NATURE, Volume 352. No. 6335, 8 August 1991.

Effects of a change in the level of inbreeding on the genetic load*

S. C. H. Barrett** and D. Charlesworth***

"The effects of inbreeding may not be as noticeable in the first generation as the invigoration immediately apparent after crossing" [1]. This statement, published in 1919, has received little attention, and has apparently never been tested empirically, although the reduction of the genetic load of populations by inbreeding is well known in theoretical terms [2-5]. Because inbreeding increases homozygosity, and hence the effectiveness of selection against recessive or partially recessive detrimental alleles, changes in levels of inbreeding can lead to a reduction in the frequencies of such mutant alleles. This results in equilibration at higher population mean fitness [6] and is referred to as 'purging' populations of their genetic load. Severe inbreeding can also reduce genetic load due to overdominant alleles, provided selection coefficients are not symmetrical at all loci, because alleles giving lower fitness will be reduced in frequency at equilibrium [7,8]. With either fitness model, however, reduction in genetic load takes time, and the initial effect of an increase in inbreeding is reduced fitness due to homozygosity. There are few data relating to the extent to whcih fitness is reduced during inbreeding in a set of lines and to how long the reduction lasts before increasing again to the initial level, or higher. Inbreeding experiments involving sib mating in mice and Drosophila subobscura [10], and successive bottlenecks in house flies [11] have yielded some evidence consistent with the purging hypothesis. Here, we report results of an experiment demonstrating a prolonged time-course of recovery of mean fitness under self-fertilization of a naturally outcrossing plant, and also compare our results with expectations derived by computer calculations. Our results show that the genetic load present in an outcrossing population can be explained only with a high mutation rate to partially recessive deleterious alleles, and that inbreeding purges the population of mutant alleles.

The experimental plant was the self-compatible annual water hyacinth, Eichhornia paniculata (Pontederiaceae). The species inhabits ephemeral pools and ditches in northeastern Brazil and the Caribbean, and exhibits a wide range of natural outcrossing rates [12]. The experiments reported here used material from two populations with contrasting breeding systems. Population B11 from Brazil had an estimated outcrossing rate of 0.94, whereas population J13 from Jamaica was highly selfing, with no within-population heterozygosity observed at 21 allozyme loci [13]. A range of life-history traits was measured on all plants during the 8-month experiment. Only data on mean flower number per plant are presented here, but results for biomass were similar. Flower number exhibits the strongest inbreeding depression in E. paniculata. It represents the outcome of many developmental steps and must be influenced by many loci. As the species is annual, this trait is likely to be highly correlated with fitness. The results of five generations of self-fertilization of plants from population B11, followed by intercrossing the inbred lines, are shown in Fig. 1. Inbreeding had no significant effect on flower number per plant from the naturally inbreeding J13 population (Fig. 1; Table 1), showing that the differences observed are unlikely to be due to ageing of the stored seeds.

The expectations for such an experiment were calculated using deterministic computer programs that model either overdominant selection at several loci [14], or mutation-selection balance at many unlinked loci [5]. There was remarkable similarity between the main features of the experimental results and the theoretical expectations of the mutational model for the inbred generations, when the initial populations before inbreeding begins were subject to a high mutation rate to moderately detrimental, partially recessive alleles (Figs. 1, 2; Table 1). Mean fitness values decreased in the first generation of inbreeding, but subsequently remained at roughly the same level. After five generations of inbreeding, the mean fitness of progeny derived from intercrossing the inbred lines was much higher than that of the inbred progeny, and also considerably higher than the fitness of the original outbred population. With strongly deleterious mutations (selection coefficient, 0.9), rapid recovery of fitness during the five generations of inbreeding was clearly evident (Fig. 2b; Table 1), unlike the experimental results. With a partially inbred initial population, the results were similar to those observed (Table 1).

With the asymmetrical overdominance model, populations at equilibrium under high selfing do not maintain variation, but will fix the allele least disadvantageous in homozygotes though an increase will occur if the set of lines has not reached equilibrium, so that some still remain with the lower fitness allele. One would thus expect no increase in fitness on intercrossing the inbred lines after the population reaches equilibrium. During inbreeding, the fitness values behaved similarly to those in the mutational model, with a slight decrease, rather than an increase, in fitness (Fig. 2c; Table 1). But the intercross mean fitness, although higher than the inbred fitness, was never higher than the fitness of the initial population, unlike the mutational load model and the observed data (Table 1).

