Abstract
The European wild rabbit (Oryctolagus cuniculus) is an introduced species in Britain, and populations have been profoundly influenced by both man and disease. In stable environmental conditions, distinct social behaviour is observed, and this social structure leads to significant genetic structuring at the intrapopulation level. In this study, European wild rabbits were sampled from 17 sites across the East Anglian region of Britain and genotyped with nine microsatellite loci. Genotypical proportions deviated significantly from Hardy–Weinberg equilibrium, reflecting a degree of population subdivision and non-random mating. Several estimates of measures of population genetic structure revealed that populations are genetically distinct and have small effective population sizes. These distinctive properties are seen to be the combined effects of the social structure and random drift acting on bottlenecked populations after myxomatosis. It is concluded that the genetic structure seen in rabbit populations today is unlikely to reflect historical structuring present before myxomatosis, but that it results from recent events.
Similar content being viewed by others
Introduction
The European wild rabbit (Oryctolagus cuniculus) originated in southern Spain and north Africa and now has a widespread global distribution, primarily because of the influence of man. It was introduced into Britain in the eleventh century by the Normans, who kept captive populations of rabbits bred as a food and fur resource. Wild populations were rare owing to lack of a suitable habitat and an abundance of predators. In the eighteenth century, the growing of winter crops combined with an increased interest in game resulted in food resources becoming available and in the control of predators. This meant that populations could survive and increase in the wild (Sheail, 1971). Within 200 years, populations had expanded over most of the UK. In 1952, the myxoma virus was introduced to France in an attempt to control rapidly increasing populations and, by 1953, the virus had spread to Britain where it caused an initial mortality of up to 99.9%. Although rabbits have developed some immunity to the disease, myxomatosis is still thought to have a considerable impact on population sizes.
Within the East Anglian region of Britain, the history of the establishment and distribution of rabbit warrens is well documented. Before the sixteenth century, rabbit warrening was largely restricted to areas of light sandy soil (Bailey, 1991). Many areas of forest and heavy soil were inaccessible to rabbits until as late as the eighteenth or nineteenth centuries when land use began to change, and populations of wild rabbits became established.
The European wild rabbit is a burrowing animal that has evolved coloniality, the benefits of which include increased protection from predators, shelter and availability of nesting sites provided by the group warren, along with desirable changes in the quality and quantity of food produced by group foraging effects (Bell, 1983). Within this system, small, stable breeding groups are formed, and linear dominance hierarchies are observed in both males and females, with dominant individuals generally exhibiting higher reproductive success (Bell, 1983). Natal dispersal is sex biased; males disperse and females tend to remain within their breeding group (Webb et al., 1995). This social structuring within a population of European wild rabbits is seen to result in higher relatedness among females within a social group than among males (Surridge et al., 1999). An overall reduction in gene flow is observed, leading to genetic structuring of the population, with breeding groups constituting genetically isolated units.
The fine-scale genetic structuring resulting from social behaviour may be expected to have a significant impact on the larger scale population structure of the European wild rabbit. This study investigates the genetic structure of rabbit populations in the East Anglian region of Britain using nine polymorphic microsatellite loci, with the aims of examining: (i) the wider scale influences of social behaviour; (ii) the historical effects of founder events as populations expanded across the East Anglian region; and (iii) the influence of the myxomatosis bottleneck on population genetic structure. Populations from areas where rabbits were established initially were sampled together with areas where rabbit populations would have become established around 700 years later. The influence of population bottlenecks and founder effects was assessed by estimating genetic diversity within these different populations.
Materials and methods
Sampling
Seventeen sites across East Anglia were sampled; between 20 and 55 individual rabbits were collected from each site. The approximate defended home territory of a rabbit is 0.25 ha; samples for this study were collected from areas of up to several hundred hectares. The distribution of sample sites across East Anglia is given in Fig. 1. DNA was extracted from small pieces of ear tissue taken from rabbits shot as part of routine control procedures. Ear tissue was preserved in a high-salt tissue preservation buffer.
Genetic analysis
DNA extraction was performed using standard techniques (overnight cell lysis using proteinase K, SDS and EDTA; protein purification using chloroform followed by isopropanol precipitation) and polymerase chain reaction (PCR) amplification and genotyping of microsatellite loci performed as described previously for the nine loci: sol03, sol08, sol30, sol33, sol44, sat5, sat7, sat8 and sat12 (Surridge et al., 1997).
