Non-African Origin of a Local Beneficial Mutation in D. melanogaster
http://www.100md.com
《分子生物学进展》
Institut für Tierzucht und Genetik, Wien, Austria
Correspondence: E-mail: christian.schloetterer@vu-wien.ac.at.
Abstract
It is well understood that the out-of-Africa habitat expansion of D. melanogaster was associated with the fixation of many beneficial mutations. Nevertheless, it is not clear yet whether these beneficial mutations segregated already in Africa or originated outside of Africa. In this article, we describe an ongoing selective sweep specific to one European population. One microsatellite allele has increased in a population from The Netherlands to a frequency of 18%, whereas it is virtually absent in 12 other European populations. The selective sweep resulted in a genomic region of more than 600 kb that is identical by descent. This is probably the first evidence of a beneficial mutation that has arisen outside of Africa and has resulted in a selective sweep localized in a population from The Netherlands.
Key Words: local adaptation ? allele frequency ? microsatellite ? partial selective sweep
Introduction
Hitchhiking mapping is an approach for the identification of genomic regions that were recently subjected to positive directional selection (reviewed in Schl?tterer [2003]). By scanning a large number of markers distributed over the genome, candidate regions could be identified without the need of experimental crosses or information about the selected trait. Hitchhiking mapping has already been applied successfully to a wide variety of species ranging from humans to maize (e.g., Payseur, Cutter, and Nachman 2002; Vigouroux et al. 2002; Kauer, Dieringer, and Schl?tterer 2003).
In Drosophila, hitchhiking mapping studies have mainly focused on the question of which genomic regions were affected by positive selection associated with the out-of-Africa habitat expansion (Harr, Kauer, and Schl?tterer 2002; Glinka et al. 2003; Kauer, Dieringer, and Schl?tterer 2003; Sch?fl and Schl?tterer 2004). On the X chromosome, approximately 10% to 20% of the surveyed markers were found to deviate from neutral expectations (Glinka et al. 2003; Kauer, Dieringer, and Schl?tterer 2003; Sch?fl and Schl?tterer 2004). Interestingly, when multiple European populations were analyzed, the same loci carried the signature of a selective sweep (Harr, Kauer, and Schl?tterer 2002; Kauer, Dieringer, and Schl?tterer 2003). This raises the question of whether the beneficial mutation originated outside of Africa or whether it was already segregating in Africa at a low frequency (Orr and Betancourt 2001). Until now, no experimental data are available to distinguish these two scenarios.
A recent survey of multiple D. melanogaster populations from North-America and Europe indicated low, but statistically significant, differences between and within continents (Caracristi and Schl?tterer 2003). Thus, the presence of population substructure in combination with a broad range of different habitats covered by non-African D. melanogaster populations raises the following question: Does local adaptation occur in non-African D. melanogaster populations? While a wealth of comparative data analyses between African and non-African populations is available, very limited data exist about selective sweeps specific to local non-African populations (Schl?tterer, Vogl, and Tautz 1997).
In this study, a pronounced differentiation among non-African (European) populations is described for an X-linked locus. At this locus, one microsatellite allele was present at a relatively high frequency in flies originating from the island of Texel (The Netherlands) but virtually absent in the remainder of the European populations investigated. The absence of molecular variability linked to this allele strongly suggests that its frequency has risen because of a local event of positive selection.
Materials and Methods
Microsatellites and Populations Samples
An X-linked microsatellite locus (XTF0) was studied in 19 (13 Europeans and six non-Europeans) D. melanogaster populations (number of lines): Naples, Italy (30*), Rome, Italy (40*), Neumarkt, Germany (30*), Katowice, Poland (16*), Texel, The Netherlands (66*), Gotheron, France (30*), Friedrichshafen, Germany (40*), Copenhagen, Denmark (30*), Sofia, Bulgaria (24*), Chaves, Portugal (18), Kreta, Greece (12*), Moscow, Russia (30*), Helsinki, Finland (14*), Sengwa Wildlife Preserve, Zimbabwe (18), Harare, Zimbabwe (28), Lake Kariba, Zimbabwe (12). Jiamusi, China (30*), La Milpa, Belize (20*), New Delhi, India (12). The asterisk (*) indicates F1 flies, which are the offspring of freshly collected flies. Other lines are inbred isofemale lines; thus, for genotyping, we randomly selected one allele (Dieringer and Schl?tterer 2003).
Further, 14 microsatellite loci were typed in lines originating from Texel, The Netherlands (28 lines) at both sides of locus XTF0. These loci span a distance of approximately 970 kb (for absolute positions, see table 1 in Supplementary Material online).
Genotyping
DNA extraction from single individuals was carried out according the high-salt extraction procedure described in Miller, Dykes, and Polesky. (1988). Genotyping of microsatellite loci followed standard protocols (Schl?tterer 1998). PCR products were separated on a 7% denaturing polyacrylamid gel (32% formamide, 5.6 M urea) and visualized by autoradiography. PCR products of known size and a slippage ladder were used to determine the allele sizes (Schl?tterer and Zangerl 1999).
DNA Sequencing
At least twelve lines collected from Texel and one line from Friedrichshafen were sequenced at three noncoding DNA fragments spanning approximately 107 kb (table 2 in Supplementary Material online). The average length of the fragments was 642 bp (see table 1). Genomic DNA was extracted from single male flies using a high-salt extraction method (see Genotyping). Each fragment was sequenced in both directions. Primers, used either for PCR amplification or for sequencing reaction, were designed on the basis of the release 3.1 of the complete D. melanogaster genomic sequence (table 2 in Supplementary Material online). Sequencing reactions were performed with the BigDye Terminator version 1.1 cycle sequencing chemistry. The sequencing products were purified with Sephadex G-50 fine (Amersham Biosciences, Sweden), and a MegaBACE automated sequencer was used for separation. Sequences were aligned by ClustalX, manually adjusted, and compared with the GenBank sequence. Sequences were deposited in GenBank under accession numbers AY689233 to AY689273.
Table 1 Summary Statistics of the Sequenced DNA Fragments
Data Analysis
Expected heterozygosity, allele-frequencies, and FST values were estimated using the software MSA version 3.14 (Dieringer and Schl?tterer 2003). Significance of observed FST values was obtained by permuting genotypes 10,000 times among groups.
Sequence diversity was estimated using either the number of segregating sites, w (Watterson 1975), or the average number of pairwise sequence differences in the sample, (Tajima 1983). To test whether the observed allele frequency spectrum was in accordance with the expectations from the neutral model of evolution, Tajima's D statistic (Tajima 1989) was calculated. All computations were carried out using the DnaSP version 4.00 software package (Rozas et al. 2003). Gaps in the sequence alignments were excluded from the analyses. Linkage disequilibrium analyses were also performed with DnaSP 4.00.
Because the 191-bp allele was overrepresented in our sample of sequenced chromosomes, we reconstructed a population sample by discarding all but one (reconstruct 1) or two (reconstruct 2) chromosomes with the 191-bp allele. Using this strategy, the frequency of the 191-bp chromosomes was similar to a randomly drawn sample from the Texel population.
Estimate of Relatedness
The degree of relatedness among lines from Texel was estimated to test whether the individuals bearing the 191-bp allele were more related to each other than to the remainder of the population. Five unlinked microsatellite loci on the third chromosome (table 1 in Supplementary Material online) were genotyped in a total of 29 individuals (eight bearing the 191-bp allele). The relatedness was estimated using both the FST parameter (between the two groups of individuals bearing/not bearing the 191-bp allele) and the BAPS version 2.0 software (Corander, Waldmann, and Sillanpaa 2003). The latter software uses a MCMC algorithm to estimate posterior probabilities for all the possible clusters produced by the individuals surveyed. The following settings were used in our analysis: burn-in of 10,000, a chain length of 10,000, and a stochastic initialization (no initial cluster partition).
Time of the Selective Event
We determined the age of the genomic region IBD (Identical by Descent) using two approaches based on (1) the detected number of lines bearing the 191-bp allele and (2) the frequency of the 191-bp allele in the population sample.
In the first approach, we calculated the probability that a stretch of DNA experiences recombination (Pr), as Pr = (1 – (1 – r)L)(1 – f191) rL (1 – f191), where r is the recombination rate in the region of interest, L is the length of the stretch of DNA, and f191 is the observed frequency of the 191-bp allele. Then, given the probability of recombination (Pr) and the number of IBD haplotypes (n), Er(= nPr) is the expected number of recombination events/generation. We inferred the number of generations, k, by using the largest integer value of k that satisfies the equation Erk < 1.
