the position of this upstream boundary by examining additional regions between Sod and 2021.
The Slow/Fast polymorphism is clearly not an old balanced polymorphism. On the other hand, the data suggest that natural selection has acted recently and strongly on variation in the neighborhood of Sod. The data appear compatible with a model in which a rare variant has recently risen rapidly in frequency, and is perhaps now subject to some form of balancing selection. The data at this point do not allow us to put an upper bound on the size of the region, which has been partially swept of variation. The 2021 region, which is 12.7 kb upstream from Sod, appears to be outside the swept region. Hence, one boundary of the swept region appears to be between the Sod locus and the 2021 region. Because the 1819 region, roughly 20 kb downstream of Sod appears to be in the swept region, we conclude that the swept region is greater than 20 kb in length. This in turn suggests that surprisingly strong selection acted on the selected site (selection coefficient on the order of 0.01 or higher). The pattern of linkage disequilibrium is like Fig. 2C, from which we infer that the time since the sweep is 25,000 generations or more.
These results force one to consider the possibility that balanced polymorphisms may typically be short-lived, arising and being maintained for a time too short to result in the strong peak of linked variation, such as that observed at Adh. The two best documented cases of long-maintained balanced polymorphisms are in the major histocompatibility complex in mammals (18, 19) and the S-locus (mating incompatibility determining locus) of some plants (20). These two cases involve large numbers of maintained alleles and remarkably old lineages (which presumably result in a peak of variation at linked sites.) These two examples may be the very rare exception. The most celebrated case of a balanced polymorphism is the sickle cell variant at the ß-globin locus of humans. This is a case where strong balancing selection is well documented, and it is also well documented that the currently segregating sickle cell variants are recently arisen (and have arisen independently in different populations.) Perhaps the sickle cell case, and the Sod case, are illustrative of the most common situation for protein polymorphisms, not the ancient stable polymorphisms that result in peaks of variation at linked sites. These short-lived polymorphisms might be compatible with models that incorporate temporal and spatial variability in selection coefficients (1).
On the other hand, it is important to consider alternatives to the partial selective, sweep hypothesis. This is particularly so given some similarities between the pattern of variation seen at Sod and the results of surveys at Est6 (21, 22) and Est-P (23), which are located approximately 1 centimorgan or 1,000 kb from Sod (24). The pattern of variation in the Sod locus and in the 1819 region in our recent study is remarkable for having two similar haplotypes that are highly diverged from the rest of the sample. (Note in Fig. 4 lines 968F and 498F in the Sod locus and lines 968F and 112 in the 1819 region.) This pattern of variation seems surprising and may not be expected under the sweep hypothesis. A very similar pattern was also observed in a recent survey at the Est-P locus (23). In that survey, which used a subset of the same lines used in our study, line 357F is highly diverged from the rest of the sample. Est-P and Sod are too far apart to be affected by the same selected sweep with normal rates of recombination. The presence of a small number of highly diverged haplotypes suggests the possibility of a distinct isolated subpopulation that has recently merged with another population. This could have been a geographically isolated population, but another possibility is that the diverged haplotype is or was part of an inversion. In(3L)P is a common and widespread inversion that contains Est-P and Sod (25), but no third chromosome inversions were found to be segregating in the El Rio population (26). It is not known if the diverged haplotypes could represent sequences that have “escaped” from an inversion, as has been previously suggested (22) for sequences at the Est6 locus, which is very closely linked to Est-P. Parenthetically, we note that two highly divergent lines were found in a survey of variation in the Adh region of D. melanogaster (4). The possibility of an inversion was investigated, and no direct evidence for such an inversion was found.
It is also worth noting that an earlier study of DNA sequence variation at Est6 found patterns that “suggest that allozyme 8 has both arisen and proliferated, relatively recently” (21). Thus, at Est6 as well as at Sod, it appears that certain variants have recently been driven to high frequency. We conclude that the patterns of DNA sequence variation at Est6 and Est-P have intriguing similarities to those at Sod. These may reflect similar independent selection events, but the possibility of some event or process affecting this large segment of chromosome 3 must also be entertained.