Statistical analysis
First, we estimated the genetic diversity within the populations sampled using the Shannon–Weaver diversity index (Shannon & Weaver, 1964). This index integrates two aspects of diversity, variant richness as well as the frequency of each variant, i.e. the number of alleles and their frequencies. For an infinite population, diversity is estimated by H′ = –∑pilnpi, where pi is the frequency of variant i, in this case allele i. The bias in H′ resulting from finite sample size is small and can be ignored in most cases (Peet, 1974). H′ is maximized when each individual sampled carries unique alleles; thus, maximum H′ is given by lnn, where n is the total number of alleles sampled. Relative diversity was calculated as the ratio of H′ to maximum H′. These diversity indices should reveal the effects of past bottlenecks and/or the founding of populations from a small number.
Secondly, we tested the genotypes at each locus for each population for deviation from the expected Hardy–Weinberg ratios using the software POPGENE (Yeh et al., 1996), which performs both chi-squared and likelihood ratio tests of statistical significance. Rabbits show a high degree of social structure and, therefore, it is unlikely that the genotypical proportions in the sampled populations will conform to Hardy–Weinberg expectations. However, assuming that any heterozygote deficit found is caused entirely by the effects of null alleles, it is possible to calculate the frequency of these null alleles from the expected (He) and observed (Ho) heterozygosity values. This can be given by:
where r is the frequency of null alleles.
Thirdly, Wright’s (1951) FST, FIT and FIS (based on an infinite allele model of mutation) were estimated in the form of θ,F and f, the sample-based, respective estimators of these parameters proposed by Weir & Cockerham (1984). These were computed using FSTAT (Goudet, 1995). Because there is no general consensus over which model of mutation is best applied to microsatellite data (Di Rienzo et al., 1994), we also obtained estimates of Slatkin’s (1995) RST (based on a stepwise mutation model) in the form of φST values calculated using AMOVA (Michalakis & Excoffier, 1996). Pairwise genetic distances (in the form of θ) were plotted against pairwise geographical distances in kilometres. A Mantel test (Mantel, 1967) was used to test for a significant relationship.
Also estimated were Nm or gene flow in the form of number of migrants exchanged per generation, and Ne, the effective population size. The relationship Nm = (1/4FST) – 0.25 (Wright, 1951) was used to estimate the number of migrants per generation from θ, the estimator of FST (where equilibrium conditions in terms of the effective size of a population over generations and the balance between drift and migration are assumed). Slatkin’s (1985) method of estimating Nm, based on the distribution of rare or private alleles, was calculated (with refinements) in GENEPOP (Raymond & Rousset, 1995).
Similarly, where steady state, neutrality and an infinite allele model are assumed, Ewens (1972) has derived the expectation of the mean number of different alleles per locus in a sample as a function of the rate of mutation μ, sample size n and the effective population size Ne. Thus, for a given sample size and observed number of alleles, the value of 4Neμ was obtained, and Ne was calculated assuming a mutation rate of 10–3, which lies between reported rates for microsatellites in rodents and humans (Dallas, 1992; Weber & Wong, 1993). The formulae for Ne derived by Crow & Kimura (1970) and Ohta & Kimura (1973) (where Ne is a function of μ and the observed heterozygosities for the infinite and single-step stepwise mutation models respectively) were used to obtain Ne values for comparative purposes.
Results
Microsatellite loci
All the microsatellite loci showed polymorphism, having between eight and 17 alleles and heterozygosities ranging from 0.24 to 0.72 (Table 1). The mean number of alleles per locus ranged from 3.45 to 7.78 in the 17 populations. Loci showing the greatest number of alleles tended to be interrupted repeats, for example sol03 (17 alleles), sat5 (17 alleles), sol33 (16 alleles) and sol30 (15 alleles). However, there appeared to be no clear correlation between number of repeats and number of alleles in our sample of microsatellite loci. Population-specific allele frequencies for all loci are given in the Appendix.
Departures from Hardy–Weinberg equilibrium
The social structure of the European wild rabbit prevents random mating. Reflecting this, all loci were seen to deviate significantly from Hardy–Weinberg equilibrium using both chi-squared and likelihood ratio tests (P < 0.001). Observed and expected heterozygosities are given in Table 1, together with the numbers of alleles observed per locus. All loci except sat12 showed a significant heterozygote deficit. Possible reasons for this deficit, apart from population subdivision and assortative mating, include selection against heterozygotes and null alleles. Estimates of r for each locus are given in Table 1 (r1 from eqn 1, r2 from eqn 2). These frequencies are high for some loci, e.g. sol44 and sat7. The frequency of blank genotypes in the data set (possible null–null homozygote genotypes) ranged from 0.022 (for sat12) to 0.052 (for sat7).
