Population Genetics of Caenorhabditis elegans: The Paradox of Low Polymorphism in a Widespread Species

Corresponding author: Jody Hey, Rutgers University, Nelson Biological Labs, 604 Allison Rd., Piscataway, NJ 08854-8082., hey{at}biology.rutgers.edu (E-mail)

Communicating editor: P. D. KEIGHTLEY

	ABSTRACT

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Caenorhabditis elegans has become one of the most widely used model research organisms, yet we have little information on evolutionary processes and recent evolutionary history of this widespread species. We examined patterns of variation at 20 microsatellite loci in a sample of 23 natural isolates of C. elegans from various parts of the world. One-half of the loci were monomorphic among all strains, and overall genetic variation at microsatellite loci was low, relative to most other species. Some population structure was detected, but there was no association between the genetic and geographic distances among different natural isolates. Thus, despite the nearly worldwide occurrence of C. elegans, little evidence was found for local adaptation in strains derived from different parts of the world. The low levels of genetic variation within and among populations suggest that recent colonization and population expansion might have occurred. However, the patterns of variation are not consistent with population expansion. A possible explanation for the observed patterns is the action of background selection to reduce polymorphism, coupled with ongoing gene flow among populations worldwide.

THE nematode Caenorhabditis elegans is among the most widely studied model organisms in current biological and biomedical research. In the area of population genetics, however, research on C. elegans lags behind that on other research organisms. For example, population genetic research on Drosophila has led to the development of tests of natural selection (HUDSON et al. 1987 ; MCDONALD and KREITMAN 1991 ; MORIYAMA and POWELL 1996 ; MCALLISTER and MCVEAN 2000 ), the mapping of quantitative trait loci (MACKAY 2001 ), and the study of diverse biological processes such as speciation (KLIMAN et al. 2000 ), codon usage bias (POWELL and MORIYAMA 1997 ), and aging (SCHMIDT et al. 2000 ). In Arabidopsis thaliana, research is increasingly drawing upon population genetic studies to elucidate a range of biochemical and developmental pathways (PURUGGANAN and SUDDITH 1998 , PURUGGANAN and SUDDITH 1999 ; MALOOF et al. 2000 , MALOOF et al. 2001 ), and to develop tools for fine-scale mapping (NORDBORG et al. 2002 ). However, in the case of C. elegans evolutionary inquiries are often limited to comparisons with the closest known relative, C. briggsae, which is thought to have diverged from C. elegans 20–120 million years ago (KENT and ZAHLER 2000 ), or with more distantly related species such as Pristionchus pacificus and Oscheius sp. (DELATTRE and FELIX 2001 ; GRANDIEN and SOMMER 2001 ). Such comparisons cannot effectively address questions at the microevolutionary level. The importance of population genetic studies and baseline data on intraspecific variation in C. elegans is only now being recognized (DELATTRE and FELIX 2001 ).

C. elegans reproduces mainly by self-fertilization, which can reduce the effective population size, and thus polymorphism levels, by up to one-half (POLLAK 1987 ). A further reduction in polymorphism at neutral loci is expected due to the effects of genetic hitchhiking (MAYNARD SMITH and HAIGH 1974 ) and background selection (CHARLESWORTH et al. 1993 ), both of which are enhanced when recombination is reduced, as will occur under high rates of self-fertilization. Notwithstanding these factors, other aspects of C. elegans population biology lead to an expectation of high levels of variation, at least on a global scale. C. elegans occurs in most parts of the world (ABDUL KADER and COTE 1996 ; HODGKIN and DONIACH 1997 ) and may well be very numerous in soils or other habitats. Even if population genetic factors associated with high rates of self-fertilization cause variation to be reduced within one population, the worldwide distribution leads to the prediction of large amounts of variation among populations. However, very low levels of DNA sequence polymorphism have been observed at four loci (three nuclear and one mitochondrial) on a global scale (KOCH et al. 2000 ; GRAUSTEIN et al. 2002 ).

A species-wide lack of genetic variation in a widespread and numerous organism is paradoxical. In this study, we examine whether these same low levels of variation are seen at numerous loci located throughout the genome. We examine variation at microsatellite loci for evidence of geographic structure, gene flow, and population expansion. Our results provide insights into the recent evolutionary history of C. elegans and identify processes that might have given rise to the observed patterns of genetic variation on a genome-wide, global scale.

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Strains:
Samples of 23 strains of C. elegans isolated from the wild were obtained from the Caenorhabditis Genetics Center (CGC; see Table 1). These represent most of the natural isolates available through the Center. These strains have been isolated at various times from the 1960s to the present and deposited at the CGC. C. elegans strains may be maintained frozen and stored indefinitely in liquid nitrogen (STIERNAGLE 1999 ). Due to the CGC's practice of freezing strains upon receipt and thawing them when requested, the strains when obtained from the CGC are generally just a few generations away from the original isolate (T. STIERNAGLE, CGC, personal communication). Specific details for each strain may be obtained from the CGC (http://biosci.umn.edu/CGC/Strains/strains.htm). Worms were grown on 6-cm petri plates at 19° on nematode growth medium, seeded with Escherichia coli.

PCR and electrophoresis:
From each strain, 20–25 worms were collected and DNA was extracted by standard methods (JOHNSTONE 1999 ). PCR was performed in 11-µl reaction volumes with 0.38 µl of each primer (1 µM), 1.11 µl of 10x buffer, 0.89 µl MgCl₂ (25 mM), 1.11 µl of dNTPs (2.5 mM), 0.44 µl of IRD M13 primer (1 µM; Licor), 0.25 units of Taq DNA polymerase, and 1 µl of genomic DNA. PCR products were separated by electrophoresis on a 7% acrylamide gel run on a Licor 4200 automated DNA sequencer. Allele sizes were scored using GeneImagIR software (Licor).