The second approach was outlined previously (Myant et al. 1997). We calculated the probability Pnr that the IBD region associated with the 191-bp allele does not recombine with another genomic region not IBD co-occurring in the population: Pnr = 1 – Pr. By solving the equation (Pnr)k = f191, we inferred k, the number of generations elapsed between the date of collection and the timepoint at which the selected variant has arisen on the ancestral chromosome.
Because recombination could only occur in females, we adjusted the estimates from both approaches by multiplying by 3/2.
Selection Coefficient
Assuming semidominance, we calculated the selection coefficient s using the equation
The frequency of the 191-bp allele in the European D. melanogaster populations (excluding Texel) was used as an estimate for the frequency at the onset of selection (p0 = 0.0032) and the observed frequency of the same allele in the Texel population as the current frequency (pt = 0.18), q = 1–p. The time t from the onset of selection to the present was conservatively estimated by the approach of Myant et al. (1997), as this yielded a longer time since the occurrence of the beneficial mutation (see Results).
Probability of Neutral Drift
Random genetic drift is the primary cause of allele frequency changes in natural populations. We tested the probability that the observed increase in frequency of the 191-bp allele could be explained by genetic drift by calculating the FST value of European D. melanogaster populations using a set of X-linked loci that have been genotyped in six European populations (including the population from Texel). This FST value provides an estimate for the drift between European populations since they expanded from Africa. Assuming an island model, it is possible to estimate the Ne/t as
We used 10,000 forward simulations based on the estimated Ne/t to determine the probability that an allele frequency could increase from p0 = 0.0032 to pt = 0.18. The fraction of simulations that provided a pt equal or larger than 0.18 was considered as the probability to obtain the increase in frequency by chance, rather than by selection.
Results
Allele Frequencies at the Locus XTF0
A recent genome scan for loci affected by positive selection in non-African D. melanogaster populations identified several candidate regions (Kauer, Dieringer, and Schl?tterer 2003). In the wake of further analyses of the molecular variability at loci flanking these candidate regions, we detected a highly divergent pattern in allelic frequencies among European populations at the microsatellite locus XTF0 (Catania et al., unpublished data).
At this locus, an allele (191 bp) was scored in the population from Texel at a frequency of 0.18, whereas it was virtually absent in the other 12 European populations (314 individuals [fig. 1]). Apart from Texel, this allele was detected only in one single line from Friedrichshafen. Interestingly, we did not detect this allele in our sample of non-European flies, including African, Asian, and American populations (120 individuals [data not shown]), indicating that the 191-bp allele is probably worldwide at a low frequency.
FIG. 1.— Allele frequencies observed at locus XTF0 for the 13 European populations investigated. The populations are indicated by numbers as follows: 1 (Texel, The Netherlands), 2 (Neumarkt, Germany), 3 (Rome, Italy), 4 (Naples, Italy), 5 (Katowice, Poland), 6 (Gotheron, France), 7 (Friedrichshafen, Germany), 8 (Copenhagen, Denmark), 9 (Sofia, Bulgaria), 10 (Chaves, Portugal), 11 (Kreta, Greece), 12 (Moscow, Russia), and 13 (Helsinki, Finland).
Of the three other classes of alleles scored in Europe (fig. 1), the most frequent ones (169 bp [0.88] and 173 bp [0.08]) were both detected in European and non-European populations. The population from Chaves (Portugal) was the only European population for which a 189-bp allele was detected. In the non-European populations from Harare (Zimbabwe), La Milpa (Belize, Central America) and Jiamusi (China), this allele was also observed.
Genotyping and Detection of Recombination Breakpoints
The frequency of the 191-bp allele in the D. melanogaster population from Texel was significantly higher than the overall frequency observed in the rest of Europe (Fisher's exact test, P < 0.0001). Given that the high frequency of an otherwise rare allele could be the indication of a selective sweep, we investigated the pattern of molecular variability in the genomic region flanking the locus XTF0.
Fourteen microsatellite loci (table 1 in Supplementary Material online), located at increasing distances from the locus XTF0, were surveyed in 28 Texel lines. Although gene diversity differed among loci, the joint analysis of all individuals from the Texel population provided no conspicuous pattern of nonneutral evolution. A different picture emerged when the lines were grouped according to allelic state at locus XTF0 (fig. 2). Individuals carrying the 191-bp allele were completely monomorphic for a genomic region of approximately 106 kb downstream and 518 kb upstream of microsatellite XTF0 (table 3 in Supplementary Material online). Outside of this interval, variability recovered on both sides, indicating that recombination occurred between the microsatellites showing no or normal levels of variability. No reduction in variation was observed for the other individuals (n = 16) from the Texel population (figure 2 and table 3 in Supplementary Material online).
FIG. 2.— Gene diversity (Hexp) of lines originating from Texel (n = 28), in the genomic regions flanking the locus XTF0 (absolute chromosomal distance is shown according to the D. melanogaster genome annotation release 3.1). Hexp values are separately displayed for the individuals bearing the allele 191 (black line) and the allele 169 (grey line) at the locus XTF0. The 12 lines bearing the allele 191 show an identical pattern of variability (Hexp = 0) that extends for approximately 624 kb. Locus XTF11 is not included because it is not informative.
Variability of DNA Sequences
To further confirm the absence of molecular variation along the surveyed region, we studied sequence variation at three noncoding regions. These DNA fragments cover a genomic region of 107 kb and are located in the center of the region, with no variation among individuals bearing the 191-bp allele.
As with the microsatellite analysis, we grouped the sequences according to the allele at locus XTF0 (table 1). Consistent with the microsatellite analysis, for all fragments, we observed complete sequence identity among the individuals carrying the 191-bp allele. Thus, the absence of variation in the 191-bp allele haplotype was verified on the sequence level. Interestingly, no fixed differences were observed between the two sets of sequences, suggesting a shared history of both allelic classes.
No significant departure from neutral expectations was detected, irrespective of whether all individuals were analyzed or we used the reconstructed population samples (see Material and Methods). We also investigated the pattern of linkage disequilibrium (LD) in the Texel population, as LD analysis is widely used to infer selection. Nevertheless, the recent colonization history of non-African D. melanogaster populations and, in particular, of the Texel population significantly complicates the inference of reliable estimators of the effective population size. As formal tests for LD need to consider the demographic past and the current effective population size, we limited our analysis to a comparison of the three polymorphic fragments. Significant pairwise LD was only detected for fragment 1. Interestingly, the same result was obtained when only nonselected chromosomes were analyzed, suggesting no increase in pairwise LD caused by the presence of the selected haplotype (table 4 in Supplementary Material online). Furthermore, we compared ZnS values (Kelly 1997) for the three fragments and obtained conflicting results. For fragment 1 we obtained the maximum value of 1, irrespective of whether only nonselected chromosomes, reconstructed data set 1, or reconstructed data set 2 was analyzed. For fragment 2, we observed an increase of LD with the number of selected haplotypes. This pattern suggests that for this locus, some deviation from neutrality may be detected if the appropriate tests were applied. The third fragment, however, displayed the opposite trend: with an increasing number of selected haplotypes, ZnS values decreased (table 1).
The 191-bp Allele in the Friedrichshafen Population
Apart from the Texel population, the 191-bp allele was also detected in a single individual collected in Friedrichshafen (figure 1 in Supplementary Material online). We genotyped six microsatellites (out of the 14 used for Texel lines) and sequenced all three fragments for this chromosome. The joined analysis of both markers indicated complete sequence identity over a maximum distance of approximately 156 kb with the Texel 191-bp chromosome. One polymorphism (a 1-base indel) was detected in the DNA fragment located 20 kb downstream from XTF0 (i.e., fragment 1) and a different allele was observed at the microsatellite locus XTF9, positioned approximately 136 kb upstream of XTF0.
High Frequency of the 191-bp Allele in Texel: an Artifact of Relatedness?
The relatively high frequency of the 191-bp allele in the flies from Texel could potentially be an artifact of relatedness among the individuals carrying this allele. Thus, in a genealogical setting, individuals carrying the 191-bp allele would have a more recent common ancestor than the remaining individuals in the population.
