A cytological study of the O5 chromosomal inversion of Drosophila subobscura (Diptera, Drosophilidae)

The O5 chromosomal inversion has been a cornerstone for understanding different aspects of the American colonization by Drosophila subobscura. To obtain more information of this evolutionary event it is important to know the pattern of bands of this inversion in detail. Comparing this pattern with that of D. melanogaster it is possible to predict which genes are located inside or close to the O5 inversion and use them as genetic markers. In this study, the complete band pattern of the O5 inversion is presented. Furthermore, the most important genes located inside it have been predicted. Finally, a constriction located close to the proximal breakage point of the O5 inversion has been observed many times and its possible genetic significance is discussed.


Introduction
Until 1978, D. subobscura was considered a species with a Palearctic area of distribution. In that year this species was first found in the Chilean locality of Puerto Montt (Brncic et al., 1981), but in 1982 it was also detected in Port Townsend (Washington) (Beckenbach and Prevosti, 1986). The species is now found in Chile from La Serena to Punta Arenas and in many localities of Argentina in the Andes mountain range. In North America the species is spread along the West Coast (Prevosti et al., 1987;1989). This phenomenon of colonization has been studied at different genetic levels: the chromosomal polymorphism (Prevosti et al., 1985;1988;, the allozyme polymorphism (Prevosti et al., 1983;Balanyà et al., 1994), the lethal genes (Mestres et al., 1990;Solé et al., 2000), the polygenes for quantitative traits (Pegueroles et al., 1995;1999;Huey et al., 2000), the mtDNA (Latorre et al., 1986;Rozas et al., 1990), the genomic DNA variation (Rozas and Aguadé, 1991) and the effect of inbreeding (Pegueroles et al., 1996).
From the beginning our research group tried to search for the origin of the colonization. The chromosomal polymorphism obtained in the American populations of D. subobscura (Prevosti et al., 1985;1988; showed a high similarity with that from the Western Mediterranean area (Prevosti et al., 1984) with the important exception of the O 5 inversion. This inversion is present in the American populations but it has never been found in the Western Mediterranean samples. The O 5 inversion from America has an interesting property, it is always completely associated with a lethal gene (Mestres et al., 1990;. In spite of its association with a lethal gene its frequency shows a clinal distribution with latitude in both colonized areas. However, in the Palearctic region the distribution of O 5 is rather erratic (Krimbas and Loukas, 1980;Krimbas, 1993;Mestres et al., 1994;Zivanovic and Mestres, 2000) and its frequency is extremely low with the exception of South Scandinavia (Sperlich, 1964;Pinsker and Sperlich, 1981). Furthermore its combination with lethal genes is variable. Thus, an O 5 inversion from Gävle (Sweden) was lethal free, while another one from Lilla Edet (also in Sweden) was found to be semilethal (Mestres et al., 1992). O 5 inversions from Taulé (France) and Zanjic (Yugoslavia), on the other hand, were associated with lethal genes. However, these lethals were neither allelic between each other nor with those from the American populations (Mestres et al., 1992;Zivanovic and Mestres, 2000). The constant association of the lethal effect and the O 5 inversion in America can be used as a tracer to detect the local Palearctic population from which the colonization started. It is most probable that a single O 5 gene arrangement with the lethal gene was among the chromosome sample of the colonizing flies . Thus, if an O 5 inversion carrying the lethal gene found in America is detected in a Palearctic population, this could be the one from which the colonization started. As the lethal genes from Taulé and Zanjic are different from those found in America, the origin of the colonization remains still unknown.
The breakage points of the O 5 inversions from Europe and America seem to be the same (Mestres et al., 1992;Zivanovic and Mestres, 2000), but the behavior of the O 5 arrangement from America and from the Palearctic region is different. In America significant latitudinal clines can be found (Prevosti et al., 1985;1988;); its frequency is higher than expected from inbreeding experiments (Pegueroles et al., 1996) and a significantly higher segregation can be observed in recombination studies (Mestres et al., 1998). All these results seem to support that the O 5 inversion is associated with an heterotic effect in America (Mestres et al., 2001). Although the breakage points of all O 5 inversions studied are identical, the possibility of recurrence in this inversion must be taken into account.
For all these reasons the O 5 inversion has a key function for understanding the American colonization by D. subobscura. The aim of the present study is to obtain more information about the cytology of this inversion. The detailed analysis of the banding pattern of the O 5 inversion will allow a comparison with the genetic and karyotypic information available in other species of Drosophila (mainly D. melanogaster), and it might be possible to recognize the genes that are located in the inversion or in its neighborhood.