Microsatellite loci:
From an exhaustive list of microsatellites in the C. elegans genome (KATTI et al. 2001 ), 20 were chosen at random from all six chromosomes (see Table 2). Using the complete genome sequence of C. elegans (available at ftp://ncbi.nlm.nih.gov/genbank/genomes/C_elegans/), primers were designed to amplify these repeat regions. One of the primers in each pair was end labeled. All the loci consisted of dinucleotide repeats, with the repeat motif occurring from 10 to 32 times in the reference sequence. In addition, one locus (2003) contained a string of G's following the GA repeat. All loci chosen were in noncoding regions of the genome, either in introns or in intergenic regions.

Statistical analyses:
Expected number of alleles under the infinite-alleles model: The expected number of alleles in a sample of n chromosomes is given by k_e = 1 + M/(M + 1) + M/(M + 2) + ... + M/(M + n - 1) (EWENS 1972 ), where M = 4Nu (N is the effective population size; u is the mutation rate per locus). At mutation-drift equilibrium, M can be estimated as M = H/(1 - H) (CROW and KIMURA 1970 ), where H is the heterozygosity. However, when using H estimates from a single locus, this formula is biased and a bias-corrected estimate is obtained by solving the equation M³ + (7 - M₀)M² + (8 - 5M₀)M - 6M₀ = 0 (CHAKRABORTY and WEISS 1991 ), where M₀ = H/(1 - H).

Expected number of alleles under the stepwise mutation model: A bias-corrected estimator of M is obtained by solving 1.7M⁴ + (25 - 1.7M₀)M³ + (24.5 - 13M₀)M² + (9 - 22.5M₀)M - 6M₀ = 0 (ESTOUP et al. 1995 ), where M₀ = 0.5((1 - H)^-2 - 1) (OHTA and KIMURA 1973 ). The expected number of alleles in a sample of n individuals is estimated using Equation 18 of KIMURA and OHTA 1975 .

Tests of population expansion: The parameter = 4Nu can be estimated from the genetic variance V and also from the homozygosity P₀ at a locus. Under a population growth model, the variance-based and homozygosity-based estimates of deviate from each other. This deviation, measured by the imbalance index ß, can be used to detect population expansion (KIMMEL et al. 1998 ; KING et al. 2000 ). ß was estimated in two ways,

and

for a sample of L loci indexed by i (KING et al. 2000 ). _V, the variance-based estimator of , and _P0, the homozygosity estimator of , were computed using Equations 2, 3, and 5 from KING et al. 2000 .

For a growing population, the expected distribution of allele lengths is smoothly peaked, as most bifurcations date back to the time of population expansion. However, in the case of a constant-sized population, most pairs of alleles are either closely related or distantly related since genealogy is expected to include a relatively ancient bifurcation associated with the most recent common ancestor of the entire sample. The expectation of a deep genealogy leads in turn to an expectation of an allele length distribution that is ragged and multipeaked. The within-locus k-test (REICH and GOLDSTEIN 1998 ) uses the shape of the allele-length distribution to differentiate between smooth and ragged distributions of allele lengths and thereby test for population expansion. The test statistic is given by

where Sig⁴ is an estimate of the variance of the allele-length distribution squared, Gam₄ is an estimate of the fourth central moment of the allele-length distribution, and S² is the sample variance of the allele-length distribution (REICH and GOLDSTEIN 1998 ).

The interlocus g-test (REICH and GOLDSTEIN 1998 ) employs the principle that the variance of the variance in allele sizes across loci is expected to be larger in a constant-sized population than in a growing one. This results from ancient bifurcations in a growing population having similar dates at all loci, but having considerable variation among loci in a constant-sized population. The test statistic is , where V_j is the variance at the jth locus, and is the average of all V_j. (Equation 3 of REICH and GOLDSTEIN 1998 ; see also REICH et al. 1999 ).

Maximum-likelihood estimates of population growth rate: BEAUMONT 1999 developed a Bayesian approach for using microsatellite data to assess recent changes in population size. The method yields posterior distributions of population mutation rates for individual loci, and for all loci collectively it provides distributions of the exponential population growth rate r, defined as the ratio of the current population size to that just prior to the period of population size change, and tf, which is the time since the population size began to change, expressed in units of the current population size. Distributions were obtained using the computer program MSVAR (BEAUMONT 1999 ), which conducts Markov chain Monte Carlo simulations. Noninformative, flat prior distributions were used for all parameters. Several chains were run, yielding 79,000 points representing 9 x 10⁷ parameter updates, and the distributions from multiple independent runs were pooled. Convergence was assessed in two ways: first, by comparing posterior distributions for parameters from independent runs with different starting parameter values, and second, by looking at plots of parameter values against time and checking for "stickiness." For each parameter the maximum-likelihood estimate was identified as the peak of the posterior distribution, and the 95% highest posterior density (HPD) credible set was obtained by finding the shortest interval that contained 95% of the posterior probability.

Analyses of population structure: The software package Arlequin V2.000 (SCHNEIDER et al. 2000 ) was used to estimate population parameters. Population substructure was examined using F_ST, which does not explicitly take into account a mutational mechanism, as well as R_ST (SLATKIN 1995 ), which does. Correlation between genetic and geographic distances was examined using a Mantel test with 10,000 permutations, also using Arlequin V2.000.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Patterns of variation at microsatellite loci:
Twenty microsatellite loci were examined in 23 strains of C. elegans. No variation was observed within any of the strains at any of the loci, and each strain was represented by a single allele at each locus, as expected if individuals are highly homozygous due to being cultured by self-fertilization for long periods. The fact that a number of worms from each strain were used to extract DNA is not intended to reflect any within-strain variation that may have existed in the original source populations. Any variation within a strain is unlikely to persist in a single plate after several generations of laboratory culture. Thus, essentially, for each locus, 23 chromosomes were sampled worldwide.

Ten loci were monomorphic among all the strains. The remaining 10 loci were polymorphic, with the number of alleles ranging from 3 to 12 (mean = 6.2). The longer loci tended to have more alleles (r² = 0.43, P < 0.05), although one of the longest loci (X002, 32 repeats) was monomorphic. Alleles at one of the polymorphic loci (X004) could not be scored unambiguously; hence this locus was excluded from subsequent analyses.