To test this hypothesis, we surveyed autosomal microsatellites (rather than X-linked ones, such as XTF0) on a sample of Texel individuals, which were partitioned according to their allelic state at locus XTF0 (alleles 169 and 191). If carriers of the 191-bp allele were closely related, then genetic differentiation (FST > 0) between the two groups would be detected. If the individuals in the population were equally related, no differentiation would be expected. Our analysis of five autosomal microsatellite loci showed no evidence for population differentiation between the two groups (FST = –0.024, P = 0.83). Consistent with this observation, we also found no clustering of individuals carrying the 191-bp allele in a tree representing the genealogical history of the population sample (figure 2 in Supplementary Material online).
Given that the high frequency of the 191-bp allele in the Texel population could be the result of a bottleneck specific to the Texel population, we tested this hypothesis by using a recently published data set of 48 microsatellite loci genotyped for six European D. melanogaster populations (including the population from Texel). Gene diversity in the Texel population (H = 0.51 ± 0.18) was in a similar range as for other European populations (table 2). Furthermore, pairwise FST values between other European populations and the Texel population were not consistently higher than pairwise comparisons not involving Texel (Caracristi and Schl?tterer 2003). This indicates that the Texel population did not experience higher levels of genetic drift than other European populations.
Table 2 Gene Diversity of Six European D. melanogaster Populations Estimated from 48 Microsatellite Loci
High Frequency of the 191-bp Allele in Texel: Drift or Selection?
Genetic drift is a major force changing allele frequencies in natural populations. Nevertheless, in large populations, only a weak correlation exists between linked sites. Thus, if one allele increases in frequency by genetic drift, this does not extend to flanking sites. An increase in frequency caused by selection will, in contrast, also affect linked sites (hitchhiking [Maynard Smith and Haigh 1974]). Given that the Texel population, like other European populations, is not in mutation-recombination-drift equilibrium, it is not possible to perform formal tests on our data set for the expected degree of linkage disequilibrium. Although the pattern of linkage disequilibrium on the chromosomes not carrying the 191-bp allele does not suggest that both allelic classes have been exposed to the same processes (table 1), no formal test could be devised. We tested, therefore, whether the increase in frequency of the 191-bp allele is consistent with random genetic drift under a neutral model of evolution. Based on the FST values among six European populations (0.05 [Caracristi and Schl?tterer 2003]), we determined Ne/t. Forward simulations indicated that genetic drift is extremely unlikely to explain the increase in frequency from 0.0032 to 0.18 (P = 0.0022). Even an FST value three times as high as the observed value could not account for the observed increase in frequency of the 191-bp allele (P = 0.0072). Therefore, we conclude that genetic drift is extremely unlikely to explain the increase in frequency of the 191-bp allele in the Texel population.
Timing and Strength of the Presumed Selective Sweep
Assuming that a selective sweep increased the frequency of the 191-bp allele, it is interesting to estimate the time since the onset of the selective sweep. We used two different approaches (see Material and Methods), both taking advantage of our observation that 12 individuals shared a 624-kb stretch of identical sequence. As recombination will reduce the genomic region identical by descent around the target of selection, the size of this genomic region can be used to obtain a rough time estimate. We used the estimated local recombination rate of 0.0027 cM/kb (Comeron, Kreitman, and Aguade 1999).
Both approaches produced similar results for the time a genomic region of 624 kb can be expected to be IBD. The first approach, which is based on the number of individuals with the 191-bp allele and the size of the genomic IBD region, indicated an age of nine generations. This is certainly an underestimate of the age of the selective sweep, as it assumes complete independence among the 191-bp chromosomes. The second approach builds on the observed frequency of the 191-bp allele in the population to infer the age of the IBD region. Using this approach, we estimate 185 generations. Assuming five generations/year, the two approaches provide a range for the estimated age of the spread of the beneficial mutation of 1.8 and 37 years.
We used the conservative estimate of 185 generations to infer the selection coefficient s that was associated with the increase in the frequency of the 191-bp allele to 18%. With an assumed starting frequency of 0.0032 (reflecting the frequency in Europe), we estimate a selection coefficient of 0.046.
Discussion
In one population from Texel, we observed a relatively high (18%) frequency of a microsatellite allele that was almost absent in the remainder of the European population. Given that non-African populations have low levels of differentiation (Caracristi and Schl?tterer 2003), such a striking difference in allele frequency suggests nonneutral evolution in the Texel population. Population genetic theory predicts that not only the target of selection increases in frequency in a population until it becomes fixed (selective sweep) but also neutral variation in the genomic region flanking the target of selection is affected (hitchhiking). Assuming that one rare microsatellite allele is linked to a beneficial mutation, the high frequency of the 191-bp allele in the Texel population could be explained by a selective sweep, which is not yet completed (partial selective sweep). The large genomic region (624 kb), which is IBD for those individuals carrying the 191-bp allele, provides further support for a selective sweep in the Texel population.
Interestingly, our results closely resemble those of Hudson et al. (1994), in which a partial selective sweep was proposed as explanation for the pattern of molecular variability observed in a region encompassing the CuZn-superoxide dismutase (Sod) locus (Hudson et al. 1994). In that case, a total of 28 sequences (out of 41 studied) including the SodS and SodF alleles (alternative forms of the enzyme) revealed no variation over 1,428 bp across three D. melanogaster populations. Conversely, the remaining 13 sequences showed a number of different haplotypes. Further analyses of the sweep in subsequent studies (Hudson, Sáez, and Ayala 1997; Sáez et al. 2003) have then shed light on the length of the swept chromosomal segment (41 to 54 kb), as well as its age (2,600 to 22,000 years) and the strength of selection (0.020 to 0.103). Yet, the target of selection has not been identified.
Despite these striking similarities between our results and the Sod data, we also note several important differences. In comparison to Hudson et al. (1994), who detected the pattern of a partial selective sweep in multiple populations, our study detected the partial selective sweep only in a single population, despite that a large number of populations were also analyzed. Furthermore, the size of the genomic IBD region is much larger in our study than for the Sod locus.
Evidence for a Strong Selective Sweep
Kaplan, Hudson, and Langley (1989) showed that for recombination rates typical of D. melanogaster, a selection coefficient equal to 0.01 could potentially remove variation at sites up to 10 kb from the target of selection (Kaplan, Hudson, and Langley 1989). Consistent with the observation that the monomorphic region in our study was significantly larger than the one observed by Kaplan, Hudson, and Langley (1989), we also inferred a much larger s (0.046). This estimate falls well within the range of s determined for the Sod locus (0.020 – 0.103) (Sáez et al. 2003) and for DDT resistance at Cyp6g1 in Drosophila (0.022 to 0.11) (Catania et al. 2004; Schlenke and Begun 2004). Interestingly, it is substantially larger than the selection coefficients inferred for two genomic regions carrying mutations facilitating the out-of-Africa habitat expansion of D. melanogaster (Harr, Kauer, and Schl?tterer 2002), which are 0.002 to 0.01 and 0.0001 to 0.002, respectively. Given that our estimate of the selection coefficient is very conservative, as we assumed that the beneficial mutation occurred 185 generations ago, rather than our second estimate of nine generations, the selective sweep in the Texel population is probably among the strongest sweeps described for D. melanogaster.
Local Adaptation?
It is conceivable that we uncovered a local selective sweep improving adaptation to the habitat of D. melanogaster on the Texel island. This is consistent with the almost complete absence of the 191-bp allele in the remaining European populations. Nevertheless, we also note that our calculations indicate a very recent origin of the beneficial mutation. Thus, it is also possible that the beneficial mutation in the Texel population has not yet spread to other European populations. Further studies of multiple, more recently collected, populations in and close to Texel are required to address this question.
Irrespective of whether the selective sweep detected in our study is confined to the island of Texel or not, our findings strongly suggest that we observe a beneficial mutation that has probably originated outside of Africa. Finally, we would mention that an alternative explanation for our observed pattern could be that the habitat has changed and a previously neutral mutation has become favorable.
Mapping the Target of Selection
Recently, it has been suggested that the comparison of populations with and without the selective sweep could be used for hitchhiking mapping (Schl?tterer 2002a, 2002b, 2003). In our case, however, the window of genomic region affected by the selective sweep is too large to be informative. Based on release 3.1 of the FlyBase, the genomic IBD region (624 kb) harbors 46 annotated genes, which precludes further testing of candidate genes. Thus, such strong recent selective sweeps are not well suited for hitchhiking mapping.
