the subspecies in their responses to inbreeding were not detected until the interactions of f with [f.sub.a] were considered.
G. Alaks and A. Walsh provided expert care for the 16,763 mice used in this study. J. Jimenez, S. Margulis, S. Murphy, T. Nowak, B. Randolph, M. Warneke, and M. Wilson provided assistance with animal care. We are grateful also to G. Alaks, B. Brewer, J. Dubach, P. McGill (all of the Chicago Zoological Society), J. Jimenez (University of Illinois-Chicago), R. Edwards (University of Florida), C. Petrick (Eglin Air Force Base), and J. Gore (Florida Game and Freshwater Fish Commission) for assistance with trapping mice. Collecting was conducted under permit WX90005 from the Florida Game and Freshwater Fish Commission, with permission kindly granted by Eglin Air Force Base and Ocala National Forest to collect mice from their lands. We thank P. Hedrick, R. Frankham, and an anonymous reviewer for their comments and suggestions on the manuscript. Funding was provided by National Science Foundation grant BSR-9024950 and Research Opportunity Funds of the Smithsonian Institution.
BALLOU, J. D. 1995. Genetic management, inbreeding depression, and outbreeding depression in captive populations. Ph.D. diss., University of Maryland, College Park, MD.
-----. 1997. Ancestral inbreeding only minimally affects inbreeding depression in mammalian populations. J. Hered. 88: 169-178.
BARRETT, S.C. H., AND D. CHARLESWORTH. 1991. Effects of a change in the level of inbreeding on the genetic load. Nature 352:522-524.
BARRETT, S. C. H., AND J. R. KOHN. 1991. Genetic and evolutionary consequences of small population size in plants: implications for conservation. Pp. 3-30 in D. A. Falk and K. E. Holsinger, eds. Genetics and conservation of rare plants. Oxford Univ. Press, Oxford.
BOWMAN, J. C., AND D. S. FALCONER. 1960. Inbreeding depression and heterosis of litter size in mice. Genet. Res. 1:262-274.
BREWER, B. A., R. C. LACY, M. L. FOSTER, AND G. ALAKS. 1990. Inbreeding depression in insular and central populations of Peromyscus mice. J. Hered. 81:257-266.
CHARLESWORTH, B., M. T. MORGAN, AND D. CHARLESWORTH. 1991. Multilocus models of inbreeding depression with synergistic selection and partial self-fertilization. Genet. Res. 57:177-194.
CHARLESWORTH, D., AND B. CHARLESWORTH 1987. Inbreeding depression and its evolutionary consequences. Annu. Rev. Ecol. Syst. 18:237-268.
CHARLESWORTH, D., M. T. MORGAN, AND B. CHARLESWORTH. 1992. The effect of linkage and population size on inbreeding depression due to mutational load. Genet. Res. 59:49-61.
CONNOR, J. L., AND M. J. BELLUCCI. 1979. Natural selection resisting inbreeding depression in captive wild housemice (Mus musculus). Evolution 33:929-940.
CROW, J. F. 1948. Alternative hypotheses of hybrid vigor. Genetics 33:477-487.
DARWIN, C. R. 1868. Variation of animals and plants under domestication. John Murray, London.
-----. 1876. The effects of cross and self fertilization in the vegetable kingdom. John Murray, London.
DOUMS, C., F. VIARD, A.-F. PERNOT, B. DELAY, AND P. JARNE. 1996. Inbreeding depression, neutral polymorphism, and copulatory behavior in freshwater snails: a self-fertilization syndrome. Evolution 50:1908-1918.
DUDASH, M. R., D. E. CARR, AND C. B. FENSTER. 1997. Five generations of enforced selfing and outcrossing in Mimulus guttatus: inbreeding depression variation at the population and family level. Evolution 51:54-65.
EHIOBU, N. G., M. E. GODDARD, AND J. F. TAYLOR. 1989. Effect of rate of inbreeding on inbreeding depression in Drosophila melanogaster. Theor. Appl. Genet. 77:123-127.
FALCONER, D. S. 1989. Introduction to quantitative genetics. 3d ed. Longman, New York.
FRANKHAM, R. 1995. Inbreeding and extinction: a threshold effect. Conserv. Biol. 9:792-799.
GARCIA, N., C. LOPEZ-FANJUL, AND A. GARCIA-DORADO. 1994. The genetics of viability in Drosophila melanogaster: effects of inbreeding and artificial selection. Evolution 48:1277-1285.