The strains TR388 from Wisconsin and N2 from England were identical at all 19 loci. Likewise, strains DH424 (El Prieto Canyon, California) and DR1349 (a subclone of strain PA-1; Pasadena, California) were identical at all loci. It is unlikely that cross-contamination of cultures has occurred since the strains have been maintained at the stock center (T. STIERNAGLE, CGC, personal communication). Although outcrossing rates in the wild are not known, it is possible that C. elegans has a low outcrossing rate and that this coupled with occasional migration of strains over long distances might explain the finding of identical genotypes in disparate locations. One way that such a clonal population structure may be revealed is by the presence of significant linkage disequilibrium between loci on different chromosomes. To test this, we classified alleles for each locus as either common or rare and looked for nonrandom associations of alleles, under a two-locus two-allele model, using Fisher's exact test. Of the 36 pairwise comparisons between loci on different chromosomes, only 2 were significant at the 5% level, and the exact probabilities were distributed evenly over the interval from zero to one. Thus, there is no evidence for the presence of any significant level of linkage disequilibrium among the loci examined.

An alternative explanation for the finding of identical strains in distant locations is very recent migration. While such movement of individuals could be effected by association with, for example, birds, it is also possible that human activity has been a factor facilitating C. elegans dispersal.

Of the nine variable loci, only two (1004 and 3004) displayed variation entirely in multiples of the motif length (2 bp; see Table 3 and Fig 1). The allele sizes suggest that mutation patterns do not conform strictly to the stepwise model of microsatellite mutation. Table 3 shows the observed number of alleles at each variable locus and the expected number of alleles under infinite-allele and stepwise mutation models. These estimates were corrected for bias (ZOUROS 1979 ; CHAKRABORTY and WEISS 1991 ; ESTOUP et al. 1995 ), and the concordance of the observed data with the predictions was tested by the nonparametric Wilcoxon signed ranks test. The observed data are not significantly different from the predictions of the infinite-alleles model (P > 0.05) but are different from the predictions of the stepwise mutation model (P < 0.01). These predictions, however, assume an equilibrium population size, while in fact the recent history of the species might have involved large changes in population size. FRISSE 1999 conducted mutation-accumulation experiments using 100 lines each started from the descendants of a single, inbred, wild-type worm, which were propagated from a single, randomly picked worm each generation to minimize the effects of selection. It was found that for microsatellites in the range of sizes used in this study, most mutations involved the addition or deletion of a single repeat unit. It is possible that these microsatellites are evolving according to the stepwise model, and the lack of concordance in number of alleles results from the demographic history of the populations.

View larger version (25K): In this window In a new window Download PPT slide   Figure 1. Distribution of allele sizes for nine microsatellite loci. The horizontal axis depicts the absolute sizes of PCR products for each locus.

Correlation between genetic and geographic distances:
The geographic origin of one strain (PB303) is not known, so this strain was excluded from the geographic analyses. The strains were assigned to regions (Europe, North America, Australia, and Hawaii) and to populations within regions (in Europe—England, France, and Germany; in North America—northern California, southern California, Vancouver, and Wisconsin; in Australia and Hawaii—one strain each). Partitioning of genetic variation within populations, among populations, and among regions was examined using analysis of molecular variance (AMOVA; EXCOFFIER et al. 1992 ), starting from a matrix of pairwise genetic distances. Using pairwise distances measured as the number of shared alleles, which corresponds to the infinite-alleles model, F_ST = 0.38 is not significant at the 0.05 level. About 62% of the total variance is contained within populations. However, using the sum of squared allele-size differences, a distance measure corresponding to the stepwise mutation model, 54% of the variance is among regions, and R_ST = 0.73 is significant (P < 0.02). Thus there is some evidence of population structure in C. elegans under the assumption of a stepwise model of mutation.

Similarity between pairs of strains was calculated as S = (the number of alleles shared over all loci)/(the number of loci). Pairwise distances were then calculated as (1 - S) (BOWCOCK et al. 1994 ). Geographic distances between the sites ("as the crow flies") were obtained using a web-based program (available at http://www.indo.com/distance/) and a 22 x 22 geographic distance matrix was constructed. No correlation was detected between the geographic and genetic distance matrices (r = 0.02, P = 0.348; Mantel test, 10,000 permutations), using the software package Arlequin V2.000 (SCHNEIDER et al. 2000 ).

Tests of population expansion:
Population expansion can leave a strong signature on allele-size distributions (KIMMEL et al. 1998 ). Populations that have been growing are expected to have distributions that are smooth and have a single peak, whereas populations that have been constant in size or have been shrinking will have more ragged, multipeaked allele-frequency distributions (see METHODS). Fig 2 provides examples of several simulated distributions under a model of constant population size and a model of recent population growth.

View larger version (18K): In this window In a new window Download PPT slide   Figure 2. Effect of population growth on the allele-size distribution at microsatellite loci. Coalescent simulations were performed using the program SIMCOAL (http://cmpg.unibe.ch/software/simcoal/). Each panel shows the distribution obtained from six independent simulation runs of sample size 23, under a given set of parameter values. In all cases, a stepwise mutation model with a mutation rate of 0.00178 per locus per generation and a limit of 20 alleles was used, and the current effective population size was set at 14,000, which is approximately the average estimate across loci for Ne under the stepwise mutation model. (A) Constant population size. (B) Exponential growth with r = 0.005.

Three statistical tests were employed to detect population expansion events from multilocus microsatellite data. None reveals a signature of population expansion in the data. Of the three tests, the imbalance index ß is the most sensitive to population increase. In particular, ß is highly responsive to expansions of recent origin (KING et al. 2000 ). Populations that have recently expanded exhibit ß < 1 (i.e., ln ß < 0). The values of ln(ß) calculated in two ways (ln ß1 and ln ß2, see METHODS) are 3.235 and 2.175, respectively. The within-locus k-test (REICH and GOLDSTEIN 1998 ; REICH et al. 1999 ) is based on the shape of the allele-size distribution. The probability of obtaining a positive value of k is constrained between 0.515 and 0.55. Population expansion is characterized by fewer loci than expected having a positive k value under the null hypothesis of constant population size. A one-tailed binomial test, with the probability of a positive k conservatively set to 0.515, can be used to test for departures from the null hypothesis (REICH and GOLDSTEIN 1998 ). By this method, we find no evidence of population expansion (P value = 0.463). The interlocus g-test (REICH and GOLDSTEIN 1998 ) compares the observed and estimated values of the variance in allele sizes across loci. Again, no sign of population expansion is detected by this test at the 0.05 significance level (table of cutoff values in REICH et al. 1999 ). These tests are based on the assumption of a single-step mutational process. Our data indicate that the loci under study might deviate from this model of mutation. When this assumption is violated due to multistep mutations, the probability of wrongly inferring expansion increases for the within-locus test and decreases for the interlocus test (REICH et al. 1999 ). Since neither of these tests detects an expansion event, we infer that these data reveal no evidence of recent population expansion.