Power to Detect Partial Selective Sweeps
Our study was motivated by the presence of one high frequency allele that was virtually absent in all other European D. melanogaster populations. However, if we had surveyed any of the other flanking microsatellite loci, no population-specific pattern would have emerged, and, thus, this selective sweep would have remained unnoticed. This raises a substantial complication for any effort to estimate the number of beneficial mutations required for adaptation to a novel habitat. Unless the density of markers in a genome scan is extremely high, the number of beneficial mutations is very likely to be underestimated in a very unpredictable manner.
A recent study of selective events that did not sweep to fixation (Sabeti et al. 2002) noted that very large sample sizes are required to identify a statistically significant deviation from neutrality. Our sequence analysis of a moderate number of chromosomes suggested that LD analysis for such small sample sizes are unlikely to be a powerful approach for the identification of partial selective sweeps.
Alternative Explanations?
We tested for the possibility that the high frequency of the 191-bp allele could be the consequence of a very close relationship of the carriers of the allele. Such a situation may occur if the population in Texel was derived from a small number of founders, with one founder carrying the 191-bp allele. Our analysis of genetic differentiation among the bearers of the 191- bp allele and the remainder of the population, however, did not provide evidence for a very recent founder event. Also, the level of variation in the Texel population is very similar to other European populations, providing no support for the idea of a recent founder event specific to the Texel population.
A second explanation for the large genomic IBD region around the 191-bp allele could be the presence of a mini-inversion in this region. However, two lines of argument provide evidence against this hypothesis: (1) we observed different breakpoints for the different chromosomes at the edges of the IBD region, and (2) for the 191-bp chromosome in the Friedrichshafen population the IBD region is smaller than in the Texel population.
Acknowledgements
We thank all the members of the CS-lab, C. Vogl, and R. Bürger for constructive suggestions and critical discussion. We are also grateful to D. Evdochenko, whose microsatellite data gave the input to the present study. One anonymous reviewer provided very helpful suggestions that improved the manuscript. This work was supported by Fonds zur F?rderung der wissenschaftlichen Forschung (FWF) grants and an EMBO young investigator award to C.S.
References
Caracristi, G., and C. Schl?tterer. 2003. Genetic differentiation between American and European Drosophila melanogaster populations could be attributed to admixture of African alleles. Mol. Biol. Evol. 20:792–799.
Catania, F., M. O. Kauer, P. J. Daborn, J. L. Yen, R. H. ffrench-Constant, and C. Schl?tterer. 2004. Worldwide survey of an Accord insertion and its association with DDT resistance in Drosophila melanogaster. Mol. Ecol. 13:2491–2504.
Comeron, J. M., M. Kreitman, and M. Aguade. 1999. Natural selection on synonymous sites is correlated with gene length and recombination in Drosophila. Genetics 151:239–249.
Corander, J., P. Waldmann, and M. J. Sillanpaa. 2003. Bayesian analysis of genetic differentiation between populations. Genetics 163:367–374.
Dieringer, D., and C. Schl?tterer. 2003. Microsatellite analyzer (MSA)—a platform independent analysis tool for large microsatellite data sets. Mol. Ecol. Notes 3:167–169.
Glinka, S., L. Ometto, S. Mousset, W. Stephan, and D. De Lorenzo. 2003. Demography and natural selection have shaped genetic variation in Drosophila melanogaster: a multi-locus approach. Genetics 165:1269–1278.
Harr, B., M. Kauer, and C. Schl?tterer. 2002. Hitchhiking mapping—a population based fine mapping strategy for adaptive mutations in D. melanogaster. Proc. Natl. Acad. Sci. USA 99:12949–12954.
Hudson, R. R., K. Bailey, D. Skarecky, J. Kwiatowski, and F. J. Ayala. 1994. Evidence for positive selection in the superoxide dismutase (Sod) region of Drosophila melanogaster. Genetics 136:1329–1340.
Hudson, R. R., A. G. Sáez, and F. J. Ayala. 1997. DNA variation at the Sod locus of Drosophila melanogaster: an unfolding story of natural selection. Proc. Natl. Acad. Sci. USA 94:7725–7729.
Kaplan, N. L., R. R. Hudson, and C. H. Langley. 1989. The "hitchhiking effect" revisited. Genetics 123:887–899.
Kauer, M. O., D. Dieringer, and C. Schl?tterer. 2003. A microsatellite variability screen for positive selection associated with the "out of Africa" habitat expansion of Drosophila melanogaster. Genetics 165:1137–1148.
Kelly, J. K. 1997. A test of neutrality based on interlocus associations. Genetics 146:1197–1206.
Maynard Smith, J., and J. Haigh. 1974. The hitch-hiking effect of a favorable gene. Genet. Res. 23:23–35.
Miller, S. A., D. D. Dykes, and H. F. Polesky. 1988. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16:1215.
Myant, N. B., S. A. Forbes, I. N. Day, and J. Gallagher. 1997. Estimation of the age of the ancestral arginine3500glutamine mutation in human apoB-100. Genomics 45:78–87.
Orr, H. A., and A. J. Betancourt. 2001. Haldane's sieve and adaptation from the standing genetic variation. Genetics 157:875–884.
Payseur, B. A., A. D. Cutter, and M. W. Nachman. 2002. Searching for evidence of positive selection in the human genome using patterns of microsatellite variability. Mol. Biol. Evol. 19:1143–1153.
Reynolds, J. B., B. S. Weir, and C. C. Cockerham. 1983. Estimation of the coancestry coefficient: basis for a short-term genetic distance. Genetics 105:767–779.
Rozas, J., J. C. Sanchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19:2496–2497.
Sabeti, P. C., D. E. Reich, J. M. Higgins et al. (17 co-authors). 2002. Detecting recent positive selection in the human genome from haplotype structure. Nature 419:832–837.
Sáez, A. G., A. Tatarenkov, E. Barrio, N. H. Becerra, and F. J. Ayala. 2003. Patterns of DNA sequence polymorphism at Sod vicinities in Drosophila melanogaster: unraveling the footprint of a recent selective sweep. Proc. Natl. Acad. Sci. USA 100:1793–1798.
Schlenke, T. A., and D. J. Begun. 2004. Strong selective sweep associated with a transposon insertion in Drosophila simulans. Proc. Natl. Acad. Sci. USA 101:1626–1631.
Schl?tterer, C. 1998. Microsatellites. Pp. 237–261 in A. R. Hoelzel, ed. Molecular genetic analysis of populations: a practical approach 2/e. Oxford University Press, Oxford.
———. 2002a. A microsatellite-based multilocus screen for the identification of local selective sweeps. Genetics 160:753–763.
———. 2002b. Towards a molecular characterization of adaptation in local populations. Curr. Opin. Genet. Dev. 12:683–687.
———. 2003. Hitchhiking mapping—functional genomics from the population genetics perspective. Trends Genet. 19:32–38.
Schl?tterer, C., C. Vogl, and D. Tautz. 1997. Polymorphism and locus-specific effects on polymorphism at microsatellite loci in natural Drosophila melanogaster populations. Genetics 146:309–320.
Schl?tterer, C., and B. Zangerl. 1999. The use of imperfect microsatellites for DNA fingerprinting and population genetics. Pp. 153–165 in J. T. Epplen and T. Lubjuhn, eds. DNA profiling and DNA fingerprinting. Birkh?user, Basel.
Sch?fl, G., and C. Schl?tterer. 2004. Patterns of microsatellite variability among X chromosomes and autosomes indicate a high frequency of beneficial mutations in non-African D. simulans. Mol. Biol. Evol. 21:1384–1390.
Tajima, F. 1983. Evolutionary relationship of DNA sequences in finite populations. Genetics 105:437–460.
———. 1989. The effect of change in population size on DNA polymorphism. Genetics 123:597–601.
Vigouroux, Y., M. McMullen, C. T. Hittinger, K. Houchins, L. Schulz, S. Kresovich, Y. Matsuoka, and J. Doebley. 2002. Identifying genes of agronomic importance in maize by screening microsatellites for evidence of selection during domestication. Proc. Natl. Acad. Sci. USA 99:9650–9655.
