Background

Apoptosis, a form of programmed cell death, is a process needed for normal development and maintenance of tissue homeostasis in multicellular organisms [1, 2]. Cells undergoing apoptosis show characteristic changes, such as chromatin and cytoplasm condensation, chromosomal DNA fragmentation, breaking up of nuclei and then of cells into fragments called apoptotic bodies [3, 4]. Individual cells apoptose, while the neighboring cells remain undamaged [3, 4]. Apoptosis is a complex process whereby a proteolytic cascade of caspases is activated in cells [5].

The occurrence of apoptosis is a feature of female germline development common to vertebrate and invertebrate species [6, 7]. In the Drosophila melanogaster ovaries, there are two checkpoints where programmed cell death occurs. One is in the germarium (region 2a/2b), where apoptosis probably regulates the proper ratio of germline cells to follicle cells [8]. The other checkpoint is located in the vitellarium (stages 7-8 of oogenesis) [9]. The number of egg chambers undergoing apoptosis increased in D. melanogaster fed a diet lacking protein [8], under the effect of 900-MHz and 1800-MHz radiation [10], and after exposure to chemical agents [11]. The normal development of mature egg is consistently associated with apoptosis of 15 nurse cells in the egg chamber [12]. It is noteworthy that apoptosis and autophagy coexist at all the above mentioned stages of oogenesis in D. melanogaster [13, 14].

It has been also hypothesized that the apoptotic process had a symbiotic origin [15]. In terms of the endosymbiotic theory, mitochondria, which play a major role at the early stages of apoptosis, evolved from the free-living prokaryotes [5]. One of the symbionts may be involved in the regulation of apoptosis in partner cells. To illustrate, extracellular parasites, particularly such worms as filarial nematodes, schistosomes and the cestode Taenia crassiceps, are able to induce apoptosis in host immune cells [16]. Bacterial pathogens (Chlamydia, Neisseria, Legionella pneumophila) can either block or induce apoptosis in host cells, depending on the stage of infection [17, 18]. At the early stage of infection, bacteria replicate in the host cell, using different mechanisms to prevent apoptosis. At the late stages of infection, the bacteria induce apoptosis in the host cell, thereby facilitating egress and ensuring infection of neighboring cells.

Wolbachia associated with various hosts in which it manipulates viability and reproduction causing parthenogenesis, feminization, male killing and cytoplasmic incompatibility, provides a unique model for studying mechanisms of symbiont interactions [19, 20]. The Wolbachia strain wMel is widely spread in natural populations of D. melanogaster [21, 22]; in contrast, wMelPop has been detected in a laboratory stock of D. melanogaster [23]. It is possibly not encountered in nature. In D. melanogaster, the wMelPop strain reduces lifespan, proliferating widely in the brain, muscle and retina cells [23]. In certain insect species, the presence of Wolbachia is required for oogenesis [24]. Removal of the Wolbachia strain wAtab3 from parasitic wasps Asobara tabida resulted in copious apoptosis of the egg chambers in the ovarioles and led to sterility [25]. The mechanisms whereby the endosymbiont Wolbachia impacts apoptosis in host cells have been poorly studied. Preferential infection and high accumulation of Wolbachia in region 2a of the germarium [26] where the checkpoint is located in Drosophila was thought-provoking. We raised the question: Can bacteria Wolbachia in region 2a of the germarium affect the frequency of apoptosis there? Using fluorescence and transmission electron microscopy we compared germaria from ovaries of two D. melanogaster stocks infected with either the wMel or wMelPop strains with germaria from two uninfected counterparts. It was established that the presence of wMel did not alter apoptosis frequency in germaria from D. melanogaster Canton S. In contrast, the number of germaria containing apoptotic cells in the checkpoint was considerably increased in the wMelPop-infected flies as compared with their uninfected counterparts. Thus, evidence was obtained indicating that the virulent Wolbachia strain wMelPop has an effect on the fate of germline cells during D. melanogaster oogenesis.