GRIFFIN, A. R., AND D. LINDGREN. 1985. Effects of inbreeding on production of filled seed in Pinus radiata - experimental results and a model of gene action. Theor. Appl. Genet. 71:334-343.
HAUSER, T. P., AND V. LOESCHCKE. 1996. Drought stress and inbreeding depression in Lychnis flos-cuculi (Caryophyllaceae). Evolution 50:1119-1126.
HEDRICK, P. W. 1994. Purging inbreeding depression and the probability of extinction: full-sib mating. Heredity 73:363-372.
HOSMER, D. W., AND S. LEMESHOW. 1989. Applied logistic regression. Wiley, New York.
HUSBAND, B. C., AND D. W. SCHEMSKE. 1996. Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50:54-70.
JIMENEZ, J. A., K. A. HUGHES, G. ALAKS, L. GRAHAM, AND R. C. LACY. 1994. An experimental study of inbreeding depression in a natural habitat. Science 266:271-273.
KARKKAINEN, K., V. KOSKI, AND O. SAVOLAINEN. 1996. Geographical variation in inbreeding depression of Scots pine. Evolution 50:111-119.
KELLER, L. F., P. ARCESE, J. N. M. SMITH, W. M. HOCHACHKA, AND S. C. STEARNS. 1994. Selection against inbred song sparrows during a natural population bottleneck. Nature 372:356-357.
KOSUDA, K. 1972. Synergistic effect of inbreeding on viability in Drosophila virilis. Genetics 72:461-468.
LACY, R. C. 1987. Loss of genetic diversity from managed populations: interacting effects of drift, mutation, immigration, selection, and population subdivision. Conserv. Biol. 1:143-158.
-----. 1997. The importance of genetic variation to the viability of mammalian populations. J. Mammal. 78:320-335.
LACY, R. C., AND B. E. HORNER. 1996. Effects of inbreeding on skeletal development of Rattus villosissimus. J. Hered. 87:277-287.
-----. 1997. Effects of inbreeding on reproduction and sex ratio of Rattus villosissimus. J. Mammal. 78:877-887.
LACY, R. C., A. M. PETRIC, AND M. WARNEKE. 1993. Inbreeding and outbreeding depression in captive populations of wild species. Pp. 352-374 in N. W. Thornhill, ed. The natural history of inbreeding and outbreeding. Univ. of Chicago Press, Chicago.
LACY, R. C., G. ALAKS, AND A. WALSH. 1996. Hierarchical analysis of inbreeding depression in Peromyscus polionotus. Evolution 50:2187-2200.
LANDE, R., AND D. W. SCHEMSKE. 1985. The evolution of self-fertilization and inbreeding depression in plants. I. Genetic models. Evolution 39:24-40.
LANDE, R., D. W. SCHEMSKE, AND S. T. SCHULTZ. 1994. High inbreeding depression, selective interference among loci, and the threshold selfing rate for purging recessive lethal mutations. Evolution 48:965-978.
LATTA, R., AND K. RITLAND. 1994. The relationship between inbreeding depression and prior inbreeding among populations of four Mimulus taxa. Evolution 48:806-817.
LATTER, B. D. H., J. C. MULLEY, D. REID, AND L. PASCOE. 1995. Reduced genetic load revealed by slow inbreeding in Drosophila melanogaster. Genetics 139:287-297.
LEVIN, D. A. 1984. Inbreeding depression and proximity-dependent crossing success in Phlox drummondii. Evolution 38:116-127.
LYNCH, C. B. 1977. Inbreeding effects upon animals derived from a wild population of Mus musculus. Evolution 31:526-537.
LYNCH, M. 1988. Design and analysis of experiments on random drift and inbreeding depression. Genetics 120:791-807.
MARGULIS, S. W. 1997. Inbreeding-based bias in parental responsiveness to litters of oldfield mice. Behav. Ecol. Sociobiol. 41: 177-184.
-----. 1998a. Differential effects of inbreeding at juvenile and adult life-history stages in Peromyscus polionotus. J. Mammal. 79:326-336.
-----. 1998b. Relationships among parental inbreeding, parental behaviour and offspring viability in oldfield mice. Anim. Behav. 55:427-438.
MARGULIS, S. W., AND J. ALTMANN. 1997. Behavioural risk factors in the reproduction of inbred and outbred oldfield mice. Anim. Behav. 54:397-408.
