Transposable elements in South American populations of Drosophila simulans

Elements transposables dans des populations sud-americaines de Drosophila simulans. Cette etude recherche la presence de quatre elements transposables (mariner. gypsy, hobo et 412) au sein de populations sud-americaines de Drosophila simulans. Le profil d'hybridation genomique de 12 populations differentes a ete determine par Southern Blot. Bien que seul un petit nombre de copies de mariner ait ete observe, chaque population presentait un profil d'hybridation caracteristique, ce qui suggere que l'element etait actif. Le nombre de copies de l'element gypsy s'est avere tres petit; avec un profil d'hybridation similaire dans toutes les lignees, ce qui indique que l'element etait inactif dans les populations etudiees. L'une des lignees ne comportait pratiquement pas d'element gypsy, ce qui n'etait encore jamais arrive chez D. simulans ni D. melanogaster. L'element hobo de 1,1 kb n'a ete detecte dans aucune lignee d'Amerique du Sud alors qu'il existe dans toutes les lignees, originaires d'autres sites geographiques, analysees jusqu'a present. Ces informations en accord avec les donnees de la litterature, suggerent que l'element hobo n'a peut-etre envahi que recemment le genome des populations de D. simulans en Amerique du Sud. Pour l'element 412, les populations etudiees ont montre des bandes specifiques. ce qui suggere que cet element peut avoir des activites de transposition dans le genome de ces mouches.


INTRODUCTION
Transposable elements (TEs) have already been shown to compose a significant fraction of the genome of a wide variety of organisms [1]. Approximately 10 % of the D. melanogaster genome consists of 50 different families of TEs [13]. The amount of middle repetitive DNA sequences in D. simulans was estimated to be about one third that present in its sibling species D. melanogaster !10). It is believed that a major portion of this repetitive DNA is composed of TEs !26). Virtually all families of TEs that have been cloned from D. melanogaster are also represented in the genome of D. simulans !6). The only exception to the similar TE composition between the two species is the presence of the P family, which is only found in D. melanogaster and mariner which is only found in D. simulans. Considering that virtually the same TE families are present in both species and that D. simulans has only one third the total repetitive DNA present in D. melanogaster, the mean number of element copies per family should be lower in D. simulans [17,20,23].
These last authors detected significant differences in copy number of TEs between those species by in situ hybridization.
Studies covering the whole distribution range of D. simulans should make it possible to better understand the differences in the number of copies of TEs between this species and D. melanogaster, in addition to providing important information on the evolutionary history of TEs. In an attempt to contribute to the resolution of this question, we investigated the occurrence of four different transposable element families (mariner, hobo, gypsy and 412), by means of Southern blot analyses, in 12 D. simulans populations originating from different locations within the American continent.

Southern blots
Genomic DNA was prepared from 40-50 adult flies, according to Jowett [16]. DNA samples (approximately 5 vg from each strain) were completely digested with restriction endonucleases, submitted to electrophoresis on 0.8 % agarose gels, transferred to nylon membranes and hybridized to nick-translated DNA probes (labeled with 32 P-a-dATP) in the presence of 50 % formamide at 42 °C. Each filter was washed three times with 0.2X SSC and 0.5 % SDS for 20 min at 42 °C.
A 0.9-kb fragment obtained by NheI/PvuII digestion of Mos 1 plasmid DNA [19] was used as a probe for the analysis of the mariner TE family. As a probe for the gypsy element, we employed the pGGHS plasmid, which contains a complete gypsy element (9). Members of the hobo family were probed with the pHX4 plasmid, which contains a XhoI 2.6-kb fragment of hobo element of D. melanogaster. The 2.6-kb fragment was removed from a complete element contained in the pHLFl plasmid [4] and subcloned into the Bluescript plasmid. The probe used to detect 412, was an element four HindIII-EcoRI fragment from cDml,l2 with a total length of 4.4-kb was used [27]. When the DNAs from D. simulans populations of diverse geographical origins were hybridized to the mariner element probe, we observed hybridization patterns that were characteristic to each one of them. The mean number of hybridizing bands varied according to the restriction enzyme used. For instance, when D. simulans DNA was digested with Sall, we saw, on average, 9.0 t 4.1 hybridizing bands (see figure 1A). The yellow strain presented the largest number of hybridization bands (16 bands); the Montevideo population presented the smallest number (one band).
Since the mariner element possesses a single site for Saff, we expected each copy of the element to produce two hybridization bands when the genomic DNA was digested with this enzyme. The number of copies per genome should have been equal to approximately half the number of observed bands. Enzymes that do not have internal sites in the mariner element sequence would yield a smaller number of bands, which should correspond to the approximate number of copies of the element. When DNA was digested with XhoI, for example, the number of hybridizing bands was on average 4.4 f 1.5. Once again, the yellow strain was the one with the largest number of bands (a total of nine) which were not seen in this Southern, but have been observed in other experiments (data not shown). Maquin6 and 1093 were the ones showing the smallest number (only three hybridization bands) ( figure 1 B).