When a model of recent population size change was fit to the data, the maximum-likelihood estimate of the growth rate, r, was 0.0027 (95% HPD region: 0.00096–0.01205). For the time parameter, tf, the peak occurred at 1.117 (95% HPD region: 0.750–1.717). Since r is the ratio of the current to ancestral population sizes, these values correspond to a 500-fold reduction in population size over the past 1.12 x N₀ generations, where N₀ is the current population size.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Despite our extensive knowledge of cellular, developmental, and genetic mechanisms in C. elegans, we do not have detailed knowledge of this species' ecology, population genetics, or evolutionary history (HODGKIN and DONIACH 1997 ; DELATTRE and FELIX 2001 ). In this study we begin to fill that gap on the basis of inferences from patterns of variation at microsatellite loci.

The patterns of genetic variation observed in a worldwide sample of C. elegans are intriguing. In general, widespread species with large population sizes are expected to harbor large amounts of genetic variation. However, in organisms such as C. elegans, which reproduce primarily by selfing, variation within a population may be reduced. High rates of self-fertilization reduce the effective population size by up to a factor of two (POLLAK 1987 ). Also, neutral loci may be expected to show reduced variability due to recurrent selection against deleterious mutations at linked loci. This background selection effect is predicted to be especially strong in highly or completely selfing species, or asexually reproducing species, as compared to outcrossing species, since recombination is effectively reduced (NORDBORG et al. 1996 ; CHARLESWORTH et al. 1997 ; NORDBORG 1997 ). Alternatively, selective sweeps leading to the repeated fixation of favorable alleles might lead to the reduction in diversity at neutral loci (MAYNARD SMITH and HAIGH 1974 ; KAPLAN et al. 1989 ). This expectation is supported by the finding that the largely self-fertilizing species C. elegans and C. briggsae harbor very low levels of nucleotide polymorphism compared to the gonochoristic C. remanei, another member of the elegans group of the genus Caenorhabditis (GRAUSTEIN et al. 2002 ). Thus a high rate of inbreeding, coupled with selection, might serve to reduce effective sizes of C. elegans populations. Using observed mutation rates for microsatellite loci (FRISSE 1999 ), estimates of the effective population size from our data range from 200 to 44,000 (Table 3). Although they are found worldwide, actual densities of C. elegans in nature have not yet been estimated. However, nematodes are generally very abundant, often attaining densities of millions of individuals per square meter (BLAXTER 1998 ). If C. elegans individuals are indeed numerous, then the polymorphism data suggest the presence of a powerful selective force acting to maintain a low level of genetic variation.

The factors that are expected to reduce variation within populations may be expected to increase variation among populations. In the absence of high levels of gene flow, the effects of local inbreeding and local selective sweeps will remove variation within populations but will not remove variation between populations. However, we find that levels of variation on a global scale are very low. Surveys of microsatellites in various species typically report polymorphism at a large proportion of the loci studied (see Table 4 for some examples). The proportion of polymorphic loci is uncommonly low in C. elegans. Our results are consistent with those from other studies of interstrain variation that have utilized various classes of genetic markers. Early studies focusing on a limited number of isolates have found low levels of variation for enzymes (BUTLER et al. 1981 ), minisatellite and simple sequence repeat motifs (UITTERLINDEN et al. 1989 ), and Tc1 transposon patterns (EGILMEZ et al. 1995 ). More recently, DNA sequence polymorphism has been found to be very low for both nuclear and mitochondrial loci (KOCH et al. 2000 ; GRAUSTEIN et al. 2002 ). This genome-wide lack of variation might underlie the absence of strong heterosis for life history traits in crosses between strains of C. elegans (JOHNSON and HUTCHINSON 1993 ).

The population genetics portrait that begins to emerge from this and previous studies is of a widespread species that exhibits little variation, either within geographic regions or on a global scale. It is unlikely that this picture has been caused by a biased or small sample of loci, as our sample consists of randomly selected microsatellites located on all the chromosomes of C. elegans. Alternatively, a low mutation rate for C. elegans microsatellites might account for the low level of polymorphism. However, the mutation rate at microsatellite loci observed from mutation-accumulation experiments ranges from 8.93 x 10^-5 to 1.85 x 10^-2 (FRISSE 1999 ) for a variety of repeat sizes and motifs. This mutation rate is not unusually low compared to other organisms (e.g., CRAWFORD and CUTHBERTSON 1996 ; GOLDSTEIN and SCHLOTTERER 1999 , p.6 and references therein; VAZQUEZ et al. 2000 ; THUILLET et al. 2002 ; YUE et al. 2002 ). In this light it is noteworthy that the microsatellite mutation rates in Drosophila are very low (VAZQUEZ et al. 2000 ), yet far more polymorphism is observed in this species (Table 4).

A. thaliana may be expected to resemble C. elegans in some respects, and in recent years it has been the subject of many population genetic studies (TODOKORO et al. 1995 ; INNAN et al. 1997 ; PURUGGANAN and SUDDITH 1998 , PURUGGANAN and SUDDITH 1999 ; MALOOF et al. 2000 , MALOOF et al. 2001 ). Like C. elegans, A. thaliana reproduces mainly by self-fertilization, with an estimated selfing rate of >99% (BERGELSON et al. 1998 ), and it has a widespread distribution. Findings in A. thaliana, however, are conflicting. On the one hand, studies using microsatellite loci have found no variation within an ecotype, with all individuals homozygous at all loci, but large amounts of variation among ecotypes (TODOKORO et al. 1995 ; INNAN et al. 1997 ). The picture is one of population structure and possibly recent expansion of the species throughout the world. On the other hand, BERGELSON et al. 1998 report low levels of both intra- and interpopulation divergence using RFLP markers and no association between geographic and genetic distance. These observations might reflect processes specific to the different classes of genetic markers used in these studies. In C. elegans, however, our study and others have observed low levels of variation for a range of different markers.