MAYER, S. S., D. CHARLESWORTH, AND B. MEYERS. 1996. Inbreeding depression in four populations of Collinsia heterophylla Nutt (Scrophulariaceae). Evolution 50:879-891.
MCCALL, C., D. M. WALLER, AND T. MITCHELL-OLDS. 1994. Effects of serial inbreeding on fitness components in Impatiens capensis. Evolution 48:818-827.
MORTON, N. E., J. F. CROW, AND H. J. MULLER. 1956. An estimate of the mutational damage in man from data on consanguineous marriages. Proc. Nat. Acad. Sci. USA 42:855-863.
NASON, J. D., AND N. C. ELLSTRAND. 1995. Lifetime estimates of biparental inbreeding depression in the self-incompatible annual plant Raphanus sativus. Evolution 49:307-316.
PRAY, L. A., AND C. J. GOODNIGHT. 1995. Genetic variation in inbreeding depression in the red flour beetle Tribolium castaneum. Evolution 49:176-188.
RALLS, K., J. D. BALLOU, AND A. TEMPLETON. 1988. Estimates of lethal equivalents and the cost of inbreeding in mammals. Conserv. Biol. 2:185-193.
RICE, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225.
SCHULTZ, S. T., AND J. H. WILLIS. 1995. Individual variation in inbreeding depression: the roles of inbreeding history and mutation. Genetics 141:1209-1223.
SELANDER, R. K., M. H. SMITH, S. Y. YANG, W. E. JOHNSON, AND G. B. GENTRY. 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I. Variation in the old-field mouse (Peromyscus polionotus). Studies in genetics. VI. Univ. of Texas Publ. 7103:49-90.
SHARP, P. M. 1984. The effect of inbreeding on competitive male-mating ability in Drosophila melanogaster. Genetics 106:601-612.
STOCKLEY, P., J. B. SEARLE, D. W. MACDONALD, AND C. S. JONES. 1993. Female multiple mating behaviour in the common shrew as a strategy to reduce inbreeding. Proc. R. Soc. Lond. B Biol. Sci. 254:173-179.
TEMPLETON, A. R. 1979. The unit of selection in Drosophila mercatorum. II. Genetic revolution and the origin of coadapted genomes in parthenogenetic strains. Genetics 92:1265-1282.
TEMPLETON, A. R., AND B. READ. 1983. The elimination of inbreeding depression in a captive herd of Speke's gazelle. Pp. 241-261 in C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and conservation: a reference for managing wild animal and plant populations. Benjamin/Cummings, Menlo Park, CA.
-----. 1984. Factors eliminating inbreeding depression in a captive herd of Speke's gazelle (Gazella spekei). Zoo Biol. 3:177-199.
-----. 1994. Inbreeding: one word, several meanings, much confusion. Pp. 91-105 in V. Loeschcke, J. Tomiuk, and S. K. Jain, eds. Conservation genetics. Birkhauser Verlag, Boston, MA.
TEMPLETON, A. R., C. F. SING, AND B. BROKAW. 1976. The unit of selection in Drosophila mercatorum. I. The interaction of selection and meiosis in parthenogenetic strains. Genetics 82:349-376.
WHITE, J. M. 1972. Inbreeding effects upon growth and maternal ability in laboratory mice. Genetics 70:307-317.
WILKINSON, L. 1994. SYSTAT for DOS. Vers. 6. SYSTAT, Evanston, IL.
WILLIS, J. H. 1993. Effects of different levels of inbreeding on fitness components in Mimulus guttatus. Evolution 47:864-876.
WILLIS, K., AND R. WIESE. 1997. Elimination of inbreeding depression from captive populations: Speke's gazelle revisited. Zoo Biol. 16:9-16.
WRIGHT, S. 1977. Evolution and the genetics of populations. Vol. 3. Experimental results and evolutionary deductions. Univ. of Chicago Press, Chicago.
|Printer friendly Cite/link Email Feedback|
|Author:||Lacy, Robert C.; Ballou, Jonathan D.|
|Date:||Jun 1, 1998|
|Previous Article:||Heritabilities of dominance-related traits in male bank voles (Clethrionomys glareolus).|
|Next Article:||Temporal change of mitochondrial DNA haplotype frequencies and female effective size in a brown trout (Salmo trutta) population.|