Population differentiation and gene flow
The estimators of population differentiation based on the infinite alleles and stepwise mutation models showed a high degree of population differentiation (θ = 0.150, P < 0.001; φST = 0.198, P < 0.001). Single-locus estimates of F, f, θ and φST are given in Table 2. All loci show values of θ and φST significantly greater than zero (P < 0.005). Estimates of gene flow, Nm, calculated from θ and φST are 1.42 and 1.01 respectively. Using the private alleles method, a higher value of 2.82 individuals per generation was obtained.
Plots of pairwise θ against geographical distance (Fig. 2) showed no apparent correlation between genetic and geographical distance (Mantel test, P > 0.05).
Effective population size
For a value of μ = 10–3, we obtain estimates of effective population size of 541 using Ewens’ formula, where the number of alleles per locus was averaged over loci and over populations. The methods based on heterozygosity values gave estimates of 253 and 380 for the infinite alleles and stepwise mutation models, respectively, where heterozygosities were averaged over loci and over populations.
Diversity within populations
The Shannon–Weaver diversity index obtained for populations ranged from 0.92 to 1.79. Relative diversity ranged from 0.082 to 0.136. Of the populations sampled, 10 could be considered ‘ancient’ populations, i.e. rabbit populations were known from historical records to be present in those areas before the sixteenth century (M.␣Bailey, personal communication), and the remaining seven were sampled from areas probably colonized by rabbits from the eighteenth century onwards, termed ‘recent’ populations. As rabbits spread out from existing populations into previously unoccupied habitats in the eighteenth or nineteenth centuries, we might expect to see a loss of genetic diversity resulting from a small number of initial founders. Mean diversity for recent populations = 0.112 (range 0.082–0.125) and, for ancient populations, mean diversity = 0.114 (range 0.096–0.136). There is no significant difference between these two sets of values. These reveal that there is no relationship between the site of the population (ancient site or recent site) and the diversity found within the population. Hence, the data show no evidence for founder effects in populations occupying areas radiating out from sites supporting ancient populations.
Discussion
The European wild rabbit exhibits a high degree of non-random mating and social structuring, and a consequence of this is a high degree of genetic structuring within a population (Surridge et al., 1999). When dealing with sampled populations, therefore, it must be considered that Hardy–Weinberg assumptions do not apply, and that samples are being taken across genetically isolated breeding subunits. This is reflected in the data presented here, a marked effect being a large deviation from expected Hardy–Weinberg genotype frequencies in the form of a heterozygote deficit. Thus, the combined effect of sampling across many breeding subunits, within which individuals are known to mate assortatively, is the major cause of the observed heterozygote deficit (Surridge et al., 1999). Null alleles could also be contributing to the deficit observed, but the frequency at which they occur remains uncertain.
A high degree of population differentiation is observed between the 17 populations, indicating that the populations are substantially isolated. Because of the low levels of gene flow, there appears to be no significant trend of isolation by distance. A wide spread of pairwise genetic distances is observed, which cannot be correlated with geographical distance.
These data are in good agreement with the results of previous studies on genetic structure of geographical populations of European wild rabbits. Richardson et al. (1980) found that there was significant variation in local populations of rabbits in New South Wales using allozyme data, but this variation showed no clear pattern and was attributed to drift because of small effective population sizes. In contrast, rabbit populations in an arid region of Queensland show no genetic differentiation and high levels of gene flow over 1600 km2. Moving into a semiarid ecosystem, populations become more structured (Fuller et al., 1997). This difference can be explained by differences in ecology and demography. It has been shown that the genetic structure of eastern cottontail populations (Sylvilagus floridanus) is not solely a result of social structure (Scribner & Chesser, 1993). In fact, it is the influence of environmental parameters that determine the social behaviour, which, in turn, influences the genetic structure. Gene flow between populations may be increased in three ways: population expansion in favourable conditions; successful dispersal into recently occupied areas after a population crash (Daly 1979); and mass emigration when resource shortage develops. So, we may suggest that, in favourable, more stable conditions such as those experienced in Britain, strict, stable social organization develops, leading to reduced gene flow and small effective population sizes. If effective population size is small, this is expected to lead to changes in the genetic structure of a population because of random drift and, hence, distinct population differentiation.
