Developmental temperature and somatic excision rate of mariner transposable element in three natural populations of Drosophila simulans

Summary - The temperature effect on the somatic excision rate of an inactive copy (peach) of the mariner transposable element, inserted in the white gene and responsible of the mutation (t!*!), was investigated in isofemale lines from 3 natural populations of D simulans. The somatic excision rate of the peach element was measured by the proportion of mosaic males in the offspring of a test-cross between a wild type male (originating from natural populations) and white peach females of a reference strain (GB1). A significant effect of the breeding temperature was detected in 2 out of the 3 populations investigated, ie Loua (Congo) and Bordeaux (France). In these 2 populations, the proportion of mosaic males increased with temperature. In the third population (Agadir, Morocco), the proportion of mosaic males was always high whatever the temperature. A slight correlation between the excision rate and the number of mariner copies was observed. Finally, this temperature effect may be related to a 14-bp sequence localized in the 5’ inverted repeat of the element showing 50-57% of homology with sequences of heat shock protein promoters. mosaïques augmente avec la température d’élevage. Dans la troisième population (Agadir), ce pourcentage est toujours élevé quelle que soit la température. Par ailleurs, bien que le taux d’excision de l’élément peach augmente avec le nombre moyen de copies de l’élément présent dans les différentes populations, une seule corrélation significative au seuil de 5% a été trouvée. Enfin, une séquence de 14 pb, située dans l’inversion répétée en # de l’élément, présentant 50 à 57% d’homologie avec des séquences de promoteurs de protéines du choc thermique, pourrait être responsable des effets température détectés.


INTRODUCTION
The genome response to environmental stresses has been investigated in many ways (Hoffmann and Parsons, 1991), and often by considering the mutator effect of various chemicals and radiations. In this context and during recent years, several authors have proposed that an environmental factor, like temperature, could modify the transcription and/or the transposition rate of transposable elements (Strand and McDonald, 1985;Junakovic et al, 1986;lVIcDonald et al, 1987). On the other hand, populational factors, like inbreeding, could be involved in such a phenomenon. However, although Bi6mont et al (1987Bi6mont et al ( , 1990 showed spontaneous mobilization of copia and P elements in 2 inbred lines, it cannot be assumed that consanguinity was responsible for these movements. Among the different transposable elements described in Drosophila, a temperature effect on the excision and/or transposition rate has been demonstrated in several of them. In the P-M system, hybrid dysgenesis, due to the mobility of P elements, and transposase production are enhanced at high temperature (Engels, 1989 and references therein). In the I-R system, the reactivity of I strains is affected by temperature (Picard et al, 1977;Bucheton, 1978). Junakovic et al (1986) and Ratner et al (1992) reported an increase of transposition rate related to temperature for copia-like elements and Strand and McDonald (1985) showed an increase of copia homologous transcript after a heat shock stress. This phenomenon is not limited to Drosophila, but has also been reported in many species like Antirrhinum majus for the element Tam3 (see Coen et al, 1989 and references therein), Saccharomyces cerevisiae for the element Ty (Boeke, 1989) and Escherichia coli for the element Tn3 (Sherratt, 1989). In Drosophila, the phenomenon related to transposition, ie the phenotypic effects of the transposition/excision, are stimulated when the breeding temperature increases. But, in other organisms and for other transposable elements such as Z'y and Tam3, the transposition was increased at low temperature.
For the first type, an interaction with promoters sensitive to temperature could exist while for the second type, it is assumed that a thermal degradation of products involved in the transposition might occur. For the mariner element some temperature effects have already been mentioned by Garza et al (1991). Such effects were observed in a D melanogaster transformed line in which the germinal excision of the non-autonomous peach element was controlled by the active element Mosl. In this line, the excision rate varied from 10.7% at 18°C to 26.4% at 29°C. In D simulans the same phenomenon was observed and the germinal excision rate of the peach element ranged from 0.2% at 18°C to 3.4% at 25°C.
In the present work, we investigated the temperature effect on the somatic excision rate of an inactive element (the peach element inserted in the white gene) when induced by active elements present in isofemale lines of 3 natural populations of D simulans. The general occurrence of active mariner elements was previously reported in these natural populations (Capy et al, 1990). A relationship between the average number of copies per line and the level of somatic excision was also analysed. We found a significant temperature effect for 2 of the 3 populations tested and a slightly significant overall correlation between the number of copies and the excision rate. Finally, an analysis of mariner sequence in 5' of the initiation site showed a 14-bp sequence with 50-57% of homology with promoter sequence of several heat shock protein (hsp) genes.