Watterson, G. A. 1975. On the number of segregating sites in genetical models without recombination. Theor. Popul. Biol. 7:256–276.(F. Catania and C. Schl?tt)
Correspondence: E-mail: christian.schloetterer@vu-wien.ac.at.
Abstract
It is well understood that the out-of-Africa habitat expansion of D. melanogaster was associated with the fixation of many beneficial mutations. Nevertheless, it is not clear yet whether these beneficial mutations segregated already in Africa or originated outside of Africa. In this article, we describe an ongoing selective sweep specific to one European population. One microsatellite allele has increased in a population from The Netherlands to a frequency of 18%, whereas it is virtually absent in 12 other European populations. The selective sweep resulted in a genomic region of more than 600 kb that is identical by descent. This is probably the first evidence of a beneficial mutation that has arisen outside of Africa and has resulted in a selective sweep localized in a population from The Netherlands.
Key Words: local adaptation ? allele frequency ? microsatellite ? partial selective sweep
Introduction
Hitchhiking mapping is an approach for the identification of genomic regions that were recently subjected to positive directional selection (reviewed in Schl?tterer [2003]). By scanning a large number of markers distributed over the genome, candidate regions could be identified without the need of experimental crosses or information about the selected trait. Hitchhiking mapping has already been applied successfully to a wide variety of species ranging from humans to maize (e.g., Payseur, Cutter, and Nachman 2002; Vigouroux et al. 2002; Kauer, Dieringer, and Schl?tterer 2003).
In Drosophila, hitchhiking mapping studies have mainly focused on the question of which genomic regions were affected by positive selection associated with the out-of-Africa habitat expansion (Harr, Kauer, and Schl?tterer 2002; Glinka et al. 2003; Kauer, Dieringer, and Schl?tterer 2003; Sch?fl and Schl?tterer 2004). On the X chromosome, approximately 10% to 20% of the surveyed markers were found to deviate from neutral expectations (Glinka et al. 2003; Kauer, Dieringer, and Schl?tterer 2003; Sch?fl and Schl?tterer 2004). Interestingly, when multiple European populations were analyzed, the same loci carried the signature of a selective sweep (Harr, Kauer, and Schl?tterer 2002; Kauer, Dieringer, and Schl?tterer 2003). This raises the question of whether the beneficial mutation originated outside of Africa or whether it was already segregating in Africa at a low frequency (Orr and Betancourt 2001). Until now, no experimental data are available to distinguish these two scenarios.
A recent survey of multiple D. melanogaster populations from North-America and Europe indicated low, but statistically significant, differences between and within continents (Caracristi and Schl?tterer 2003). Thus, the presence of population substructure in combination with a broad range of different habitats covered by non-African D. melanogaster populations raises the following question: Does local adaptation occur in non-African D. melanogaster populations? While a wealth of comparative data analyses between African and non-African populations is available, very limited data exist about selective sweeps specific to local non-African populations (Schl?tterer, Vogl, and Tautz 1997).
In this study, a pronounced differentiation among non-African (European) populations is described for an X-linked locus. At this locus, one microsatellite allele was present at a relatively high frequency in flies originating from the island of Texel (The Netherlands) but virtually absent in the remainder of the European populations investigated. The absence of molecular variability linked to this allele strongly suggests that its frequency has risen because of a local event of positive selection.
Materials and Methods
Microsatellites and Populations Samples
An X-linked microsatellite locus (XTF0) was studied in 19 (13 Europeans and six non-Europeans) D. melanogaster populations (number of lines): Naples, Italy (30*), Rome, Italy (40*), Neumarkt, Germany (30*), Katowice, Poland (16*), Texel, The Netherlands (66*), Gotheron, France (30*), Friedrichshafen, Germany (40*), Copenhagen, Denmark (30*), Sofia, Bulgaria (24*), Chaves, Portugal (18), Kreta, Greece (12*), Moscow, Russia (30*), Helsinki, Finland (14*), Sengwa Wildlife Preserve, Zimbabwe (18), Harare, Zimbabwe (28), Lake Kariba, Zimbabwe (12). Jiamusi, China (30*), La Milpa, Belize (20*), New Delhi, India (12). The asterisk (*) indicates F1 flies, which are the offspring of freshly collected flies. Other lines are inbred isofemale lines; thus, for genotyping, we randomly selected one allele (Dieringer and Schl?tterer 2003).
Further, 14 microsatellite loci were typed in lines originating from Texel, The Netherlands (28 lines) at both sides of locus XTF0. These loci span a distance of approximately 970 kb (for absolute positions, see table 1 in Supplementary Material online).
Genotyping
DNA extraction from single individuals was carried out according the high-salt extraction procedure described in Miller, Dykes, and Polesky. (1988). Genotyping of microsatellite loci followed standard protocols (Schl?tterer 1998). PCR products were separated on a 7% denaturing polyacrylamid gel (32% formamide, 5.6 M urea) and visualized by autoradiography. PCR products of known size and a slippage ladder were used to determine the allele sizes (Schl?tterer and Zangerl 1999).
DNA Sequencing
At least twelve lines collected from Texel and one line from Friedrichshafen were sequenced at three noncoding DNA fragments spanning approximately 107 kb (table 2 in Supplementary Material online). The average length of the fragments was 642 bp (see table 1). Genomic DNA was extracted from single male flies using a high-salt extraction method (see Genotyping). Each fragment was sequenced in both directions. Primers, used either for PCR amplification or for sequencing reaction, were designed on the basis of the release 3.1 of the complete D. melanogaster genomic sequence (table 2 in Supplementary Material online). Sequencing reactions were performed with the BigDye Terminator version 1.1 cycle sequencing chemistry. The sequencing products were purified with Sephadex G-50 fine (Amersham Biosciences, Sweden), and a MegaBACE automated sequencer was used for separation. Sequences were aligned by ClustalX, manually adjusted, and compared with the GenBank sequence. Sequences were deposited in GenBank under accession numbers AY689233 to AY689273.
Table 1 Summary Statistics of the Sequenced DNA Fragments
Data Analysis
Expected heterozygosity, allele-frequencies, and FST values were estimated using the software MSA version 3.14 (Dieringer and Schl?tterer 2003). Significance of observed FST values was obtained by permuting genotypes 10,000 times among groups.
Sequence diversity was estimated using either the number of segregating sites, w (Watterson 1975), or the average number of pairwise sequence differences in the sample, (Tajima 1983). To test whether the observed allele frequency spectrum was in accordance with the expectations from the neutral model of evolution, Tajima's D statistic (Tajima 1989) was calculated. All computations were carried out using the DnaSP version 4.00 software package (Rozas et al. 2003). Gaps in the sequence alignments were excluded from the analyses. Linkage disequilibrium analyses were also performed with DnaSP 4.00.
Because the 191-bp allele was overrepresented in our sample of sequenced chromosomes, we reconstructed a population sample by discarding all but one (reconstruct 1) or two (reconstruct 2) chromosomes with the 191-bp allele. Using this strategy, the frequency of the 191-bp chromosomes was similar to a randomly drawn sample from the Texel population.
Estimate of Relatedness
The degree of relatedness among lines from Texel was estimated to test whether the individuals bearing the 191-bp allele were more related to each other than to the remainder of the population. Five unlinked microsatellite loci on the third chromosome (table 1 in Supplementary Material online) were genotyped in a total of 29 individuals (eight bearing the 191-bp allele). The relatedness was estimated using both the FST parameter (between the two groups of individuals bearing/not bearing the 191-bp allele) and the BAPS version 2.0 software (Corander, Waldmann, and Sillanpaa 2003). The latter software uses a MCMC algorithm to estimate posterior probabilities for all the possible clusters produced by the individuals surveyed. The following settings were used in our analysis: burn-in of 10,000, a chain length of 10,000, and a stochastic initialization (no initial cluster partition).
Time of the Selective Event
We determined the age of the genomic region IBD (Identical by Descent) using two approaches based on (1) the detected number of lines bearing the 191-bp allele and (2) the frequency of the 191-bp allele in the population sample.
In the first approach, we calculated the probability that a stretch of DNA experiences recombination (Pr), as Pr = (1 – (1 – r)L)(1 – f191) rL (1 – f191), where r is the recombination rate in the region of interest, L is the length of the stretch of DNA, and f191 is the observed frequency of the 191-bp allele. Then, given the probability of recombination (Pr) and the number of IBD haplotypes (n), Er(= nPr) is the expected number of recombination events/generation. We inferred the number of generations, k, by using the largest integer value of k that satisfies the equation Erk < 1.