Estimates of effective population size corresponded well with others obtained previously for the European wild rabbit. From ecological studies, a figure of 120 has been obtained (Daly, 1979) and, using an isolation by distance calculation from distance moved between birth and breeding, Richardson (1981) estimated an effective population size of 170. Over non-arid regions of Australia, effective population size has been calculated at between 220 and 340 using coalescence theory (Zenger, 1996), showing close agreement with our estimates of 253–380 based on heterozygosity under an infinite alleles model and a stepwise mutation model. However, caution must be applied when interpreting estimates of effective population sizes obtained using models that have certain assumptions, such as fixed population size and random mating, which are unrealistic in a species displaying pronounced social behaviour such as the European wild rabbit.
Our data show no evidence for founder effects reducing genetic variation within populations as rabbits expanded into unoccupied habitats. A possible reason for this is the effects of the population bottleneck experienced as rabbit populations were exposed to myxomatosis later on in their history. The resulting survivors in a population would have undergone loss or fixation of different alleles because of drift, leading to genetic differentiation of geographically isolated populations and the obscuring of much genetic structure induced by past events.
It is often assumed that, when a population goes through a severe bottleneck, random genetic drift will induce a massive loss of genetic variability. Despite the extremely large mortality induced by myxomatosis, much variation is still observed within rabbit populations, with average heterozygosity being 0.520. The reduction in heterozygosity induced by a population bottleneck depends not only on bottleneck size but also on the rate of population growth after the bottleneck, with rapid growth limiting the severe effects of drift to a few generations. However, the loss of alleles is largely dependent on bottleneck size only (Nei et al., 1975), because the bottleneck tends to eliminate many low-frequency alleles from the population. It is proposed that the rabbit’s short generation time, together with high fecundity leading to a rapid population growth rate, could account for the degree of heterozygosity observed in present populations, despite the severity of the bottleneck. On the other hand, despite the degree of variation still observed in European wild rabbit populations, there can be no doubt that a disease such as myxomatosis resulting in the death of 99.9% of a population must have a significant effect on the genetic structure of that population.
To conclude, the present-day distribution and abundance of the European wild rabbit has been greatly modified by both man and disease. The genetic structure of populations reflects the social structure of the rabbit and is influenced by genetic drift. Hence, any genetic structuring present in European wild rabbit populations before the myxomatosis outbreak, for example influences of man, such as introductions or translocations, or those induced by natural colonization of populations into unoccupied habitats, is not evident in present populations.
References
Bailey, M. (1991). The rabbit and the Mediaeval East Anglian economy. Agric Hist Rev, 36: 1–20.
Bell, D. J. (1983). Mate choice in the European rabbit. In: Bateson, P. P. (ed.) Mate Choice pp. 211–223. Cambridge University Press, Cambridge.
Brookfield, J. F. Y. (1996). A. simple new method for estimating null allele frequency from heterozygote deficiency. Mol Ecol, 5: 453–455.
Chakraborty, R., Deandrade, M., Daiger, S. P. and Budowle, B. (1992). Apparent heterozygote deficiencies observed in DNA typing data and their implications in forensic applications. Ann Hum Genet, 56: 45–57.
Crow, J. F. and Kimura, M. (1970). An Introduction to Population Genetics Theory. Harper and Row, New York.
Dallas, J. F. (1992). Estimation of microsatellite mutation rates in recombinant strains of mouse. Mamm Genome, 3: 452–456.
Daly, J. (1979). The Ecological Genetics of the European Wild Rabbit (Oryctolagus cuniculus (L.)) in Australia. Ph.D. Thesis, Australian National University.
Dirienzo, A., Peterson, A. C., Garza, J. C., Valdes, A. M., Slatkin, M. and Freimer, N. B. (1994). Mutational processes of simple-sequence repeat loci in human populations. Proc Natl Acad Sci USA, 91: 3166–3170.
Ewens, W. J. (1972). The sampling theory of selectively neutral alleles. Theor Pop Biol, 3: 87–112.
Fuller, S. J., Wilson, J. C. and Mather, P. B. (1997). Patterns of differentiation among wild rabbit populations Oryctolagus cuniculus L. in arid and semiarid ecosystems of north-eastern Australia. Mol Ecol, 6: 145–153.
Goudet, J. (1995). FSTAT, version 1.2: a computer program to calculate F-statistics. J Hered, 86: 485–486.
Mantel, N. (1967). The detection of disease clustering and a generalised regression approach. Cancer Res, 27: 209–220.
Michalakis, Y. and Excoffier, L. (1996). A generic estimation of population subdivision using distances between alleles with special reference for microsatellite loci. Genetics, 142: 1061–1064.
Nei, M., Maruyama, T. and Chakraborty, R. (1975). The bottleneck effect and genetic variability in populations. Evolution, 29: 1–10.