We thank Kevin Bailey, Evgeniy S.Balakirev, and Eladio Barrio for contributions to the early stages of this project; Shiliang Qin and John W.Jacobs for use of a laboratory of the Hitachi Chemical Research Center, Inc. at the University of California, Irvine; and Evgeniy S. Balakirev, Jordi Bascompte, and Francisco Rodriguez-Trelles for discussions; and Nelsson Becerra for technical assistance. This work was supported by National Institutes of Health Grant GM-42397 to F.J.A. and by a postdoctoral fellowship to A.G.S. from the Spanish Ministry of Education and Science.
1. Gillespie, J.H. (1991) The Causes of Molecular Evolution (Oxford Univ. Press, New York).
2. Kimura, M. (1983) The Neutral Theory of Molecular Evolution (Cambridge Univ. Press, Cambridge, U.K.).
3. Hudson, R.R. & Kaplan, N.L. (1988) Genetics 120, 831–840.
4. Kreitman, M. & Hudson, R.R. (1991) Genetics 127, 565–582.
5. Lee, Y.M., Misra, H.P. & Ayala, F.J. (1981) Proc. Natl. Acad. Sci. USA 78, 7052–7055.
6. Singh, R.S., Hickey, D.A. & David, J. (1982) Genetics 101, 235–256.
7. Peng, T., Moya, A. & Ayala, F.J. (1986) Proc. Natl. Acad. Sci. USA 83, 684–687.
8. Peng, T.X., Moya, A. & Ayala, F.J. (1991) Genetics 128, 381–391.
9. Tyler, R.H., Brar, H., Singh, M., Latorre, A., Graves, J.L., Mueller, L.D., Rose, M.R. & Ayala, F.J. (1993) Genetica 91, 143–149.
10. Lee, Y.M. & Ayala, F.J. (1985) FEBS Lett. 179, 115–119.
11. Hudson, R.R., Bailey, K., Skarecky, D., Kwiatowski, J. & Ayala, F.J. (1994) Genetics 136, 1329–1340.
12. Tajima, F. (1989) Genetics 123, 585–595.
13. Hudson, R.R., Kreitman, M. & Asuadé, M. (1987) Genetics 116, 153–159.
14. Kaplan, N.L., Hudson, R.R. & Langley, C.H. (1989) Genetics 123, 887–899.
15. Chovnick, A., Gelbart, W. & McCarron, M. (1977) Cell 11, 1–10.
16. Sharp, P.M. & Li, W.-H. (1989) J. Mol. Evol. 28, 398–402.
17. Rowan, R.G. & Hunt, J.A. (1991) Mol. Biol. Evol. 8, 49–70.
18. Klein, J. (1986) Natural History of the Major Histocompatibility Complex (Wiley, New York).
19. Takahata, N. (1993) in Mechanisms of Molecular Evolution, eds. Takahata, N. & Clark, A.G. (Japan Sci. Soc., Tokyo), pp. 1–21.
20. Clark, A.G., (1993) in Mechanisms of Molecular Evolution, eds. Takahata, N. & Clark, A.G. (Japan Sci. Soc., Tokyo), pp. 79–108.
21. Cooke, P.H. & Oakeshott, J.G. (1989) Proc. Natl. Acad. Sci. USA 86, 1426–1430.
22. Odgers, W.A., Healy, M.J. & Oakeshott, J.G. (1995) Genetics 141, 215–222.
23. Balakirev, E.S. & Ayala, F.J. (1996) Genetics 144, 1511–1518.
24. Hartl, D.L. & Lozovskaya, E.R. (1995) The Drosophila Genome Map: A Practical Guide (Landes, Austin, TX).
25. Voelker, R.A., Cockerham, C.C., Johnson, F.M., Schaffer, H.E., Mukai, T. & Mettler, L.E. (1978) Genetics 88, 515–527.
26. Smit-McBride, Z., Moya, A. & Ayala, F.J. (1988) Genetics 120, 1043–1051.