Material and Methods
All O 5 chromosomal lines available carry a lethal gene and no homokariotypic strain O 5 /O 5 could be established by inbreeding. However, the lethal gene of the O 5 chromosomal lines are not allelic in all cases. For instance, as previously mentioned, the lethal gene associated with the American O 5 inversion is not allelic with that from Taulé (France). Thus, it is possible to obtain O 5 /O 5 individuals with O chromosomes of different origin. For this purpose the T15 (from Taulé) and the G7A (from Gilroy, California), both O 5 chromosomal lines with different lethal genes, were crossed. The strategy of these crosses is presented in Figure 1. One Va/O 5 male from Taulé was crossed with Va/O 5 virgin females from Gilroy. The parental individuals are heterokaryotypic balanced over a chromosome Va (Varicose) from the Va/Ba balanced lethal strain (Sperlich et al., 1977). The Varicose chromosome carries the X-ray induced overlapping inversions O VIII + 210 and the naturally occurring O 3+4 region. Among the offspring the karyotypes: Va/O 5 , O 5 /O 5 and Va/Va are expected, but Va/Va genotypes die because the Va gene is a recessive lethal. The problem is the time of lethal effect of the Va gene. Since the Va gene of D. subobscura shows the same behavior as the Delta gene of D. melanogaster, it is generally assumed that they are homologous genes (Krimbas, 1993). In D. melanogaster, the Delta gene is embryonic lethal (Lindsley and Zimm, 1992). However, if the Va/Va genotypes can survive the larval stadium, then both homokaryotypes (O 5 /O 5 and Va/Va) might be present in the F 1 larval offspring. Certainly, it would be possible to distinguish these genotypes by their pattern of bands, but this would slow down the analysis. To overcome this problem, one possibility is to perform a more complex strategy of crosses preventing the appearance of Va/Va individuals (Mestres et al., 1992). Another option is to determine the time of the lethal effect of  crosses between the lethal chromosomal lines T15 and G7A all homokaryotypes must be O 5 /O 5 . The polytene chromosomes were stained and squashed in aceto-orcein solution and observed using a light microscope.

Results and Discussion
The pattern of bands in the O 5 /O 5 homokaryotypes is presented in Figure 3. The pairing is accurate indicating the full coincidence of the breakage points confirming previous observations (Mestres et al., 1992;Zivanovic and Mestres, 2000). Although it is not possible to exclude a recurrent origin of new O 5 inversions by mutation, the probability of this phenomenon is expected to be very low. Using the method proposed by Sperlich and Pfriem (1986) the estimate of this recurrence would be between 1.368 x 10 -7 and 9.488 x 10 -8  assuming that the breakage points are randomly distributed along the O chromosome of D. subobscura. However, this is most unlikely, because many "hot spots" can be identified in this chromosome (Kunze-Mühl and Müller, 1958;Krimbas, 1993)  chromosomes of D. subobscura where breaks occur more frequently than in other regions and might be discussed in the context of a possible recurrence of the O 5 inversion.
As the pattern of bands obtained in the homokaryotype O 5 /O 5 is very clear, it could be directly compared with that of the 3R -chromosome arm of D. melanogaster . The region covered by the O 5 inversion corresponds approximately to the section 86 -A to 89 -F of the 3R chromosome of D. melanogaster. According to Lindsley and Zimm (1992) the most important genes located in this segment are: Odh (Octanol dehydrogenase), Sdh-2 (Sorbitol dehydrogenase), ants (antennas), cu (curled), man (mandarin), Hsp 70A (Heat -shock protein), ttr (tetrapter), ry (rosy), Ace (acetyl cholinesterase), aur (aurora), red (red Malpighian tubes), trx (trithorax), cv-c (crossveinless c), Tm1 (Tropomyosin 1), jvl (javelinlike), Po (Pyridoxal oxidase), Sb (Stubble), ss (spineless), Ubx (Ultrabithorax), abd-A (abdominal A), Abd-B (Abdominal B) and cal (coal). Thus, some of these genes could be used as molecular markers for the O 5 inversions from American and Palearctic populations. The DNA of the Odh (86 -D 1-4) and the Sdh-2 (86 -D) genes has been sequenced in D. melanogaster (Luque et al., 1994;Luque et al., 1998). Both genes are located inside, but close to the proximal breakage point of the O 5 inversion in D. subobscura. The DNA sequence of a fragment of the Odh gene of D. subobscura has been recently used to characterize the O 5 and other inversions in American and Palearctic populations of D. subobscura (Abad, 2000). Homeotic genes such as trx (88 -B3), Ubx (89 -E1), abd-A (89 -E2) and Adb-B (89 -E2) have been also characterized at the molecular level in D. melanogaster (Martin et al., 1995). These genes map inside and close to the distal breakage point of the O 5 inversion, thus it might become possible to cover the two ends of this inversion and use them for future investigations.