Ohta, T. and Kimura, M. (1973). A model of mutation appropriate to estimate the number of electrophoretically detectable alleles in a finite population. Genet Res, 82: 201–204.
Peet, R. K. (1974). The measurement of species diversity. Ann Rev Ecol Syst, 5: 285–307.
Raymond, M. and Rousset, F. (1995). GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J Hered, 86: 248–249.
Richardson, B. J. (1981). The genetic structure of rabbit populations. In: Myers, K. and MacInnes, C. D. (eds) Proceedings of the World Lagomorph Conference.. pp. 37–52. University of Guelph, Guelph.
Richardson, B. J., Rogers, P. M. and Hewitt, G. M. (1980). Ecological genetics of the wild rabbit in Australia. II. Protein variation in British, French and Australian rabbits and the geographical distribution of the variation in Australia. Aust J Biol Sci, 33: 371–383.
Scribner, K. T. and Chesser, R. K. (1993). Environmental and demographic correlates of spatial and seasonal genetic structure in the eastern cottontail (Sylvilagus floridanus). J␣ Mammal, 74: 1026–1044.
Shannon, C. E. and Weaver, W. (1964). The Mathematical Theory of Communication.. University of Illinois Press, Urbana.
Sheail, J. (1971). Rabbits and their History.. David and Charles, Newton Abbott.
Slatkin, M. (1985). Rare alleles as indicators of gene flow. Evolution, 39: 53–65.
Slatkin, M. (1995). A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139: 457–462.
Surridge, A. K., Bell, D. J., Rico, C. and Hewitt, G. M. (1997). Polymorphic microsatellite loci in the European wild rabbit (Oryctolagus cuniculus) are also amplified in other lagomorph species. Anim Genet, 28: 302–305.
Surridg E, A. K., Ibrahim, K. M., Bell, D. J., Webb, N. J., Rico, C. and Hewitt, G. M. (1999). Fine-scale genetic structuring in a natural population of European wild rabbits (Oryctolagus cuniculus). Mol Ecol, 8: 299–307.
Webb, N. J., Ibrahim, K., Bell, D. J. and Hewitt, G. M. (1995). Natal dispersal and genetic structure in a population of the European wild rabbit (Oryctolagus cuniculus). Mol Ecol, 4: 239–247.
Weber, J. L. and Wong, C. (1993). Mutation of human short tandem repeats. Hum Mol Genet, 2: 1123–1128.
Weir, B. S. and Cockerham, C. C. (1984). Estimating F-statistics for the analysis of population structure. Evolution, 38: 1358–1370.
Wright, S. (1951). The genetical structure of populations. Ann Eugen, 15: 323–354.
Yeh, F. C., Rongcai, Y., Mao, J., Zhihong, Y. and Boyle, T. J. B. (1996). POPGENE, the Microsoft Windows-based user-friendly software for population genetics analysis of co-dominant and dominant markers and quantitative traits. Department of Natural Resources, University of Alberta, Alberta, Canada.
Zenger, K. (1996). The Genetic Variation and Evolution of the Mitochondrial DNA Non-Coding Region in Australian Wild Rabbit Populations ( Oryctolagus cuniculus (L.)). M.Sc. Thesis, University of Western Sydney.
Acknowledgements
We wish to thank D. Bullock, N. Mackinnon, P. Scott, K. Turner and The National Trust, Harvey’s Game Butcher, A. S. Kerslake, J. Farmer, J. Rudderham and M. Wright. We also thank L. Excoffier for the program used to calculate Ewens’ theta. This work was supported by grants from the BBSRC (D. J. Bell and G. M. Hewitt) and the Leverhulme Trust (K. M. Ibrahim).
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
(See Table 3 below)
Rights and permissions
About this article
Cite this article
Surridge, A., Bell, D., Ibrahim, K. et al. Population structure and genetic variation of European wild rabbits (Oryctolagus cuniculus) in East Anglia. Heredity 82, 479–487 (1999). https://doi.org/10.1038/sj.hdy.6885110
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/sj.hdy.6885110
Keywords
This article is cited by
-
Survival costs of fast exploration during juvenile life in a small mammal
Behavioral Ecology and Sociobiology (2015)
-
Genetic diversity comparison of the DQA gene in European rabbit (Oryctolagus cuniculus) populations
Immunogenetics (2015)
-
Diversity and evolutionary history of the MHC DQA gene in leporids
Immunogenetics (2008)
-
Genetic structure, lack of sex-biased dispersal and behavioral flexibility in the pair-living fat-tailed dwarf lemur, Cheirogaleus medius
Behavioral Ecology and Sociobiology (2007)