Natural populations and breeding temperature
The 3 populations used in this work came from Loua (Congo), Agadir (Morocco) and Bordeaux (France). Season and average temperature at the time of capture are the following: for Loua, November 1989, at the end of the dry season and the begining of the wet season, the average daily temperature was ! 26-27°C; for Agadir, spring 1989 with an average daily temperature of 21°C and Bordeaux, autumn 1989, more precisely during the vintage season and an average temperature of 17°C. Flies were caught by attractive fermenting traps or directly by sweeping on natural breeding sites. Isofemale lines were initiated from each population (25 for Loua, 22 for Agadir and 31 for Bordeaux) and < 5 generations were kept at 25°C in the laboratory conditions before their analysis. The isofemale line method was used to keep most of the original variability of the sample and to reduce the selection effect of laboratory conditions (Capy, 1987). The experiments were performed at 3 breeding temperatures: 17, 25 and 29°C.

Somatic excisions
The transposable element mariner can be excised either somatically or in the germline. Somatic excisions are phenotypically observed when an active element determines the excision of the non-autonomous element peach inserted in the white gene (mutation white peach, w!'°h). In this case, flies with mosaic eyes are observed.
A mosaic eye is a white peach phenotype with 1 or several red spots (see, for example: Bryan and Hartl, 1988;Hartl, 1989). Each red spot corresponds to an excision event, and the size of a red spot is related to the excision time during the development. To quantify the rate of excision in a single individual, 4 classes of mosaic flies (Mos-0, Mos-1, Mos-2 and Mos-3) were defined according to Capy et al (1990s) ; see also table I for a definition of the classe.

Estimation of the somatic excision rate
The somatic excision rate, in each isofemale line, was estimated after a test-cross between 5 males of the line tested and 5 females of the GB1 strain (figure 1). The latter strain of D simulans, built by Glenn Bryan (Bryan and Hartl, 1988), contains a single inactive element (the peach element) inserted in the white gene. This element was introduced in D simulans from the white peach strain of D mauritiana, after an interspecific cross followed by several backcrosses. The GB1 strain is extremely stable and no revertant has been observed. The somatic excision rate was estimated by the ratio of mosaic males observed in the F 1 of the test-cross over the total number of F, males examined, ie at least 10 males/line but when possible 50 males/line.

Southern blots
For each isofemale line, DNA of 25-30 individuals was prepared, as described by Junakovic et al (1984), completely digested with the restriction enzymes BamHI and HindIII, which do not cut in the element, and loaded in a single lane of a 0.8% gel agarose. To detect all elements (complete and deleted), filters were then hybridized with a mixture of the SspI-Xhol and of the Xholl-Nhel fragments of mariner (see . These 2 fragments cover 1.1 kb of the total element (total length of the element = 1.286 kb). These probes were labelled by nick translation according to Maniatis et al (1982). Hybridization and washing procedures were as Junakovic et al (1984). Agadir. The differences between the total PMF at 17 and 29°C are statistically significant only for the first 2 populations. Concerning Agadir, the PMF seems to be independent of the temperature. Such a result can also be due to the method used to quantify the amount of independent excisions in each individual. With this method, it is almost impossible to discriminate between individuals having a high level of somatic excision, since all of them are grouped in the Mos-3 class. In other words, it is possible that a temperature effect still existed in the Agadir population but such effect could not be quantified because the basic excision rate, in this population, remained high whatever the breeding temperature. Table II shows that in Bordeaux and Loua both temperature and isofemale line effects were detected, but not in Agadir. Again, a more detailed analysis, class by class, showed that the temperature effect was higher in population in which the main class at 17°C was the Mos-1, ie Loua and Bordeaux. In these populations, the average number of Mos-1 individuals decreases at 25°C and 29°C, while the average number of Mos-2 and Mos-3 individuals increases.

Average number of copies per isofemale line
The average number of copies (ANC) per isofemale line was estimated by counting the number of bands on Southern filters (fig 2). This average number includes both active and inactive copies. A large polymorphism was observed both within and between populations. In each population, it is possible to find some isofemale lines with a high or a small number of copies (see fig 2). For instance, for Agadir, line 21 presents an ANC of 17 elements while a single element was detected in line 12. The same phenomenon can be observed for the 2 other populations but the highest ANCs per line were observed for Agadir. On the other hand, it must be stressed that no line was found to be totally free of mariner, and that the average number of copies were under-estimated since co-migration of bands and degradation of high molecular weight bands may occur.
In this analysis 30,22 and 21 isofemale lines were tested for Bordeaux, Loua and Agadir, respectively, and the ANC per line are 7.9 ±1.2 for Agadir, 5.2 t 0.5 for Bordeaux and 4.7 ! 0.6 for Loua. The classification of populations is the same as regards the PMF and the ANC, ie Agadir, Bordeaux and Loua, from the highest to the lowest values. Within each population, a Spearman rank correlation test between the PMF and the ANC of the different lines for each breeding temperature, showed only 1 significant correlation. Seven out of 9 coefficients were positive, but not all coefficients were independent since the numbers of mariner copies per line were used for the 3 temperatures. Therefore, although some tendencies are suggested, it remains difficult to conclude that a relationship exists between the excision rate and the number of copies in wild populations, and such a result has to be confirmed from a larger set of data.

