Heredily 61(1988)121—131 The Genetical Society of Great Britain Received 20 October 1987 Temporal surveys of French populations of Drosophila melanogaster: P—M system, enzymatic polymorphism and infection by the sigma virus D. Anxolabéhère*, * Mécanismesmoléculaires de Ia spéciation, Université P. & M. Curie, 4 place Jussieu, L. Charles-Palabost,t 75005 Paris, France. A. Fleuriett and t Institut de Biocénotique Expérimentale des G. Periquett Agrosystèmes, Faculté des Sciences, Parc Grandmont, 37200 Tours, France. t Laboratoire de Génétique, Université de Clermont Ferrand II, B.P. 45, 63170 Aubière, France. About 120 strainsderived from natural populations of D. melanogaster collected between 1983 and 1986 in the Languedoc—Roussillon area (France) were tested for their characteristics in the P—M system, enzymatic polymorphism and infection by the sigma virus. Over these four years, a clear geographical pattern was found for the P—M system over a distance of 250 km, with weak P populations in the northern zone of the area studied, Qpopulationsin the centre and M' populations in the south. This pattern was stable for P sterility potential but a decrease of the M cytotype level was observed in the southernmost zone. During the same period, the frequency of flies infected by the sigma virus also showed a regular geographical pattern with the highest values in the central zone. The frequency of infected flies increased steadily during this period. In the enzymatic study, neither geographical nor temporal trends were detected for the loci analyzed. Correlations among the different systems are discussed. INTRODUCTION as either completely lacking P sequences (M-type) or containing structurally altered P elements (M' Overthe last decade, several studies of genetic type) (Bingham eta!.,1982). In contrast, variation in natural Drosophila populations have individuals of the P type or of the Qtype(which involved non-Mendelian systems. In D. melano- induce all dysgenic traits except sterility) have gaster,suchstudies have centred on the P-M sys- 30-50 copies of P sequences (Bingham eta!., tem of transposable elements (see Bregliano and 1982), some of which are fully intact elements, Kidwell, 1983 and Engels, 1983) and on the verti- 29 kb long, that can code for a transposase (Karess cally transmitted sigma virus (L'Héritier, 1970; and Rubin, 1984) and the others, often a majority, Brun and Plus, 1980). There is a geographical are deleted P elements that, in dysgenic crosses, variation in the distribution of both systems in can be mobilized in response to the transposase Europe. The pattern of this variation points to the produced by intact elements of the same genome Languedoc-Roussillon, a region in the south of (O'Hare and Rubin, 1983). P and Qindividuals France, as an intermediate zone of special interest also possess a cellular state called the P cytotype forevolutionary studies (Anxolabéhère eta!., 1985; which represses P element activity (Engels, 1979; Fleuriet, 1986). Simmons and Bucholz, 1985). In the P— M system of Drosophila me!anogaster, Surveys of natural populations have shown pro- the syndrome known as hybrid dysgenesis occurs nounced geographical differences (Anxolabéhère when members of the P family of transposable eta!., 1982, 1984; Kidwell, 1983; Boussy and Kid- elements are activated, It is manifested in a number well, 1987). P strains predominate in the Americas of traits (such as mutation, male recombination and Central Africa whereas M cytotype strains and gonadal sterility) when active P elements are predominate in Europe, North Africa and Asia. introduced into the M cytotype individuals Recent investigations (Anxolabéhère et a!., 1985) (Engels, 1979; Kidwell, 1985). These are defined reveal a gradient from Western Europe, where 122 D. ANXOLABEHERE ET AL. most strains are Q,toMid-Asian areas, where M' strains predominate. These results concur with the hypothesis of an invasion of Eurasian populations by P elements. In the Drosophila-sigma system, the virus con- fers CO2 sensitivity on its carriers; the sigma virus is not contagious but is transmitted from fly to fly through gametes without being integrated into the chromosomes (Brun and Plus, 1980). In natural French populations of D. melanogaster, about 20 per cent of the individuals are infected by sigma. In France, the characteristics of the system differ from those observed in the rest of Europe (Fleuriet, 1986). This paper focuses on a review of the Languedoc-Roussillon area, a region of transition from P-Q to M cytotype populations and sigma virus characteristics, over four consecutive years. We examine for pattern modifications in this transition area, as well as possible correlations between the geographical patterns of the P-M system and of sigma virus. Populations of this region were also characterized for their poly- morphism at six enzymatic loci. Figure 1 Geographical location of the D.melanogaster popu- lations sampled in the Languedoc-Roussillon area from 1983 to 1986. 