The second approach was outlined previously (Myant et al. 1997). We calculated the probability Pnr that the IBD region associated with the 191-bp allele does not recombine with another genomic region not IBD co-occurring in the population: Pnr = 1 – Pr. By solving the equation (Pnr)k = f191, we inferred k, the number of generations elapsed between the date of collection and the timepoint at which the selected variant has arisen on the ancestral chromosome.
Because recombination could only occur in females, we adjusted the estimates from both approaches by multiplying by 3/2.
Selection Coefficient
Assuming semidominance, we calculated the selection coefficient s using the equation
The frequency of the 191-bp allele in the European D. melanogaster populations (excluding Texel) was used as an estimate for the frequency at the onset of selection (p0 = 0.0032) and the observed frequency of the same allele in the Texel population as the current frequency (pt = 0.18), q = 1–p. The time t from the onset of selection to the present was conservatively estimated by the approach of Myant et al. (1997), as this yielded a longer time since the occurrence of the beneficial mutation (see Results).
Probability of Neutral Drift
Random genetic drift is the primary cause of allele frequency changes in natural populations. We tested the probability that the observed increase in frequency of the 191-bp allele could be explained by genetic drift by calculating the FST value of European D. melanogaster populations using a set of X-linked loci that have been genotyped in six European populations (including the population from Texel). This FST value provides an estimate for the drift between European populations since they expanded from Africa. Assuming an island model, it is possible to estimate the Ne/t as
We used 10,000 forward simulations based on the estimated Ne/t to determine the probability that an allele frequency could increase from p0 = 0.0032 to pt = 0.18. The fraction of simulations that provided a pt equal or larger than 0.18 was considered as the probability to obtain the increase in frequency by chance, rather than by selection.
Results
Allele Frequencies at the Locus XTF0
A recent genome scan for loci affected by positive selection in non-African D. melanogaster populations identified several candidate regions (Kauer, Dieringer, and Schl?tterer 2003). In the wake of further analyses of the molecular variability at loci flanking these candidate regions, we detected a highly divergent pattern in allelic frequencies among European populations at the microsatellite locus XTF0 (Catania et al., unpublished data).
At this locus, an allele (191 bp) was scored in the population from Texel at a frequency of 0.18, whereas it was virtually absent in the other 12 European populations (314 individuals [fig. 1]). Apart from Texel, this allele was detected only in one single line from Friedrichshafen. Interestingly, we did not detect this allele in our sample of non-European flies, including African, Asian, and American populations (120 individuals [data not shown]), indicating that the 191-bp allele is probably worldwide at a low frequency.
FIG. 1.— Allele frequencies observed at locus XTF0 for the 13 European populations investigated. The populations are indicated by numbers as follows: 1 (Texel, The Netherlands), 2 (Neumarkt, Germany), 3 (Rome, Italy), 4 (Naples, Italy), 5 (Katowice, Poland), 6 (Gotheron, France), 7 (Friedrichshafen, Germany), 8 (Copenhagen, Denmark), 9 (Sofia, Bulgaria), 10 (Chaves, Portugal), 11 (Kreta, Greece), 12 (Moscow, Russia), and 13 (Helsinki, Finland).
Of the three other classes of alleles scored in Europe (fig. 1), the most frequent ones (169 bp [0.88] and 173 bp [0.08]) were both detected in European and non-European populations. The population from Chaves (Portugal) was the only European population for which a 189-bp allele was detected. In the non-European populations from Harare (Zimbabwe), La Milpa (Belize, Central America) and Jiamusi (China), this allele was also observed.
Genotyping and Detection of Recombination Breakpoints
The frequency of the 191-bp allele in the D. melanogaster population from Texel was significantly higher than the overall frequency observed in the rest of Europe (Fisher's exact test, P < 0.0001). Given that the high frequency of an otherwise rare allele could be the indication of a selective sweep, we investigated the pattern of molecular variability in the genomic region flanking the locus XTF0.
Fourteen microsatellite loci (table 1 in Supplementary Material online), located at increasing distances from the locus XTF0, were surveyed in 28 Texel lines. Although gene diversity differed among loci, the joint analysis of all individuals from the Texel population provided no conspicuous pattern of nonneutral evolution. A different picture emerged when the lines were grouped according to allelic state at locus XTF0 (fig. 2). Individuals carrying the 191-bp allele were completely monomorphic for a genomic region of approximately 106 kb downstream and 518 kb upstream of microsatellite XTF0 (table 3 in Supplementary Material online). Outside of this interval, variability recovered on both sides, indicating that recombination occurred between the microsatellites showing no or normal levels of variability. No reduction in variation was observed for the other individuals (n = 16) from the Texel population (figure 2 and table 3 in Supplementary Material online).
FIG. 2.— Gene diversity (Hexp) of lines originating from Texel (n = 28), in the genomic regions flanking the locus XTF0 (absolute chromosomal distance is shown according to the D. melanogaster genome annotation release 3.1). Hexp values are separately displayed for the individuals bearing the allele 191 (black line) and the allele 169 (grey line) at the locus XTF0. The 12 lines bearing the allele 191 show an identical pattern of variability (Hexp = 0) that extends for approximately 624 kb. Locus XTF11 is not included because it is not informative.
Variability of DNA Sequences
To further confirm the absence of molecular variation along the surveyed region, we studied sequence variation at three noncoding regions. These DNA fragments cover a genomic region of 107 kb and are located in the center of the region, with no variation among individuals bearing the 191-bp allele.
As with the microsatellite analysis, we grouped the sequences according to the allele at locus XTF0 (table 1). Consistent with the microsatellite analysis, for all fragments, we observed complete sequence identity among the individuals carrying the 191-bp allele. Thus, the absence of variation in the 191-bp allele haplotype was verified on the sequence level. Interestingly, no fixed differences were observed between the two sets of sequences, suggesting a shared history of both allelic classes.
No significant departure from neutral expectations was detected, irrespective of whether all individuals were analyzed or we used the reconstructed population samples (see Material and Methods). We also investigated the pattern of linkage disequilibrium (LD) in the Texel population, as LD analysis is widely used to infer selection. Nevertheless, the recent colonization history of non-African D. melanogaster populations and, in particular, of the Texel population significantly complicates the inference of reliable estimators of the effective population size. As formal tests for LD need to consider the demographic past and the current effective population size, we limited our analysis to a comparison of the three polymorphic fragments. Significant pairwise LD was only detected for fragment 1. Interestingly, the same result was obtained when only nonselected chromosomes were analyzed, suggesting no increase in pairwise LD caused by the presence of the selected haplotype (table 4 in Supplementary Material online). Furthermore, we compared ZnS values (Kelly 1997) for the three fragments and obtained conflicting results. For fragment 1 we obtained the maximum value of 1, irrespective of whether only nonselected chromosomes, reconstructed data set 1, or reconstructed data set 2 was analyzed. For fragment 2, we observed an increase of LD with the number of selected haplotypes. This pattern suggests that for this locus, some deviation from neutrality may be detected if the appropriate tests were applied. The third fragment, however, displayed the opposite trend: with an increasing number of selected haplotypes, ZnS values decreased (table 1).
The 191-bp Allele in the Friedrichshafen Population
Apart from the Texel population, the 191-bp allele was also detected in a single individual collected in Friedrichshafen (figure 1 in Supplementary Material online). We genotyped six microsatellites (out of the 14 used for Texel lines) and sequenced all three fragments for this chromosome. The joined analysis of both markers indicated complete sequence identity over a maximum distance of approximately 156 kb with the Texel 191-bp chromosome. One polymorphism (a 1-base indel) was detected in the DNA fragment located 20 kb downstream from XTF0 (i.e., fragment 1) and a different allele was observed at the microsatellite locus XTF9, positioned approximately 136 kb upstream of XTF0.
High Frequency of the 191-bp Allele in Texel: an Artifact of Relatedness?
The relatively high frequency of the 191-bp allele in the flies from Texel could potentially be an artifact of relatedness among the individuals carrying this allele. Thus, in a genealogical setting, individuals carrying the 191-bp allele would have a more recent common ancestor than the remaining individuals in the population.