These results support the partial dominance (mutational) hypothesis [2, 7] for genetic load, and suggest that the mutation rate per genome in E. paniculata must be very high (of the order of one per generation). This is consistent with most other available data [15, 16]. The failure of fitness to recover during inbreeding, and intercrossed fitness levels higher than those of the original outbred strains, are also seen in maize [17] and mice [9] (Table 1). This suggests that purging of partially recessive mutations accounts for an important component of heterosis. When the fitness values under inbreeding are plotted against the inbreeding coefficient, the partial dominance model yields an increase with high F values, due to purging, but the overdominance model predicts a monotonic decline. The partial dominance model can thus also explain the observed patterns in inbreeding experiments with conifers [18].

The occurence of natural selection during the inbreeding experiment, consistent with purging, was indicated by excess heterozygosity of five (presumably neutral) allozyme marker loci scored in the experimental plants, each generation of inbreeding (data not shown). Under inbreeding, heterozygotes tent to be produced by outcrosses. Surviving adults in the progeny generation therefore tend to be heteozygotes [19, 20], and excess heterozygosity compared with the neutral expectation is frequently found in inbreeding plants [21].

These results are also relevant to populations undergoing disturbance by humans. Inbreeding of formerly outbred populations occurs in many zoo populations, as well as domesticated populations and populations subjected to severe reductions in size as a result of human alterations of the environment. Extreme inbreeding has been recommended to purge genetic load and force the adaptation of endangered populations to the inbreeding regime they will experience under human management [22]. This assumes that lowered fitness inevitably caused by increased homozygosity during this process will not be too severe or prolonged. The validity of this assumption depends on the severity and dominance of the mutant alleles in the population being inbred [23, 5]. Our results indicate that purging by the most severe inbreeding, self-fertilization, could decrease fitness considerably, with little recovery under inbreeding, and that fitness is restored only when when the inbred lines are intercrossed.

SUPPORTING DATA:

FIG. 1 Mean flower numbers of plants in various inbred and intercrossed generations. [a] Naturally outbred population B11, NE Brazil. [b] Naturally inbred population J13, Jamaica. METHODS. Seed from open-pollinated families were randomly sampled from the two populations, B11 and J13. Lines were established in the glasshouse from each of 30 families from B11, and from each of 10 families from J13. One plant from each line was selfed and randomly outcrossed to other plants from the same population, to give the S1 and O1 generations, respectively. A small amount of S1 seed was sown and a single randomly chosen plant was self-pollinated to give the S2 generation. The remaining S1 and O1 seeds were stored dry at room temperature. The selfing was repeated for four successive generations with seeds from the S generations being stored. Plants from the lines were then selfed and outcrossed (within populations) to give the S5 and O5 generations. Stored seeds (maximum age 3 years) from all generations were then sown and the multi-generation plants grown in the glasshouse under uniform conditions. Fitness components were compared among generations in a randomized block design with all lines represented. Sample sizes were n = 1007 for population B11 (average 4.8 plants per line per generation), and n = 378 for J13 (average 5.4 plants per line per generation). Means and standard errors are shown and the letters above the bars indicate statistically significant differences (5% level).

FIG. 2 Theoretical results for different selfed and intercrossed generations, assuming unlinked loci. [a] Mutational model: selection coefficient s = 0.2; dominance coefficient h = 0.2; mutation rate per diploid genome U = 1.0. [b] Mutational model: s = 0.9; h = 0.2; U = 0.1. [c] Overdominance model: selection coefficients s1 = 0.1 and s2 = 0.2 at four loci. METHODS. The computer calculations for the mutational model were obtained using the infinite population size model of Kondrashov [24] programmed as described [5]. In this model, the fitness of a homozygote for a mutant allele at one locus was denoted by 1 - s, where s is the selection coefficient, and the fitness of heterozygotes by 1 - hs, where h is the dominance coefficient. For the overdominance model, a five-locus deterministic program was used, with symmetric selection coefficients s1 and s2 (with the biologically implausible assumption of symmetrical overdominance, the mean fitness declines with inbreeding even though variation is maintained with both alleles at frequencies of 0.5, and the fitness on intercrossing inbred lines will not exceed the original outbred value). To combine the fitness effects of different loci, multiplicative fitnesses were assumed in both models. For each model populations were run until equilibrium was reached under almost complete outcrossing (outcrossing rate 0.99), and then the selfing rate was altered to a high value (outcrossing rate 0.01). The time-course of changes in the population was then followed. Each generation, the fitnesses of inbred and outbred progeny produced in the population were calculated.