DISCUSSION AND CONCLUSION
Our results show that in some natural populations of D simulans: 1) there is a temperature effect on the somatic excision rate of an inactive copy of the manner transposable element, 2) this temperature effect may vary from one population to another and 3) the between-population variability of the excision rate measured by the PMF could be related to the average number of mariner copies per population.
The relationship between the excision rate and the breeding temperature suggests that a promoter acting on the transposase production is sensitive to temperature. This could be an internal promoter of the transposase, included in the transposable element itself, or an external promoter, close to the insertion point of the transposable element. In this respect, the role of the genetic environment seems to be an important component of mariner element activity as already demonstrated by the position effects frequently observed (Garza et al,'1991;Medhora et al, 1991).
A comparison between the mariner sequence in 5' of the initiation site and the promoter sequence of several heat shock genes was made (table III). Homology of 57% between a 14-bp sequence in the inverted repeat of the manner transposable element and a putative promoter of hsp-70 was detected (see: Strand and McDonald, 1985; and references therein). Homologies of 50% were also observed with other sequences of heat shock promotors and with a 14-bp sequence of the copia transposable element (sequence localized in position &mdash; 110 bp from the transcriptional start, ie in the LTR of the element). For this latter element, a temperature effect on the transcription rate was also described (Strand and McDonald, 1985). In the case of mariner, it is thus possible that the thermosensitivity of the transcription rate is an intrinsic property of its 5' inverted repeat.
Another mechanism which could explain a relationship between the breeding temperature and an excision rate would be a thermal degradation of a product involved in the repression of the element mobility. In mariner, several position effects have been described ), but we do not know whether there is a positive or a negative effect of the genetic environment on the transposase production by autonomous elements. In case of a negative effect, a thermal sensitivity of a product responsible for such a repression should result in an increase of transposase production and in an increase of excision rate. For the moment, data are available on neither the transposition nor regulation mechanisms of the mariner element to test this hypothesis.
In the 3 populations investigated, our results suggest that a relationship exists between the average number of copies and the activity level of the mariner elements measured by the excision rate. Although the presence of active elements is these populations is confirmed, it is however not possible to estimate the number of active copies in each population and each isofemale line.
The activity of the mariner element may vary according to its own sequence and to its genomic position Medhora et al, 1991;Capy et al, 1992). Therefore, a given excision rate can be due to a single element highly active or to several elements with a low activity. In this latter case, a dosage effect could exist. Such a dosage effect, for the mariner element, has already been stressed by Garza et al (1991). In the present work, it is likely that the transposable production is higher in Agadir than in Bordeaux and Loua. In other words, it is possible that Bordeaux and Loua were genetically more stable than Agadir at the moment of their sampling.
The variability of the temperature effect observed between the three natural populations raises the following question: does this variability correspond to genetic drift between geographically distant populations or to different local adaptations?
Very few data are available on transposable elements in natural populations originating from different biogeographic areas. For numerous genetical traits, like morphological traits or enzymatic polymorphism, it seems that in D melanogaster (a sibling species of D simulans), an influence of the average temperature is likely for the occurrence of latitudinal clines (Lemeunier et al, 1986;David and Capy, 1988). D simulans, on the other hand, generally exhibits a lower geographic differentiation, although regular latitudinal variations have been observed in several cases (Parsons, 1983;Lemeunier et al, 1986;Hoffmann and Parsons, 1991). Therefore, such a climatic factor is able to induce genetic variations among natural populations of cosmopolitan species. To detect a variability of temperature effect on excision rate of transposable elements according to the average temperature of the geographic area of natural populations, many more populations should be investigated.
In conclusion, it seems that temperature is one of the environmental factors able to induce the somatic transposition of the mariner element in natural populations of D simulans. However, populations with high basic excision rate could remain insensitive to this factor.
For the moment, the data available are not sufficient to check this conclusion.
But, it must be stressed again that for Drosophila, temperature is a factor which shows seasonal and daily variations in several places. Therefore, it is possible that temperature can induce genomic stresses strong enough to stimulate mobilisation and rearrangement of transposable elements. If so, the reasons and the mechanisms of these remain to be investigated.