To test this hypothesis, we surveyed autosomal microsatellites (rather than X-linked ones, such as XTF0) on a sample of Texel individuals, which were partitioned according to their allelic state at locus XTF0 (alleles 169 and 191). If carriers of the 191-bp allele were closely related, then genetic differentiation (FST > 0) between the two groups would be detected. If the individuals in the population were equally related, no differentiation would be expected. Our analysis of five autosomal microsatellite loci showed no evidence for population differentiation between the two groups (FST = –0.024, P = 0.83). Consistent with this observation, we also found no clustering of individuals carrying the 191-bp allele in a tree representing the genealogical history of the population sample (figure 2 in Supplementary Material online).
Given that the high frequency of the 191-bp allele in the Texel population could be the result of a bottleneck specific to the Texel population, we tested this hypothesis by using a recently published data set of 48 microsatellite loci genotyped for six European D. melanogaster populations (including the population from Texel). Gene diversity in the Texel population (H = 0.51 ± 0.18) was in a similar range as for other European populations (table 2). Furthermore, pairwise FST values between other European populations and the Texel population were not consistently higher than pairwise comparisons not involving Texel (Caracristi and Schl?tterer 2003). This indicates that the Texel population did not experience higher levels of genetic drift than other European populations.
Table 2 Gene Diversity of Six European D. melanogaster Populations Estimated from 48 Microsatellite Loci
High Frequency of the 191-bp Allele in Texel: Drift or Selection?
Genetic drift is a major force changing allele frequencies in natural populations. Nevertheless, in large populations, only a weak correlation exists between linked sites. Thus, if one allele increases in frequency by genetic drift, this does not extend to flanking sites. An increase in frequency caused by selection will, in contrast, also affect linked sites (hitchhiking [Maynard Smith and Haigh 1974]). Given that the Texel population, like other European populations, is not in mutation-recombination-drift equilibrium, it is not possible to perform formal tests on our data set for the expected degree of linkage disequilibrium. Although the pattern of linkage disequilibrium on the chromosomes not carrying the 191-bp allele does not suggest that both allelic classes have been exposed to the same processes (table 1), no formal test could be devised. We tested, therefore, whether the increase in frequency of the 191-bp allele is consistent with random genetic drift under a neutral model of evolution. Based on the FST values among six European populations (0.05 [Caracristi and Schl?tterer 2003]), we determined Ne/t. Forward simulations indicated that genetic drift is extremely unlikely to explain the increase in frequency from 0.0032 to 0.18 (P = 0.0022). Even an FST value three times as high as the observed value could not account for the observed increase in frequency of the 191-bp allele (P = 0.0072). Therefore, we conclude that genetic drift is extremely unlikely to explain the increase in frequency of the 191-bp allele in the Texel population.
Timing and Strength of the Presumed Selective Sweep
Assuming that a selective sweep increased the frequency of the 191-bp allele, it is interesting to estimate the time since the onset of the selective sweep. We used two different approaches (see Material and Methods), both taking advantage of our observation that 12 individuals shared a 624-kb stretch of identical sequence. As recombination will reduce the genomic region identical by descent around the target of selection, the size of this genomic region can be used to obtain a rough time estimate. We used the estimated local recombination rate of 0.0027 cM/kb (Comeron, Kreitman, and Aguade 1999).
Both approaches produced similar results for the time a genomic region of 624 kb can be expected to be IBD. The first approach, which is based on the number of individuals with the 191-bp allele and the size of the genomic IBD region, indicated an age of nine generations. This is certainly an underestimate of the age of the selective sweep, as it assumes complete independence among the 191-bp chromosomes. The second approach builds on the observed frequency of the 191-bp allele in the population to infer the age of the IBD region. Using this approach, we estimate 185 generations. Assuming five generations/year, the two approaches provide a range for the estimated age of the spread of the beneficial mutation of 1.8 and 37 years.
We used the conservative estimate of 185 generations to infer the selection coefficient s that was associated with the increase in the frequency of the 191-bp allele to 18%. With an assumed starting frequency of 0.0032 (reflecting the frequency in Europe), we estimate a selection coefficient of 0.046.
Discussion
In one population from Texel, we observed a relatively high (18%) frequency of a microsatellite allele that was almost absent in the remainder of the European population. Given that non-African populations have low levels of differentiation (Caracristi and Schl?tterer 2003), such a striking difference in allele frequency suggests nonneutral evolution in the Texel population. Population genetic theory predicts that not only the target of selection increases in frequency in a population until it becomes fixed (selective sweep) but also neutral variation in the genomic region flanking the target of selection is affected (hitchhiking). Assuming that one rare microsatellite allele is linked to a beneficial mutation, the high frequency of the 191-bp allele in the Texel population could be explained by a selective sweep, which is not yet completed (partial selective sweep). The large genomic region (624 kb), which is IBD for those individuals carrying the 191-bp allele, provides further support for a selective sweep in the Texel population.
Interestingly, our results closely resemble those of Hudson et al. (1994), in which a partial selective sweep was proposed as explanation for the pattern of molecular variability observed in a region encompassing the CuZn-superoxide dismutase (Sod) locus (Hudson et al. 1994). In that case, a total of 28 sequences (out of 41 studied) including the SodS and SodF alleles (alternative forms of the enzyme) revealed no variation over 1,428 bp across three D. melanogaster populations. Conversely, the remaining 13 sequences showed a number of different haplotypes. Further analyses of the sweep in subsequent studies (Hudson, Sáez, and Ayala 1997; Sáez et al. 2003) have then shed light on the length of the swept chromosomal segment (41 to 54 kb), as well as its age (2,600 to 22,000 years) and the strength of selection (0.020 to 0.103). Yet, the target of selection has not been identified.
Despite these striking similarities between our results and the Sod data, we also note several important differences. In comparison to Hudson et al. (1994), who detected the pattern of a partial selective sweep in multiple populations, our study detected the partial selective sweep only in a single population, despite that a large number of populations were also analyzed. Furthermore, the size of the genomic IBD region is much larger in our study than for the Sod locus.
Evidence for a Strong Selective Sweep
Kaplan, Hudson, and Langley (1989) showed that for recombination rates typical of D. melanogaster, a selection coefficient equal to 0.01 could potentially remove variation at sites up to 10 kb from the target of selection (Kaplan, Hudson, and Langley 1989). Consistent with the observation that the monomorphic region in our study was significantly larger than the one observed by Kaplan, Hudson, and Langley (1989), we also inferred a much larger s (0.046). This estimate falls well within the range of s determined for the Sod locus (0.020 – 0.103) (Sáez et al. 2003) and for DDT resistance at Cyp6g1 in Drosophila (0.022 to 0.11) (Catania et al. 2004; Schlenke and Begun 2004). Interestingly, it is substantially larger than the selection coefficients inferred for two genomic regions carrying mutations facilitating the out-of-Africa habitat expansion of D. melanogaster (Harr, Kauer, and Schl?tterer 2002), which are 0.002 to 0.01 and 0.0001 to 0.002, respectively. Given that our estimate of the selection coefficient is very conservative, as we assumed that the beneficial mutation occurred 185 generations ago, rather than our second estimate of nine generations, the selective sweep in the Texel population is probably among the strongest sweeps described for D. melanogaster.
Local Adaptation?
It is conceivable that we uncovered a local selective sweep improving adaptation to the habitat of D. melanogaster on the Texel island. This is consistent with the almost complete absence of the 191-bp allele in the remaining European populations. Nevertheless, we also note that our calculations indicate a very recent origin of the beneficial mutation. Thus, it is also possible that the beneficial mutation in the Texel population has not yet spread to other European populations. Further studies of multiple, more recently collected, populations in and close to Texel are required to address this question.
Irrespective of whether the selective sweep detected in our study is confined to the island of Texel or not, our findings strongly suggest that we observe a beneficial mutation that has probably originated outside of Africa. Finally, we would mention that an alternative explanation for our observed pattern could be that the habitat has changed and a previously neutral mutation has become favorable.
Mapping the Target of Selection
Recently, it has been suggested that the comparison of populations with and without the selective sweep could be used for hitchhiking mapping (Schl?tterer 2002a, 2002b, 2003). In our case, however, the window of genomic region affected by the selective sweep is too large to be informative. Based on release 3.1 of the FlyBase, the genomic IBD region (624 kb) harbors 46 annotated genes, which precludes further testing of candidate genes. Thus, such strong recent selective sweeps are not well suited for hitchhiking mapping.