Results

Frequency of apoptosis in germaria from ovaries of the uninfected and Wolbachia-infected D. melanogaster

Two parts are distinguished in the Drosophila ovariole: the germarium made up of four regions (1, 2a, 2b, 3) and the vitellarium (Figure 1A, B) [27, 28]. The region 2a/2b, where apoptosis can occur, contains 16-cell cysts, somatic stem cells (SSCs), which contact with the somatic stem cell niche (SSCN) and follicle cells (Figure 1B). Cell death in this region of the germarium was detected by two methods, acridine orange (AO)-staining and TUNEL assay. Fluorescence microscopy of AO-stained ovarioles demonstrated that apoptotic cells were located as large yellow or orange spots in region 2a/2b of the germarium from D. melanogaster (Figure 2A, C, E, G). Germaria containing no apoptotic cells fluoresced homogeneous green (Figure 2B, D, F, H). It should be noted that wMel- and wMelPop-infected flies, besides bright spots in region 2a/2b (Figure 2C, G), showed weak punctuate fluorescence both in regions 2a/2b and 1 of the germarium (Figure 2C, D, G, H). Such fluorescent puncta were not observed following TUNEL, thereby indicated that they were not caused by apoptosis.

Figure 1
figure 1

A schematic representation of an ovariole of D. melanogaster . A, an ovariole of D. melanogaster consisting of the germarium (g) and the vitellarium. B, a detailed scheme of the germarium structure composed of regions 1, 2a, 2b, 3. The checkpoint is framed (red). C, a 16-cell cyst; SSCN, a somatic stem cell niche; SSC, a somatic stem cell; FC, a follicle cell.

Full size image
Figure 2
figure 2

Visualisation of acridine orange (AO)- and TUNEL-stained germarium cells of D. melanogaster . A, C, E, G, germaria containing apoptotic cells in region 2a/2b from 5 day-old uninfected (A, E) and Wolbachia-infected (C, G) females (AO staining). B, D, F, H, germaria not containing apoptotic cells from the same fly stocks (AO staining). Arrows indicate small punctate AO-staining in regions 1 and 2a/2b (C, D, G, H). I, relative proportion of germaria containing apoptotic cells from ovaries of the uninfected (w1118T, Canton ST) and Wolbachia-infected (w1118, Canton S) flies. The total number of examined germaria is indicated by blue number; bars show the average percentage per experiment ± s. e. m. J, L, germaria containing apoptotic cells in region 2a/2b in the wMelPop- and wMel-infected fly stocks, respectively (TUNEL). K, M, germaria not containing apoptotic cells from the same fly stocks. Region 2a/2b of the germarium is indicated by red brackets. Scale bars: 20 μm.

Full size image

The percentage of germaria containing apoptotic cells was 41.8±4.1% in the uninfected D. melanogasterw1118T maintained on standard food, whereas it increased to 70.6±5.3% in the wMelPop-infected flies (Figure 2I). Analysis performed with the wMel-infected D. melanogaster Canton S revealed no significant differences from their uninfected counterparts (Figure 2I, Table 1).The next step was to exclude the possible effect of insufficient nutrition on the current results. To do so, we conducted experiments in which flies were raised on rich food source taking into account that it decreases the number of germaria containing apoptotic cells [8, 29]. We found that rich food causes a decrease in the relative proportion of apoptotic germaria in both w1118T and w1118 flies; however, the difference between these two groups was significant (Figure 2I, Table 1). The percentage of germaria containing apoptotic cells did not change under the effect of rearing D. melanogaster Canton S on different food. Based on analysis of apoptotic cell death by TUNEL, three groups of germaria were distinguished: TUNEL-negative, TUNEL-positive with 1-2 distinct puncta in region 2a/2b and TUNEL-positive with a cluster of bright spots (Additional file 1). There was no evidence for variation in the frequency of apoptosis between wMel-infected (Canton S) and uninfected (Canton ST) flies (Table 2; χ2=1.3, df=1, P=0.25); however, there was evidence for a difference in the frequency of apoptosis between the w1118T and w1118 flies (Table 2; χ2=25.3, df=1, P<0.0001). The total percentage of germaria containing apoptotic cells in D. melanogaster agreed well with the one obtained with AO-staining. Thus, TUNEL confirmed the results of AO-staining.