Footnotes and References:

1. East, E. M. & Jones, D. F. Inbreeding and Outbreeding: Their Genetic and Sociological Significance 147 (Lippincott, Philadelphia, 1919).

2. Crow, J. F. Genetics 33, 477-487 (1948).

3. Wright, S. Evolution and the Genetics of Populations Vol. 3 (Univ. of Chicago Press, 1977).

4. Lande, R. & Schemske, D. W. Evolution 39, 24-40 (1985).

5. Charlesworth, D., Morgan, M. T. & Charlesworth, B. Evolution 44, 1469-1489 (1990).

6. Crow, J. F. in Mathematical Models in Population Genetics (eds Kojima, K. -I) 128-177 (Springer-Verlag, Berlin, 1970).

7. Charlesworth, D. & Charlesworth, B. A. Rev. ecol. Syst. 18, 237-268 (1987).

8. Ziehe, M. & Roberds, J. H. Genetics 121, 861-868 (1989).

9. Falconer, D. S. Genet. Res. Camb. 17, 215-235 (1972).

10. Hollingsworth, M. J. & Maynard-Smith, J. J. Genet. 53, 295-314 (1955).

11. Bryant, E. H., Meffert, L. M. & McCommas, S. A. Amer. Nat. 136, 542-549 (1990).

12. Barrett, S. C. H. & Husband, B. C. Pl. Spec. Biol. 5, 41-55 (1990).

13. Glover, D. E. & Barrett, S. C. H. Heredity 59, 7-17 (1987).

14. Charlesworth, D. & Charlesworth, B. Evolution 44, 870-888 (1990).

15. Kondrashov, A. Nature 336, 435-440 (1988).

16. Charlesworth, B., Charlesworth, D. & Morgan, M. T. Nature 347, 380-382 (1990).

17. Shull, G. H. in Heterosis (ed. Gowen, J. W.) 14-48 (Iowa University Press, Ames, 1950).

18. Griffin, A. R. & Lindgren, D. Theor. appl. Genet. 71, 334-343 (1985).

19. Ohta, T. & Cockerham, C. C. Genet. Res. Camb. 23, 191-200 (1974).

20. Sing, C. F., Brewer, G. J. & Thirtle, B. Genetics 75, 381-404 (1973).

21. Brown, A. H. D. Theor. appl. Genet. 15, 1-42 (1979).

22. Templeton, A. R. & Read, B. Zoo Biol. 3, 177-199 (1984).

23. Wilton, A. N., Joseph, M. G. & Svend, J. Genet. Res. Camb. 54, 129-140 (1989).

24. Kondrashov, A. S. Genetics 111, 635-653 (1985).

ACKNOWLEDGEMENTS. This research was funded in part by an operating grant from the Natural Sciences and Engineering Research Council of Canada (to S.C.H.B.) and in part by an NSF grant to D. C. We thank W. Cole for technical assistance and advice, and B. Charlesworth for his comments on the manuscript.

* Zanzara Afghans Reg'd CKC would like to thank Nature magazine, Dr. S. C. H. Barrett and Dr. D. Charlesworth for kindly granting permission to reprint this very important study/article on our website. We are thankful for the continued efforts of such excellent scientists and to Nature magazine for providing such a wonderful forum for Scientific work.

** Dr. Spencer C. H. Barrett. Canada Research Chair. Department of Botany, University of Toronto.

*** Dr. Deborah Charlesworth. Institute of Cell, Animal and Population Biology, University of Edinburgh.