A possible explanation to account for the low levels of variation in our data and those from other studies is that C. elegans has only recently spread throughout the world. If this has been the case then we should see evidence of recent population growth in the allelic distributions of microsatellite loci. We conducted several tests that were designed to detect recent population expansion on the basis of microsatellite data and that are capable of detecting recent expansion events with a high probability (KING et al. 2000 ). None of the tests revealed a signature of such an event in the recent history of the organism. We also fit the data to a model of recent exponential size change (BEAUMONT 1999 ). The resulting maximum-likelihood estimates suggest that the global population of C. elegans has actually been shrinking and are consistent with a recent 500-fold reduction in size. The method assumes that mutation occurs under the stepwise mutation model (SMM; BEAUMONT 1999 ), and so the fact that the data do not correspond to this model may partly explain why the allele-frequency distributions depart so strongly from that expected under population growth. However, the tests are not all similarly sensitive to this assumption (KING et al. 2000 ). On balance, it seems unlikely that present C. elegans genetic diversity is the result of recent colonization and expansion events.

The patterns of variation can also be considered in light of models of natural selection. If recombination is low, due to low rate of outcrossing, then both selective sweeps and background selection can greatly reduce levels of genetic variation at neutral loci. However, two points argue against selective sweeps as the main cause of low polymorphism. First, selective sweeps that are associated with local adaptation will remove variation only over the local geographic area and will effectively promote population differentiation at the expense of more widespread polymorphisms. However, C. elegans reveals low levels of polymorphism globally and only moderate amounts of population structure. Second, a recent selective sweep exerts a very similar effect on patterns of variation at linked neutral loci as does a recent population expansion. In effect, a selective sweep causes a severe bottleneck for linked loci. However, none of the analyses of recent population history found evidence for recent population growth, as would be expected if much of the genome was linked to regions that had recently experienced selective sweeps on a global scale.

C. elegans presents us with an interesting paradox. The species occurs almost worldwide in areas with widely different climates and environments. Yet only a low level of genetic variation exists on a global scale. There appears to be some population structure, but this has not led to appreciable divergence between geographic regions. We also find that the allele-frequency distributions are not as expected if C. elegans had recently spread throughout the world or if there had been recent selective sweeps that reduced variation. It is possible that background selection (CHARLESWORTH et al. 1993 ), greatly enhanced in effect due to very low levels of recombination, is the major cause of low levels of variation.

One possible factor that could explain the low level of variation worldwide and the absence of evidence of population expansion is ongoing gene flow over long distances among populations of C. elegans worldwide. During periods of environmental stress, L2 larvae of C. elegans enter a state of diapause, the dauer larva. During this facultative life stage, larvae display behaviors not seen in other stages. The dauer larva can survive long periods without feeding and resist desiccation. The dauer stage may be an adaptation not only for survival during periods of stress, but also for dispersal by birds, insects, or other animals (RIDDLE 1988 ; HODGKIN and DONIACH 1997 ). If we assume an equilibrium migration model then the F_ST estimate and the relationship, F_ST = 1/(1 + 4N_em) (WRIGHT 1951 ), where m is the migration rate among populations, lead to an estimate of N_em of 0.41 individuals migrating among populations per generation. Taken together, the data and analyses provide indirect evidence for movement over long distances of C. elegans individuals among populations. This movement may be association with human activity; however, it does not appear to have been solely a recent phenomenon, as we find no evidence for population expansion. Another interpretation of the pattern observed here is that C. elegans occurs in very low numbers in nature. Given the rapid growth rate observed in laboratory cultures, the ability to go into the dauer state in unfavorable environments, and the general observation that nematodes usually attain very high densities, this seems an unlikely explanation. However, the gaps in our knowledge of the ecology of this species preclude us from entirely ruling out this possibility.

	ACKNOWLEDGMENTS

We thank Christian Schlotterer, Garth Patterson, Chi-hua Chiu, Carlos Machado, Andrew Singson, and two anonymous reviewers for useful discussions and comments on earlier versions of the manuscript. All the C. elegans strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health National Center for Research Resources.

Manuscript received June 26, 2002; Accepted for publication October 17, 2002.

	LITERATURE CITED

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

ABDUL KADER, N. and M. G. COTE, 1996  Isolation, identification and characterization of some strains of Caenorhabditis elegans (Maupas, 1900) from Quebec. Fundam. Appl. Nematol. 19:381-389.

BEAUMONT, M. A., 1999  Detecting population expansion and decline using microsatellites. Genetics 153:2013-2029.[Abstract/Free Full Text]

BELL, C. J. and J. R. ECKER, 1994  Assignment of 30 microsatellite loci to the linkage map of Arabidopsis. Genomics 19:137-144.[Medline]

BERGELSON, J., E. STAHL, S. DUDEK, and M. KREITMAN, 1998  Genetic variation within and among populations of Arabidopsis thaliana.. Genetics 148:1311-1323.[Abstract/Free Full Text]

BLAXTER, M., 1998  Caenorhabditis elegans is a nematode. Science 282:2041-2046.[Abstract/Free Full Text]

BOWCOCK, A. M., A. RUIZ-LINARES, J. TOMFOHRDE, E. MINCH, and J. R. KIDD et al., 1994  High resolution of human evolutionary trees with polymorphic microsatellites. Nature 368:455-457.[Medline]

BUTLER, M. H., S. M. WALL, K. R. LUEHRSEN, G. E. FOX, and R. M. HECHT, 1981  Molecular relationships between closely related strains and species of nematodes. J. Mol. Evol. 18:18-23.[Medline]