Table 1 Details of statistical analysis (two-way ANOVA)
Full size table
Table 2 Quantification of TUNEL staining in region 2a/2b of the germaria
Full size table

Ultrastructure of germaria from ovaries of the uninfected and the Wolbachia-infected D. melanogaster

For an ultrastructural analysis of germarium cells, we first chose under the light microscope those longitudinal sections that enabled us to define region 2a/2b of the germarium (Figure 3A, B, red brackets). Cyst cells in region 2a/2b were interconnected by ring canals and consisted of nuclei that exhibited numerous invaginations, protrusions, and cytoplasm rich in organelles (Figure 3C, D, Additional file 2). Our ultrastructural data for germarium cells of the uninfected and the Wolbachia-infected flies allowed us to identify cysts in region 2a/2b showing characteristic features of apoptotic death (Figure 4 and Additional file 3). The cytoplasm was more electron-dense in such cystocytes, some mitochondria became markedly swollen (Figs. 4A and Additional file 3A). The matrix of mitochondria was light and just a few small cristae were discerned at the periphery (Figs. 4B and Additional file 3B). We observed also cells with electron-dense cytoplasm, which had lost contact with their neighboring cells (Additional file 3C). In such cells, chromatin appeared condensed in apoptotic nuclei and the lumen of the nuclear envelope was dilated (Figs. 4C and Additional file 3C). At the last stage of apoptosis, cells disaggregated into large and small fragments, or apoptotic bodies, with characteristic electron-dense cytoplasm containing ribosomes, endoplasmic reticulum membranes, and frequently intact mitochondria (Figs. 4D and Additional file 3D).

Figure 3
figure 3

Visualisation of germarium cells in semi-thin and ultra-thin sections. A, B, longitudinal semi-thin sections of germaria stained with methylene blue. C, D, ultrastructure of cyst cells from the uninfected and the wMelPop-infected flies. Arrows point to bacteria; arrowheads denote ring canals between neighboring cells. Scale bars correspond to 10 μm (A, B) and 2 μm (C, D), respectively.

Full size image
Figure 4
figure 4

Morphology of apoptotic cystocytes in region 2a/2b of the germarium from the wMelPop-infected D. melanogasterw1118. A, swollen mitochondria (black arrows) in the cytoplasm of cyst cells. White arrows indicate bacteria. B, a fragment of a cyst cell with two mitochondria: one is normal, the other is swollen with the matrix of low electron density and the disintegrated cristae. C, a cyst cell, the cytoplasm appears dense, the nucleus is pyknotic. D, apoptotic bodies (ab) containing intracellular organelles. Scale bars: 1 μm.

Full size image

Analysis of germarium cystocytes of wMel- and wMelPop-infected flies showed that individual bacteria were distributed throughout all the cytoplasm, occasionally occurring as small groups (Figs 3D and Additional file 2). Large accumulations of adjacent bacteria were observed in apoptotic cyst cells (Figure 5A, B, Additional file 4). In these masses, the bacteria varied in morphological appearance (5C, D and Additional file 4B). Some endosymbionts showed normal ultrastructural features: a three-layered envelope, a matrix with many ribosomes and dispersed chromatin. In contrast, most bacteria were surrounded by a three-layered envelope, the matrix was of low electron density with a few ribosomes. Disrupting bacteria were also encountered. These were not enclosed by an envelope, their matrix was loose, light, devoid of ribosomes. The follicle cells surrounding the cysts in region 2b of the germarium showed a normal morphology and low levels of Wolbachia with normal structure (Additional file 5).

Figure 5
figure 5

Ultrastructure of the Wolbachia strain wMelPop in apoptotic cystocytes in region 2a/2b of the germarium. A, B, Wolbachia accumulations in apoptotic cyst cells, low magnification view. C,D, bacteria framed in panels A, B depicted at higher magnification. Bacteria showing normal morphology (arrows), bacteria with matrix of low electron density (white arrowheads), bacteria with matrix of low electron density and disrupted cell wall (black arrowheads) in the cytoplasm of dying cysts. Scale bars: 2 μm.

Full size image

At the periphery of the germarium, fragments of degrading cells were frequently seen in region 1, precisely where AO-staining of the germaria from the Wolbachia-infected flies was punctate (Figure 2C, D, G, H). These fragments were filled with multilayered membranes, nuclear remnants, mitochondria, and bacteria with normal and abnormal morphology (Figure 6A-C, Additional file 6). The cell organelles and bacteria were often engulfed by autophagosomes. Besides bacteria with light matrix, like those in apoptotic cysts (Figure 6C, D), the autophagosomes occasionally enclosed electron-dense bacteria-like structures 0.2-0.3 μm in diameter (Figure 6D, E) or similar smaller ones (Figure 6F). At the periphery of the germaria, autophagosomes containing individual bacteria with normal morphology were observed (Figure 6G).