Power to Detect Partial Selective Sweeps
Our study was motivated by the presence of one high frequency allele that was virtually absent in all other European D. melanogaster populations. However, if we had surveyed any of the other flanking microsatellite loci, no population-specific pattern would have emerged, and, thus, this selective sweep would have remained unnoticed. This raises a substantial complication for any effort to estimate the number of beneficial mutations required for adaptation to a novel habitat. Unless the density of markers in a genome scan is extremely high, the number of beneficial mutations is very likely to be underestimated in a very unpredictable manner.
A recent study of selective events that did not sweep to fixation (Sabeti et al. 2002) noted that very large sample sizes are required to identify a statistically significant deviation from neutrality. Our sequence analysis of a moderate number of chromosomes suggested that LD analysis for such small sample sizes are unlikely to be a powerful approach for the identification of partial selective sweeps.
Alternative Explanations?
We tested for the possibility that the high frequency of the 191-bp allele could be the consequence of a very close relationship of the carriers of the allele. Such a situation may occur if the population in Texel was derived from a small number of founders, with one founder carrying the 191-bp allele. Our analysis of genetic differentiation among the bearers of the 191- bp allele and the remainder of the population, however, did not provide evidence for a very recent founder event. Also, the level of variation in the Texel population is very similar to other European populations, providing no support for the idea of a recent founder event specific to the Texel population.
A second explanation for the large genomic IBD region around the 191-bp allele could be the presence of a mini-inversion in this region. However, two lines of argument provide evidence against this hypothesis: (1) we observed different breakpoints for the different chromosomes at the edges of the IBD region, and (2) for the 191-bp chromosome in the Friedrichshafen population the IBD region is smaller than in the Texel population.
Acknowledgements
We thank all the members of the CS-lab, C. Vogl, and R. Bürger for constructive suggestions and critical discussion. We are also grateful to D. Evdochenko, whose microsatellite data gave the input to the present study. One anonymous reviewer provided very helpful suggestions that improved the manuscript. This work was supported by Fonds zur F?rderung der wissenschaftlichen Forschung (FWF) grants and an EMBO young investigator award to C.S.
References
Caracristi, G., and C. Schl?tterer. 2003. Genetic differentiation between American and European Drosophila melanogaster populations could be attributed to admixture of African alleles. Mol. Biol. Evol. 20:792–799.
Catania, F., M. O. Kauer, P. J. Daborn, J. L. Yen, R. H. ffrench-Constant, and C. Schl?tterer. 2004. Worldwide survey of an Accord insertion and its association with DDT resistance in Drosophila melanogaster. Mol. Ecol. 13:2491–2504.
Comeron, J. M., M. Kreitman, and M. Aguade. 1999. Natural selection on synonymous sites is correlated with gene length and recombination in Drosophila. Genetics 151:239–249.
Corander, J., P. Waldmann, and M. J. Sillanpaa. 2003. Bayesian analysis of genetic differentiation between populations. Genetics 163:367–374.
Dieringer, D., and C. Schl?tterer. 2003. Microsatellite analyzer (MSA)—a platform independent analysis tool for large microsatellite data sets. Mol. Ecol. Notes 3:167–169.
Glinka, S., L. Ometto, S. Mousset, W. Stephan, and D. De Lorenzo. 2003. Demography and natural selection have shaped genetic variation in Drosophila melanogaster: a multi-locus approach. Genetics 165:1269–1278.
Harr, B., M. Kauer, and C. Schl?tterer. 2002. Hitchhiking mapping—a population based fine mapping strategy for adaptive mutations in D. melanogaster. Proc. Natl. Acad. Sci. USA 99:12949–12954.
Hudson, R. R., K. Bailey, D. Skarecky, J. Kwiatowski, and F. J. Ayala. 1994. Evidence for positive selection in the superoxide dismutase (Sod) region of Drosophila melanogaster. Genetics 136:1329–1340.
Hudson, R. R., A. G. Sáez, and F. J. Ayala. 1997. DNA variation at the Sod locus of Drosophila melanogaster: an unfolding story of natural selection. Proc. Natl. Acad. Sci. USA 94:7725–7729.
Kaplan, N. L., R. R. Hudson, and C. H. Langley. 1989. The "hitchhiking effect" revisited. Genetics 123:887–899.
Kauer, M. O., D. Dieringer, and C. Schl?tterer. 2003. A microsatellite variability screen for positive selection associated with the "out of Africa" habitat expansion of Drosophila melanogaster. Genetics 165:1137–1148.
Kelly, J. K. 1997. A test of neutrality based on interlocus associations. Genetics 146:1197–1206.
Maynard Smith, J., and J. Haigh. 1974. The hitch-hiking effect of a favorable gene. Genet. Res. 23:23–35.
Miller, S. A., D. D. Dykes, and H. F. Polesky. 1988. A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16:1215.
Myant, N. B., S. A. Forbes, I. N. Day, and J. Gallagher. 1997. Estimation of the age of the ancestral arginine3500glutamine mutation in human apoB-100. Genomics 45:78–87.
Orr, H. A., and A. J. Betancourt. 2001. Haldane's sieve and adaptation from the standing genetic variation. Genetics 157:875–884.
Payseur, B. A., A. D. Cutter, and M. W. Nachman. 2002. Searching for evidence of positive selection in the human genome using patterns of microsatellite variability. Mol. Biol. Evol. 19:1143–1153.
Reynolds, J. B., B. S. Weir, and C. C. Cockerham. 1983. Estimation of the coancestry coefficient: basis for a short-term genetic distance. Genetics 105:767–779.
Rozas, J., J. C. Sanchez-DelBarrio, X. Messeguer, and R. Rozas. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19:2496–2497.
Sabeti, P. C., D. E. Reich, J. M. Higgins et al. (17 co-authors). 2002. Detecting recent positive selection in the human genome from haplotype structure. Nature 419:832–837.
Sáez, A. G., A. Tatarenkov, E. Barrio, N. H. Becerra, and F. J. Ayala. 2003. Patterns of DNA sequence polymorphism at Sod vicinities in Drosophila melanogaster: unraveling the footprint of a recent selective sweep. Proc. Natl. Acad. Sci. USA 100:1793–1798.
Schlenke, T. A., and D. J. Begun. 2004. Strong selective sweep associated with a transposon insertion in Drosophila simulans. Proc. Natl. Acad. Sci. USA 101:1626–1631.
Schl?tterer, C. 1998. Microsatellites. Pp. 237–261 in A. R. Hoelzel, ed. Molecular genetic analysis of populations: a practical approach 2/e. Oxford University Press, Oxford.
———. 2002a. A microsatellite-based multilocus screen for the identification of local selective sweeps. Genetics 160:753–763.
———. 2002b. Towards a molecular characterization of adaptation in local populations. Curr. Opin. Genet. Dev. 12:683–687.
———. 2003. Hitchhiking mapping—functional genomics from the population genetics perspective. Trends Genet. 19:32–38.
Schl?tterer, C., C. Vogl, and D. Tautz. 1997. Polymorphism and locus-specific effects on polymorphism at microsatellite loci in natural Drosophila melanogaster populations. Genetics 146:309–320.
Schl?tterer, C., and B. Zangerl. 1999. The use of imperfect microsatellites for DNA fingerprinting and population genetics. Pp. 153–165 in J. T. Epplen and T. Lubjuhn, eds. DNA profiling and DNA fingerprinting. Birkh?user, Basel.
Sch?fl, G., and C. Schl?tterer. 2004. Patterns of microsatellite variability among X chromosomes and autosomes indicate a high frequency of beneficial mutations in non-African D. simulans. Mol. Biol. Evol. 21:1384–1390.
Tajima, F. 1983. Evolutionary relationship of DNA sequences in finite populations. Genetics 105:437–460.
———. 1989. The effect of change in population size on DNA polymorphism. Genetics 123:597–601.
Vigouroux, Y., M. McMullen, C. T. Hittinger, K. Houchins, L. Schulz, S. Kresovich, Y. Matsuoka, and J. Doebley. 2002. Identifying genes of agronomic importance in maize by screening microsatellites for evidence of selection during domestication. Proc. Natl. Acad. Sci. USA 99:9650–9655.
Watterson, G. A. 1975. On the number of segregating sites in genetical models without recombination. Theor. Popul. Biol. 7:256–276.(F. Catania and C. Schl?tt)