CHAKRABORTY, R. and K. M. WEISS, 1991  Genetic variation of the mitochondrial DNA genome in American Indians is at mutation-drift equilibrium. Am. J. Phys. Anthropol. 86:497-506.[Medline]

CHARLESWORTH, B., M. T. MORGAN, and D. CHARLESWORTH, 1993  The effect of deleterious mutations on neutral molecular variation. Genetics 134:1289-1303.[Abstract]

CHARLESWORTH, B., M. NORDBORG, and D. CHARLESWORTH, 1997  The effects of local selection, balanced polymorphism and background selection on equilibrium patterns of genetic diversity in subdivided populations. Genet. Res. 70:155-174.[Medline]

CRAWFORD, A. M. and R. P. CUTHBERTSON, 1996  Mutations in sheep microsatellites. Genome Res. 6:876-879.[Abstract/Free Full Text]

CROW, J. F., and M. KIMURA, 1970 An Introduction To Population Genetics Theory. Harper & Row, New York.

DELATTRE, M. and M. A. FELIX, 2001  Microevolutionary studies in nematodes: a beginning. Bioessays 23:807-819.[Medline]

DIETRICH, W. F., J. MILLER, R. STEEN, M. A. MERCHANT, and D. DAMRON-BOLES et al., 1996  A comprehensive genetic map of the mouse genome. Nature 380:149-152.[Medline]

EGILMEZ, N. K., R. H. EBERT, and R. J. SHMOOKLER REIS, 1995  Strain evolution in Caenorhabditis elegans: transposable elements as markers of interstrain evolutionary history. J. Mol. Evol. 40:372-381.[Medline]

ENGLAND, P. R., D. A. BRISCOE, and R. FRANKHAM, 1996  Microsatellite polymorphisms in a wild population of Drosophila melanogaster. Genet. Res. 67:285-290.[Medline]

ESTOUP, A., L. GARNERY, M. SOLIGNAC, and J. M. CORNUET, 1995  Microsatellite variation in honey bee (Apis mellifera L.) populations: hierarchical genetic structure and test of the infinite allele and stepwise mutation models. Genetics 140:679-695.[Abstract]

EWENS, W. J., 1972  The sampling theory of selectively neutral alleles. Theor. Popul. Biol. 3:87-112.[Medline]

EXCOFFIER, L., P. E. SMOUSE, and J. M. QUATTRO, 1992  Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131:479-491.[Abstract]

FIELD, D. and C. WILLS, 1998  Abundant microsatellite polymorphism in Saccharomyces cerevisiae, and the different distributions of microsatellites in eight prokaryotes and S. cerevisiae, result from strong mutation pressures and a variety of selective forces. Proc. Natl. Acad. Sci. USA 95:1647-1652.[Abstract/Free Full Text]

FRISSE, L. M., 1999 Understanding the mechanisms of microsatellite formation and mutation using the model organism Caenorhabditis elegans. Ph.D. Dissertation, University of Missouri, Kansas City, MO.

GOLDSTEIN, D. B. and A. G. CLARK, 1995  Microsatellite variation in North American populations of Drosophila melanogaster. Nucleic Acids Res. 23:3882-3886.[Abstract/Free Full Text]

GOLDSTEIN, D. B., and C. SCHLOTTERER, (Editors), 1999 Microsatellites: Evolution and Applications. Oxford University Press, Oxford.

GRANDIEN, K. and R. J. SOMMER, 2001  Functional comparison of the nematode Hox gene lin-39 in C. elegans and P. pacificus reveals evolutionary conservation of protein function despite divergence of primary sequences. Genes Dev. 15:2161-2172.[Abstract/Free Full Text]

GRAUSTEIN, A., J. M. GASPAR, J. R. WALTERS, and M. F. PALOPOLI, 2002  Levels of DNA polymorphism vary with mating system in the nematode genus Caenorhabditis. Genetics 161:99-107.[Abstract/Free Full Text]

HENNEQUIN, C., A. THIERRY, G. F. RICHARD, G. LECOINTRE, and H. V. NGUYEN et al., 2001  Microsatellite typing as a new tool for identification of Saccharomyces cerevisiae strains. J. Clin. Microbiol. 39:551-559.[Abstract/Free Full Text]

HODGKIN, J. and T. DONIACH, 1997  Natural variation and copulatory plug formation in Caenorhabditis elegans. Genetics 146:149-164.[Abstract]

HUDSON, R. R., M. KREITMAN, and M. AGUADÉ, 1987  A test of neutral molecular evolution based on nucleotide data. Genetics 116:153-159.[Abstract/Free Full Text]

INNAN, H., R. TERAUCHI, and N. T. MIYASHITA, 1997  Microsatellite polymorphism in natural populations of the wild plant Arabidopsis thaliana. Genetics 146:1441-1452.[Abstract]

JOHNSON, T. E. and E. W. HUTCHINSON, 1993  Absence of strong heterosis for life span and other life history traits in Caenorhabditis elegans. Genetics 134:465-474.[Abstract]

JOHNSTONE, I. L., 1999 Molecular biology, pp. 201–225 in C. elegans—A Practical Approach, edited by I. A. HOPE. Oxford University Press, London.

KAPLAN, N., R. R. HUDSON, and C. H. LANGLEY, 1989  The "hitchhiking effect" revisited. Genetics 123:887-899.[Abstract/Free Full Text]

KATTI, M. V., P. K. RANJEKAR, and V. S. GUPTA, 2001  Differential distribution of simple sequence repeats in eukaryotic genome sequences. Mol. Biol. Evol. 18:1161-1167.[Abstract/Free Full Text]

KAUER, M., B. ZANGERL, D. DIERINGER, and C. SCHLOTTERER, 2002  Chromosomal patterns of microsatellite variability contrast sharply in African and non-African populations of Drosophila melanogaster.. Genetics 160:247-256.[Abstract/Free Full Text]

KENT, W. J. and A. M. ZAHLER, 2000  Conservation, regulation, synteny, and introns in a large-scale C. briggsae-C. elegans genomic alignment. Genome Res. 10:1115-1125.[Abstract/Free Full Text]