Figure 6
figure 6

Ultrastructure of the germarium cells at the periphery of region 1 in wMelPop-infected D. melanogasterw1118 . A, a fragment of region 1 of the germarium, low magnification view. Normal cells and two fragments of cells (brackets), whose cytoplasm is filled with autophagosomes, bacteria and multilayered membranes. B, multilayered membranes and fragments of a disintegrated nucleus (white arrowhead). C, a fragment of a cell with electron-dense cytoplasm containing Wolbachia of two types: one normal (black arrows), the other with matrix of low density (white arrows). D, electron-dense bacteria-like structures engulfed by autophagosome. E, higher magnification of the bacteria-like structure framed in panel D. F, an autophagosome containing electron-dense structures and vesicles . G, autophagosomes enclosed individual bacteria. Arrowheads indicate autophagosome membranes. Scale bars: 1 μm.

Full size image

Discussion

This is, to our knowledge, the first study that demonstrated by using AO- and TUNEL staining an increase in the frequency of apoptosis in the germarium checkpoint in wMelPop-infected D. melanogasterw1118. This increase is possibly caused by the specific effect of the Wolbachia strain wMelPop, since it was not observed in wMel-infected D. melanogaster Canton S. Our current electron microscopic observations allowed us to identify changes in Wolbachia morphology in apoptotic germline cells.

Morphological evidence of apoptosis in germarium cells

The ultrastructural features of apoptosis in the cyst cells of higher eukaryotes have gained wide recognition. They include cytoplasmic and nuclear condensation (pyknosis); nuclear fragmentation (karyorrhexis); normal morphological appearance of cytoplasmic organelles; an intact plasma membrane [3, 4]. The ultrastructural changes we identified here in D. melanogaster cyst cells are consistent with the above hallmarks. Furthermore, we revealed mitochondria of two types: intact morphology in one type and markedly swollen with a few cristae in the other. A similar heterogeneity of mitochondrial ultrastructure has been observed during apoptosis in granulose cells of Japanese quail (Coturnix coturnix japonica) [30], lymphocytes from leukemia patients [31], and megakaryocytes from patients with idiopathic thrombocytopenic purpura [32]. It has been suggested that the swollen mitochondria release cytochrome c, which activates a cascade of proteolytic reactions, while the normal ones retain their capacity for ATP synthesis, a process apoptosis requires [30, 31, 33]. According to our qualitative analysis using EM, morphological evidence of apoptosis was revealed in germline cells from uninfected flies and those infected with wMel and wMelPop. Thus, there are reasons for inferring that the endosymbiont Wolbachia in D. melanogaster cystocytes has no effect on sequential passage of intracellular organelles through apoptosis. To reveal the possible differences between the effect of the wMel and wMelPop strains on apoptosis in the germaria, additional morphometric analysis of the number of apoptotic structures and of Wolbachia density in the cystocytes is required.

Structural features of Wolbachiain apoptotic cysts

Wolbachia with matrix of moderate and low electron density in apoptotic cells in region 2a/2b of the germarium have been previously encountered in other types of D. melanogaster ovaries [34] and they presumably reflect different functional states of bacteria. Wolbachia with disrupted envelopes and light matrix are possibly dying bacteria in apoptotic cells. Such appearance has not been observed in Wolbachia injured or killed by heat stress [35] and tetracycline [36]. The electron-dense bacteria-like structures at the periphery of region 1 of the germarium may be evidence of changes in dying Wolbachia. Large masses of structures of this kind resembling the bacteria endospores have been found in the brain cells of the wMelPop-infected D. melanogasterw1118[23]. In our view, the electron-dense structures, which we revealed at the periphery of region 1 of the germarium, are presumably autophagosome encapsulated dying Wolbachia. A supporting line of evidence came from Wright and Barr [37], who on the basis of their observations on degenerating germaria cysts from mosquitoes Aedes scutellaris suggested that these structures represented degenerating Wolbachia.

Cell fragments containing dying bacteria and autophagosomes and appearing as numerous smaller puncta in regions 2a/2b and 1 of the germarium may represent autophagy, not apoptosis. This appears plausible when recalling that AO stains not only apoptotic cells, also lysosomes [38]. TUNEL did not reveal such puncta in these regions.