Hobo family
The cleavage sites for XhoI in the hobo element are close to the inverted terminal repeats, and digestion with this enzyme yields a 2.6-kb-long fragment when a complete hobo copy is present, and smaller bands if deleted elements are present. All strains analysed possessed the 2.6-kb band, indicating the occurrence of complete elements (figure !). Boussy and Daniels [5] have shown the occurrence of a 0.7-kb band (corresponding to a deleted element very common in D. simulans) in the majority of the strains they analysed. In the present study, we detected this band in the North American strains (1093), but we did not find it in any of the South American populations.
Some strains, such as dpp and yellow presented a very weak signal for the 2.6-kb band, indicating that possibly only a few copies of the complete element were present in these strains. Bands greater than 2.6 kb, as seen in figure 2, were also encountered in other species of the melanogaster subgroup. According to Boussy and Daniels [5] the occurrence of such bands can be explained by the presence of other TEs which carry sequences related to hobo. Hobo elements bearing a deleted or altered XhoI site are also interpreted as old sequences of hobo localized in heterochromatin.

Gypsy family
Hybridizing the membranes to the gypsy element probe resulted in a small number of hybridization bands, with an average of 7.8 ! 2.8 bands. A highly conserved hybridization pattern was observed throughout the D. simulans populations. When the fly DNA was digested with Sall or EcoRI, the majority of the bands was common to almost all populations -few bands were specific to each population ( figure 3). Based on the known sequence of D. melanogaster's gypsy, this element should have a single cleavage site for EcoRI and none for Sall. Therefore, if we assume the same element structure for gypsy of D. simulans, we might have expected to see different hybridization bands among the different studied populations in the case of the occurrence of insertion site polymorphism.
The yellow strain was an important exception. It seemed to bear very few copies of gypsy in its genome, as can be seen in figure 3A (lane 8). It should be noted that the Southern membrane shown in figure 3A is the same as the one used in figure  1A, where yellow was one of the strains showing the largest number of hybridized bands with the mariner element.

412 family
The 412 element of D. melanogaster carries four XhoI restriction sites (27!. We thus expected to obtain bands corresponding to 2.6, 2.1 and 0.9-kb in size in the presence of a complete copy of l,12. As shown in figure 4A, all the analysed populations carried the expected bands, in addition to greater-sized bands, some of which were specific for each population. The 1.4-and 0.6-kb fragments expected with an EcoRI digestion were found in all populations, in addition to greater-sized bands, some of which were specific for each population (figure 4B). 4 [18] and Giraud and Capy [14] have analysed several D. sim!alans populations by Southern Blot, and their results showed that the number of mariner copies per genome varies from 0 to 15. Our present findings on the number of mariner copies occurring in the genome of D. simulans, as inferred from the number of bands encountered in the Southern blot assays, were very much in agreement with values found in previous studies about the occurrence of mariner in strains of this species from other origins. [18] have detected hybridization bands that are specific for each strain, indicating the existence of insertion site polymorphism for the mariner element. Nevertheless, these authors have also observed some genomic sites common to strains from diverse geographical origins, indicating that at least some of them are long-existing insertion sites and may be fixed within the species. Giraud and Capy [14] described three banding patterns in mariner Southern blot analysis: i) populations from different geographical origins which share the same three bands, these populations having few bands; ii) populations with more bands, some of these common to all the lines of a population; iii) populations with few bands different among populations. In the present study, we did not find any hybridization bands common to the different populations; each population showed a characteristic hybridization profile, in agreement with Giraud and Capy's pattern 3. Laboratory populations show greater variability in mariner copy number than natural populations (15!. This fact might explain the high copy number occurring in our yellow strain, as this population has been kept in the laboratory since 1982. All South American strains employed in the present study were recently collected in the wild, except for the yellow and dpp-like strains, and none of them were mariner-free. This apparent non-existence of mariner-free populations among recently collected ones is also supported by the findings of Capy et al. [8] and Giraud and Capy (14!. Periquet et al. [21] recorded the fact that almost all natural populations of D. melanogaster possess copies of a 1.5-kb-long internally deleted hobo element. They called this element Th. These authors suggest that the accumulation of these Th elements in the genome might have some regulatory role in hobo transposition. Such a role has been suggested for the KP element of the P family of D. melanogaster [3]. Boussy and Daniels [5] analysed 19 D. melanogaster and 31 D. simulans strains by Southern blot. E (empty of hobo) and H (with hobo) strains were found in both species. In several D. melanogaster strains they observed the occurrence of the 1.5-kb (Th) element. In D. simulans, only four strains (one from Peru; one from Colombia; and one of unknown origin) did not have the hybridization band that indicates the presence of a complete element. The three South American strains also did not bear the 0.7-kb band that corresponds to an internally deleted element of 1.1 kb. All other strains, originating from North America, Australia and South Africa, as well as the one with unknown origin, showed a very strong hybridization signal corresponding to the 0.7-kb band.