KIKUCHI, S. and Y. ISAGI, 2002  Microsatellite genetic variation in small and isolated populations of Magnolia sieboldii ssp. japonica. Heredity 88:313-321.[Medline]

KIMMEL, M., R. CHAKRABORTY, J. P. KING, M. BAMSHAD, and W. S. WATKINS et al., 1998  Signatures of population expansion in microsatellite repeat data. Genetics 148:1921-1930.[Abstract/Free Full Text]

KIMURA, M. and T. OHTA, 1975  Distribution of allelic frequencies in a finite population under stepwise production of neutral alleles. Proc. Natl. Acad. Sci. USA 72:2761-2764.[Abstract/Free Full Text]

KING, J. P., M. KIMMEL, and R. CHAKRABORTY, 2000  A power analysis of microsatellite-based statistics for inferring past population growth. Mol. Biol. Evol. 17:1859-1868.[Abstract/Free Full Text]

KLIMAN, R. M., P. ANDOLFATTO, J. A. COYNE, F. DEPAULIS, and M. KREITMAN et al., 2000  The population genetics of the origin and divergence of the Drosophila simulans complex species. Genetics 156:1913-1931.[Abstract/Free Full Text]

KOCH, R., H. G. VAN LUENEN, M. VAN DER HORST, K. L. THIJSSEN, and R. H. PLASTERK, 2000  Single nucleotide polymorphisms in wild isolates of Caenorhabditis elegans.. Genome Res. 10:1690-1696.[Abstract/Free Full Text]

KRIEGER, M. J. B. and L. KELLER, 1999  Low polymorphism at 19 microsatellite loci in a French population of Argentine ants (Linepithema humile). Mol. Ecol. 8:1078-1080.

LANZARO, G. C., L. ZHENG, Y. T. TOURE, S. F. TRAORE, and F. C. KAFATOS et al., 1995  Microsatellite DNA and isozyme variability in a west African population of Anopheles gambiae.. Insect Mol. Biol. 4:105-112.[Medline]

MACKAY, T. F., 2001  Quantitative trait loci in Drosophila. Nat. Rev. Genet. 2:11-20.[Medline]

MALOOF, J. N., J. O. BOREVITZ, D. WEIGEL, and J. CHORY, 2000  Natural variation in phytochrome signaling. Semin. Cell Dev. Biol. 11:523-530.[Medline]

MALOOF, J. N., J. O. BOREVITZ, T. DABI, J. LUTES, and R. B. NEHRING et al., 2001  Natural variation in light sensitivity of Arabidopsis. Nat. Genet. 29:441-446.[Medline]

MAYNARD SMITH, J. and J. HAIGH, 1974  The hitch-hiking effect of a favourable gene. Genet. Res. 23:23-35.[Medline]

MCALLISTER, B. F. and G. A. MCVEAN, 2000  Neutral evolution of the sex-determining gene transformer in Drosophila. Genetics 154:1711-1720.[Abstract/Free Full Text]

MCDONALD, J. H. and M. KREITMAN, 1991  Adaptive protein evolution at the Adh Locus in Drosophila. Nature 351:652-654.[Medline]

MORIYAMA, E. N. and J. R. POWELL, 1996  Intraspecific nuclear DNA variation in Drosophila. Mol. Biol. Evol. 13:261-277.[Abstract]

NOOR, M. A., M. D. SCHUG, and C. F. AQUADRO, 2000  Microsatellite variation in populations of Drosophila pseudoobscura and Drosophila persimilis.. Genet. Res. 75:25-35.[Medline]

NORDBORG, M., 1997  Structured coalescent processes on different time scales. Genetics 146:1501-1514.[Abstract]

NORDBORG, M., B. CHARLESWORTH, and D. CHARLESWORTH, 1996  Increased levels of polymorphism surrounding selectively maintained sites in highly selfing species. Proc. Biol. Sci. 263:1033-1039.[Abstract/Free Full Text]

NORDBORG, M., J. O. BOREVITZ, J. BERGELSON, C. C. BERRY, and J. CHORY et al., 2002  The extent of linkage disequilibrium in Arabidopsis thaliana.. Nat. Genet. 30:190-193.[Medline]

OHTA, T. and M. KIMURA, 1973  A model of mutation appropriate to estimate the number of electrophoretically detectable alleles in a finite population. Genet. Res. 22:201-204.[Medline]

PASCUAL, M., M. D. SCHUG, and C. F. AQUADRO, 2000  High density of long dinucleotide microsatellites in Drosophila subobscura. Mol. Biol. Evol. 17:1259-1267.[Abstract/Free Full Text]

POLLAK, E., 1987  On the theory of partially inbreeding finite populations. I. Partial selfing. Genetics 117:353-360.[Abstract/Free Full Text]

POWELL, J. R. and E. N. MORIYAMA, 1997  Evolution of codon usage bias in Drosophila. Proc. Natl. Acad. Sci. USA 94:7784-7790.[Abstract/Free Full Text]

PURUGGANAN, M. D. and J. I. SUDDITH, 1998  Molecular population genetics of the Arabidopsis CAULIFLOWER regulatory gene: nonneutral evolution and naturally occurring variation in floral homeotic function. Proc. Natl. Acad. Sci. USA 95:8130-8134.[Abstract/Free Full Text]

PURUGGANAN, M. D. and J. I. SUDDITH, 1999  Molecular population genetics of floral homeotic loci: departures from the equilibrium-neutral model at the APETALA3 and PISTILLATA genes of Arabidopsis thaliana. Genetics 151:839-848.[Abstract/Free Full Text]

REED, K. M. and C. W. BEATTIE, 2001  Isolation of 105 microsatellite loci from an ovine genomic library enriched for microsatellites. Anim. Biotechnol. 12:77-86.[Medline]

REICH, D. E. and D. B. GOLDSTEIN, 1998  Genetic evidence for a Paleolithic human population expansion in Africa. Proc. Natl. Acad. Sci. USA 95:8119-8123.[Abstract/Free Full Text]

REICH, D., M. FELDMAN, and D. GOLDSTEIN, 1999  Statistical properties of two tests that use multilocus data sets to detect population expansions. Mol. Biol. Evol. 16:453-466.