The possible role of the Wolbachiastrain wMelPop in programmed cell death in region 2a/2b of the germarium

Our current estimates of apoptosis in region 2a/2b of the germarium from the ovaries of the uninfected D. melanogasterw1118T raised on standard food are consistent with those reported elsewhere [14]. It is of interest that apoptosis level in the germaria decreased in D. melanogasterw1118T, but not in D. melanogaster Canton ST after transfer to rich food. This may be indicative of differences in sensitivity to changes in food composition between different fly stocks. AO- and TUNEL staining demonstrated that the virulent Wolbachia strain wMelPop increased the percentage of germaria containing apoptotic cells in D. melanogasterw1118 ovaries, while wMel strain was without such an effect. The effect of wMelPop on cystocytes in ovaries was observed in flies maintained on standard and rich food. Evidence was provided that the effect of Wolbachia on D. melanogaster is not general, being rather specific to the pathogenic strain wMelPop.

What pathways may be envisaged for the Wolbachia strain wMelPop caused increase in the number of germaria whose cysts undergo apoptosis? On the one hand, bacteria may have a direct effect on germline cells (Figure 7A, B). In fact, one of 16 cyst cells becomes the oocyte, the other 15 differentiate into nurse cells in region 2a of the germarium. This is associated with transport of 15 centrioles into the pre-oocyte, where the microtubule-organizing center forms [39, 40]. Wolbachia distribution is dependent upon microtubules during oogenesis and bacteria show mislocalization in the egg chambers treated with colchicine which causes depolymerization of microtubules [41]. Evidence has been obtained indicating that Wolbachia are evenly distributed throughout the oocyte and nurse cells during stages 1-2 of oogenesis, becoming concentrated at the oocyte anterior during stages 3-6 [41]. With this in mind, the high levels of Wolbachia in cystocytes during differentiation into oocyte and nurse cells in region 2a of the germarium may possibly lead to impairment at the structural and/or molecular level, the cyst may undergo apoptosis as a consequence (Figure 7B).

Figure 7
figure 7

Pathways along which Wolbachia may affect egg chamber formation in region 2a/2b of the germarium. A, localization of regions in the germarium (framed) where the bacteria may interfere with normal function of cells. B, the bacteria disturb the differentiation of cystocytes (white) into the oocyte (light orange) and the nurse cells (light violet). C, the bacteria skew the proper ratio of germline cells to follicle cells. Crescent shape, SSCN; green circle, SSC; green ovals, follicle cells. Red points represent the bacteria.

Full size image

On the other hand, the increase in the number of germaria containing apoptotic cysts may result from the action of the bacteria on the SSCs, which gives rise to follicle cells in region 2b of the germarium (Figure 7A, C). Drummond-Barbosa and Spradling [8] have suggested that apoptosis in region 2a/2b of the germarium serves to maintain the proper ratio of germline cells to somatic follicle cells. In poorly fed flies, follicle cells slow down their proliferation, the germline cells to somatic follicle cell ratio becomes skewed, resulting in cyst apoptosis in region 2a/2b which corrects this ratio [8]. It has been established that stem cells are maintained in specialized microenvironment called the niche [42]. The abundance of Wolbachia in the SSCN [26] is of interest in this context. Thus reasoning, it may be assumed that the presence of Wolbachia in the SSCN decreases the SSC proliferation rate, the ratio of germline cells to follicle cells becomes imbalanced and, as a consequence, cysts undergo apoptotic death. Judging from our current data, the ultrastructural appearance of follicle cells in region 2b of the germarium from ovaries of wMelPop-infected D. melanogasterw1118 was normal, thereby indicating that Wolbachia presumably did not negatively affect follicle cells. It should be noted that the fecundity of the wMelPop infected D. melanogaster w1118 was not decreased as compared with their uninfected counterparts [43, 44]. This was evidence of insect plasticity, rendering them capable to adapt to diverse factors.

Taken together, our findings clearly demonstrated that the Wolbachia strain wMelPop has an effect on the egg chamber formation in the D. melanogaster germarium. However, the underlying mechanism is still unclear. We intend to perform a comparative morphometric analysis of apoptotic structures and bacteria in cystocytes of wMel- and wMelPop-infected flies. The results would be helpful in deciding whether the increase in apoptosis frequency is due to high bacterial density or to particular pathogenic effect of the Wolbachia strain wMelPop on female germline cells.

Conclusions

The results of this study showed that the presence of the Wolbachia strain wMelPop in D. melanogaster ovaries led to an increase in the frequency of apoptosis in the germarium checkpoint. Two possible pathways along which Wolbachia affect egg chamber formation in region 2a/2b of the germarium have been suggested. Future research should be conducted to clarify the mechanism underlying this phenomenon.