1. Les Fumades, 2. Ales, 3. Vermeil, 4. MATERIALSAND METHODS Vézénobres, 5. Ledignan, 6. St Hippolyte, 7. Crespian, 8. St Martin de Londres, 9. St Mathieu, 10. Boisseron, 11. Morethan 100 wild samples, collected in the Aniane, 12. Vendargues, 13. St Laurent, 14. Montpellier, Languedoc-Roussillon region over a 250 km x 15. Perols, 16. Paulhan, 17. Gigean, 18. Mèze, 19. Servian, 50km area (fig. 1) were investigated. Each con- 20. Cazouls, 21. Marseillan, 22.Cers,23. Nissan, 24. Ven- tenac, 25.Coursan,26. Ferrals, 27. Sigean, 28. Durban, 29. sisted of a large number of flies (from 60 to 250) Tuchan, 30. Salses, 31. Tautavel, 32. Rivesaltes, 33. Thuir, collected in October 1983, 1984, 1985 and 1986. 34. Bages, 35. Argelès, 36. Le Boulou, 37. Aubiry, 38. Banyuls, 39. Cerbère. Analysisof the P—M system Determinationof P-M status within a population. 05N NaOH—25M NaCl, neutralised in Following the procedures of Periquet (1980) P 3M NaCH3COO, prehybridized in 2 x SSC —5 x potential activity was measured at 285°C by deter- FPG solution for two hours and hybridized over- mining the percentage of dysgenic gonads (GD night at 65°C in 2 x SCC —Il x FPT and 10 per cent sterility) in 50 Fl females resulting from a cross dextran sufate solution with 32P labelled Hind III between females from the Canton-S reference restriction fragment (P1 probe), or with labelled strain and males from the tested strain (Cross A). Hind Ill—Sal I restriction fragment (P2 probe) of M cytotype level was measured at 285°C by deter- the pir251 plasmid. After hybridization, the filter mining the percentage of dysgenic gonads in 50 was washed for one hour at 65°C in 2 x SSC, 01 per Fl females resulting from a cross between males cent SDS and then rapidly at room temperature in from the Harwich reference strain and females 01 x SSC, Ol per cent SDS. from the strain being tested (Cross A*). Hybridization techniques. The presence of P homo- Analysisof enzymatic polymorphism logous sequences in individual flies was detected Allozymevariation was studied by starch gel elec- by the squash-blot technique (Tchen eta!., 1985; trophoresis using the techniques described by Anxolabéhère eta!., 1985). Each fly was crushed Charles-Palabost eta!. (1986). Six polymorphic on a nylon filter membrane (Pall Biodyne) and the enzyme loci were analyzed: Acph (acid phos- DNA fixed on the filter was denatured in phatase; 3-1014), Adh (atchol dehydrogenase; P ELEMENTS IN POPULATIONS 123 2—501), Est—C (esterase-C; 3—476), Est-6 RESULTS (esterase-6; 3—36 8), a-Gpdh (ct-glycerophosphate dehydrogenase; 2-20.5) and Pgm (phospho- The P—M system glucomutase; 3-434). Fig.2 presents the P—M status of collected strains, tested for their induced GD sterility when crossed with Harwich and Canton S. Spearman rank corre- Analysisof the Drosophila-sigma system lation tests clearly show the existence of a Thefrequency of flies infected by the sigma virus latitudinal variation. For each of the four years was estimated for each sample by standard CO2 studied, both P sterility potential and M cytotype sensitivity tests (Plus, 1954), performed on about level are significantly correlated with populations 50 adults collected in the wild. ranked from North to South (the eighty coefficients P POTENTIAL M CYTOTYPE % GD STERILITY GD STERILITY 1983 IFl0Kn 30 30 *:\ t 0• I % I" • ' 10 $1 -10 17 3 39 1984 1985 20 10— 20 39 I I I 1 ALES MONTPELLIER NARBONNE PERPIGNAN SPAIN Figure 2 Geographical and temporal distributions of strains collected in Languedoc-Roussillon according to their potential for the P—M system. Solid lines =Ppotential activity; dashed lines =Mcytotype level. 124 D. ANXOLABEHERE ET AL. are within the interval 034—079 and all are sig- P and Qpopulationsappear to be stable in the nificant with P<005 to P<0001). There is a northern and central zones, a slow change might marked difference between the northern P be occuring in the south, leading to Qstrains. cytotype populations and the southern populations However this trend is slight and one must wait for with intermediate levels of M cytotype, a pattern new collections in the future to see if there is a clearly paralleled by the significantly higher P ster- real increase in the number of P elements in ility potential of northern strains as compared to southern populations. southern strains. In other European populations, P homologous To differentiate this pattern, multivariate analy- sequences have been found in strains with inter- ses were performed by component analyses. Two mediate values of M cytotype potential principal components emerged. The first is associ- (Anxolabéhère etal., 1985). In the Languedoc- ated with the cytotype and differentiates P from Roussillon area, direct hybridizations with P1 and M cytotype strains, while the second is associated P2 probes on 22 strains, selected at random from with P potential activity and differentiates P from the 1984 and 1985 sets, again demonstrated the Qstrains.These components explain 83—92 per presence of P homologues in all the individuals cent of the total geographic variation, depending tested.