RIDDLE, D. L., 1988 The dauer larva, pp. 393–412 in The Nematode Caenorhabditis elegans, edited by W. WOOD. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SCHLOTTERER, C. and B. HARR, 2000  Drosophila virilis has long and highly polymorphic microsatellites. Mol. Biol. Evol. 17:1641-1646.[Abstract/Free Full Text]

SCHMIDT, P. S., D. D. DUVERNELL, and W. F. EANES, 2000  Adaptive evolution of a candidate gene for aging in Drosophila. Proc. Natl. Acad. Sci. USA 97:10861-10865.[Abstract/Free Full Text]

SCHNEIDER, S., D. ROESSLI and L. EXCOFFIER, 2000 Arlequin: A Software for Population Genetics Data Analysis. Department of Anthropology, University of Geneva, Geneva.

SLATKIN, M., 1995  A measure of population subdivision based on microsatellite allele frequencies. Genetics 139:457-462.[Medline]

STIERNAGLE, T., 1999 Maintenance of C. elegans, pp. 51–68 in C. elegans—A Practical Approach, edited by I. A. HOPE. Oxford University Press, London.

TARAMINO, G. and S. TINGEY, 1996  Simple sequence repeats for germplasm analysis and mapping in maize. Genome 39:277-287.[Medline]

THUILLET, A.-C., D. BRU, J. DAVID, P. ROUMET, and S. SANTONI et al., 2002  Direct estimation of mutation rate for 10 microsatellite loci in durum wheat, Triticum turgidum (L.) Thell. ssp durum desf. Mol. Biol. Evol. 19:122-125.[Free Full Text]

TODOKORO, S., R. TERAUCHI, and S. KAWANO, 1995  Microsatellite polymorphisms in natural-populations of Arabidopsis thaliana in Japan. Jpn. J. Genet. 70:543-554.

UITTERLINDEN, A. G., P. E. SLAGBOOM, T. E. JOHNSON, and J. VIJG, 1989  The Caenorhabditis elegans genome contains monomorphic minisatellites and simple sequences. Nucleic Acids Res. 17:9527-9530.[Abstract/Free Full Text]

VAZQUEZ, J. F., T. PEREZ, J. ALBORNOZ, and A. DOMINGUEZ, 2000  Estimation of microsatellite mutation rates in Drosophila melanogaster.. Genet. Res. 76:323-326.[Medline]

WEISSENBACH, J., G. GYAPAY, C. DIB, A. VIGNAL, and J. MORISSETTE et al., 1992  A second-generation linkage map of the human genome. Nature 359:794-801.[Medline]

WRIGHT, S., 1951  The genetical structure of populations. Ann. Eugen. 15:323-354.

YUE, G. H., P. BEECKMANN, and H. GELDERMANN, 2002  Mutation rate at swine microsatellite loci. Genetica 114:113-119.[Medline]

ZOUROS, E., 1979  Mutation rates, population sizes and amounts of electrophoretic variation of enzyme loci in natural populations. Genetics 92:623-646.[Abstract/Free Full Text]

This article has been cited by other articles:

				R. Jovelin, J. P. Dunham, F. S. Sung, and P. C. Phillips High Nucleotide Divergence in Developmental Regulatory Genes Contrasts With the Structural Elements of Olfactory Pathways in Caenorhabditis Genetics, April 1, 2009; 181(4): 1387 - 1397. [Abstract] [Full Text] [PDF]

				N. Phillips, M. Salomon, A. Custer, D. Ostrow, and C. F. Baer Spontaneous Mutational and Standing Genetic (Co)variation at Dinucleotide Microsatellites in Caenorhabditis briggsae and Caenorhabditis elegans Mol. Biol. Evol., March 1, 2009; 26(3): 659 - 669. [Abstract] [Full Text] [PDF]

				J. H. Thomas Genome evolution in Caenorhabditis Brief Funct Genomic Proteomic, June 23, 2008; (2008) eln022v1. [Abstract] [Full Text] [PDF]

				A. D. Cutter, J. D. Wasmuth, and N. L. Washington Patterns of Molecular Evolution in Caenorhabditis Preclude Ancient Origins of Selfing Genetics, April 1, 2008; 178(4): 2093 - 2104. [Abstract] [Full Text] [PDF]

				A. D. Cutter Multilocus Patterns of Polymorphism and Selection Across the X Chromosome of Caenorhabditis remanei Genetics, March 1, 2008; 178(3): 1661 - 1672. [Abstract] [Full Text] [PDF]

				H. S. Seidel, M. V. Rockman, and L. Kruglyak Widespread Genetic Incompatibility in C. Elegans Maintained by Balancing Selection Science, February 1, 2008; 319(5863): 589 - 594. [Abstract] [Full Text] [PDF]

				A. D. Cutter, M.-A. Felix, A. Barriere, and D. Charlesworth Patterns of Nucleotide Polymorphism Distinguish Temperate and Tropical Wild Isolates of Caenorhabditis briggsae Genetics, August 1, 2006; 173(4): 2021 - 2031. [Abstract] [Full Text] [PDF]

				C. Rydholm, G. Szakacs, and F. Lutzoni Low Genetic Variation and No Detectable Population Structure in Aspergillus fumigatus Compared to Closely Related Neosartorya Species. Eukaryot. Cell, April 1, 2006; 5(4): 650 - 657. [Abstract] [Full Text] [PDF]

				A. D. Cutter Nucleotide Polymorphism and Linkage Disequilibrium in Wild Populations of the Partial Selfer Caenorhabditis elegans Genetics, January 1, 2006; 172(1): 171 - 184. [Abstract] [Full Text] [PDF]

				M. Haber, M. Schungel, A. Putz, S. Muller, B. Hasert, and H. Schulenburg Evolutionary History of Caenorhabditis elegans Inferred from Microsatellites: Evidence for Spatial and Temporal Genetic Differentiation and the Occurrence of Outbreeding Mol. Biol. Evol., January 1, 2005; 22(1): 160 - 173. [Abstract] [Full Text] [PDF]