Methods

Drosophilastocks and maintenance

The Drosophila melanogaster Canton S infected with the Wolbachia strain wMel (IC&G, Russia) and D. melanogasterw1118 infected with wMelPop (a kind gift from prof. S. O’Neill, The University of Queensland, Australia) were used in these experiments. Flies were maintained at 25 °C either on a standard yeast-agar medium or on daily replaced rich food (standard medium covered with wet yeast paste). To obtain uninfected D. melanogasterw1118T, flies were raised on food supplemented with tetracycline at 0.03% for two generations, then on standard food for more than three generations [43]. Confirmation of the infection status of each stock was provided by PCR. For this purpose, total DNA extracted from fly ovaries and wsp 81F/wsp 691R primers for amplifying a Wolbachia surface protein gene fragment were used [45].

Acridine orande staining

Acridine orange (AO), a vital stain highly specific to apoptotic nuclei, was used [46]. Ovaries were dissected from 5-day old flies in EBR buffer (130 mM NaCl, 4.7 mM KCl, 1.9 mM CaCl2, 10 мM Hepes pH 6.9), stained with AO (Merck), 5 μg/ml, in 0.1 M sodium phosphate buffer, pH 7.2, for 3 min at room temperature [12, 47]. Samples were placed onto glass slides and covered with halocarbon oil (KMZ Chemicals Ltd.). They were viewed under an Axioscop 2 plus fluorescence microscope (Zeiss) using an appropriate filter (Zeiss filter set 02). Time elapsed from dissection to the end of viewing was restricted, 20 min. Staining of nuclei varied from bright yellow to brilliant orange, depending on the stage of degeneration [46]. The percentage of AO-staining germaria was expressed as the ratio of the number of AO-stained germaria containing apoptotic cells to the total number of analysed germaria. Three experiments were performed for each of the 4 D. melanogaster groups (w1118, w1118T stocks, standard food; w1118, w1118T, rich food). In each replicate, ovaries were dissected from 6 flies, 7-12 germaria per fly were analysed. In all, about 1350 AO-stained germaria were analysed. Bartlett’s test was used to check homogeneity of variances. Two-way ANOVA was used to determine the significance of the difference between the frequency of apoptosis of the uninfected and Wolbachia-infected flies maintained on different food.

TUNEL assay

TUNEL was the independent assay of detection of apoptotic cells. TUNEL is advantageous because preferentially labeling apoptotic cells relatively late in the apoptotic process [48]. Ovaries were dissected from 5-day old flies in phosphate-buffered saline (PBS), fixed in PBS containing 4% formaldehyde plus 0.1% Triton X-100 for 25 min. Then, they were separated into individual ovarioles, rinsed briefly in PBS twice and washed in PBS three times for 5 min each. Ovarioles were made permeable with 20 μg/ml proteinase K in PBS for 20 min at room temperature, this was followed by 3 washes in PBS for 5 min each. The TUNEL reaction and all the subsequent steps were performed using the FragEL DNA Fragmentation Detection Kit (Calbiochem) according to the manufacturer’s protocol. Samples were viewed with an Axioscop 2 plus fluorescent microscope (Zeiss), images were captured with a high resolution microscopy camera AxioCam HRc and AxioVision software. Germaria from ovaries of 10 flies were counted in each of the 4 groups. The total number of germaria analysed was about 850. The data were compared using a Chi-square test (χ2).

Electron microscopy

Fixation of the D. melanogaster ovaries was carried out using the method described previously [49, 35]. Briefly, 5 day-old females were dissected in 0.1 M phosphate buffer, pH 7.4, fixed in 2.5% glutaraldehyde (Sigma) in 0.1 M sodium cacodylate buffer, pH 7.4, for 2.5 h. This was followed by washings in the same buffer and postfixation in 1% OsO4 and 0.8% potassium ferrocyanide for 1 h. After washings, samples were placed in 1% aqueous solution of uranyl acetate (Serva) for 12 h at 4 °C. Then they were dehydrated in ethanol series and acetone, finally samples were embedded in Agar 100 Resin (Agar Scientific Ltd.). Ultra-thin sections were stained with uranyl acetate and Reynolds lead citrate. They were examined with a transmission electron microscope (JEM 100 SX, JEOL). The number of flies analysed in each of the 4 groups was 8-12.