Maruyama and Hartl
In the present study, we found two D. simulans strains with very weak hybridization signals for the 2.6-kb band (the existence of which represents the occurrence of a complete element), indicating that these strains carry few copies of the hobo element. Such strains have been kept in the laboratory for a longer period of time (14 and 6 years) than all the other populations, which were recently collected in nature and have strong hybridization signals for the 2.6-kb band. Furthermore, the 0.7-kb band (corresponding to an internally deleted element) was encountered only in the two North American populations; all the South American populations are devoid of this element. Considering our findings of few hobo sequences in older fly stocks and the results obtained by Boussy and Daniels [5], as a whole, we may suggest that hobo has been recently scattered in the genome of South American populations of D. simulans. Another argument in favor of this possible event is the reduced number of copies of deleted elements in the South American populations, most of all, the absence of the 1.1-kb elements which are so common to populations of other geographical origins. If the deleted elements do accumulate in the genome as a function of time after invasion, and if they are truly involved in the regulation of hobo activity !11!, then we might suppose that hobo is active in the South American populations of D. simulans but not enough time has elapsed for a sufficient accumulation of deleted hobo copies in their genomes.
All D. melanogaster strains contain inactive copies of gypsy located at the same positions in the pericentromeric heterochromatin [22]. Even non-related strains, when compared by Southern blot analysis, show very similar hybridization patterns, suggesting that the gypsy element has invaded the genome of D. melanogaster very early in the species evolutionary history [7]. Nevertheless, a low copy number of putative active gypsy with no fixed sites were detected by in situ hydridization in D. melanogaster chromosomes arms [2,23] and in D. simulans [23].
The South American populations of D. simulans presently examined bear very similar hybridization patterns, in the same way as D. melanogaster strains carrying few gypsy copies do. According to our estimates, as judged by the number of Southern hybridization bands, South American populations of D. simulans also bear a reduced number of gypsy copies. However, this Southern pattern may represent the ancient inactive copies of gypsy in the heterochromatin. If there exist active copies of gypsy in a low number in polymorphic insertions sites, these can be related to the weak bands specifically detected in some populations or these copies were not able to be detected by the Southern technique, since every individual presents one specific band.
In D. melanogaster, there is no evidence of a strain not presenting a hybridization signal when probed by gypsy sequences. Our finding of a D. simulans strain almost devoid of gypsy copies is most interesting. This strain has been kept in the laboratory for 14 years. Stochastic loss [12] may be an explanation for the extremely low number of gypsy sequences it carries. This hypothesis is supported by the fact that all other studied strains that were derived from recently collected population samples do possess the gypsy element.
The 412 element is a retrotransposon, 7.6 kb in length, with 481-pb-long LTRs (27]. This element, similar to other retrotransposons, has apparently been inhabiting the genome of drosophilids for a very long period of time. Vieira and Bi6mont [24,25] determined, by in situ hybridization to polytene chromosomes, the number of copies of the 41! element occurring in natural populations of D. simulans. A gradient in copy number was observed, varying from 20 in Europe, to 3-9 in Africa, with the same tendencies between North and South America. Some transposition bursts were observed within certain local populations that possess a high copy number of 412. These authors suggest this element has recently invaded the genome of D. simulans populations. In the present study we observed that all the studied populations carried copies of an apparently complete 412 element, and that the EcoRI and XhoI sites were conserved in relation to the D. melanogaster 412 element. The occurrence of a few bands that were characteristic of each strain may suggest that this element is active in South American D. simulans populations.