Wolbachia Bacteria Reside in Host Golgi-Related Vesicles Whose Position Is Regulated by Polarity Proteins

Wolbachia pipientis are intracellular symbiotic bacteria extremely common in various organisms including Drosophila melanogaster, and are known for their ability to induce changes in host reproduction. These bacteria are present in astral microtubule-associated vesicular structures in host cytoplasm, but little is known about the identity of these vesicles. We report here that Wolbachia are restricted only to a group of Golgi-related vesicles concentrated near the site of membrane biogenesis and minus-ends of microtubules. The Wolbachia vesicles were significantly mislocalized in mutant embryos defective in cell/planar polarity genes suggesting that cell/tissue polarity genes are required for apical localization of these Golgi-related vesicles. Furthermore, two of the polarity proteins, Van Gogh/Strabismus and Scribble, appeared to be present in these Golgi-related vesicles. Thus, establishment of polarity may be closely linked to the precise insertion of Golgi vesicles into the new membrane addition site.


Introduction
Wolbachia pipientis are maternally inherited symbiotic bacteria that are widespread among most insects including laboratory stocks of Drosophila melanogaster, as well as filarial nematodes and crustaceans [1][2][3][4]. Wolbachia belong to the Richettsial family responsible for the deadly human diseases such as typhus, Rocky Mountain spotted fever, and Q fever, but themselves are not involved in any known human diseases [5]. Wolbachia bacteria are best known for their ability to induce reproductive alterations in hosts such as male killing, feminization, parthenogenesis, and cytoplasmic incompatibility, all of which result in increased number of infected female offspring and thereby helping vertical transfer of Wolbachia [6]. These reproductive alterations may promote speciation in extreme cases. Because of these intriguing properties, Wolbachia have been extensively studied for entomology, agriculture and evolution.
Despite Wolbachia's unique role in host reproduction and physiology, their underlying cellular mechanisms are yet to be addressed. Studies with electron microscopy have revealed that Wolbachia bacteria are strictly present in vesicular structures in the cytoplasm of host cells [7,8]. These Wolbachia vesicles are attached to astral microtubules near centrosomes by short electron-dense bridges, and their centrosomal localization is dependent on microtubules but not actin [7]. Wolbachia bacteria are enclosed within three layers of membranes: the outer layer is host origin and two inner layers are bacterial cell wall and bacterial plasma membrane [9]. Since parasitic bacteria and enveloped mammalian viruses often utilize a variety of subcellular organelles such as endoplasmic reticulum and Golgi apparatus during their life cycles [10][11][12], Wolbachia may also be present in a host organelle that can aid the replication and propagation of Wolbachia. Identification of this host organelle is critical for understanding the Wolbachia's ability in changing host physiology.
We report here that Wolbachia reside in a group of Golgi-related vesicles. These Golgi-related vesicles distinctly localized near the site of membrane biogenesis in the embryo cortex, and appeared to contain two polarity proteins, Van Gogh/Strabismus (Vang hereafter) and Scribble (Scrib) as well as cis-Golgi GM130 protein. Furthermore, Wolbachia vesicles were mislocalized in mutant embryos defective in cell/planar polarity genes such as discs-large (dlg), Van Gogh (Vang)/strabismus (stbm), frizzled (fz) and dishevelled (dsh). These observations raise an interesting possibility that Wolbachia may mark the unique group of Golgi vesicles linked to membrane biogenesis. The additional finding that localization of Wolbachia vesicles is regulated by genes involved in cell/tissue polarity also provided a surprising new potential activity for these polarity genes in Golgi localization.

Results
It has been known that majority of fly laboratory strains is infected by Wolbachia. We have previously reported that numerous polyclonal antisera generated against fusion proteins expressed in E. coil exhibit cross-reactivity toward Wolbachia proteins in immunocytochemisitry, because of impurity in the antisera that have reactivity to E. coli proteins and also to the related Wolbachia proteins [13]. Wolbachia appear as vesicular structures with these antisera, and these false vesicular patterns can be avoided by using Wolbachia-free laboratory strains [13].
During the course of this previous study, we discovered a link between Wolbachia and Golgi-related vesicles, which is a focus of were visualized with anti-Sdt antisera. Wolbachia, recognized by antisera (blue) and DNA marker (red), appear as pink. In these images, structures appeared as blue do not contain DNA and should be considered to have endogenous Sdt (A) or Vang (B-G except D). DNAs in A, B, C and G were visualized with propidium iodide, and DNAs in D, E, and F were visualized with Toplo-3. (A) In a preblastoderm stage embryo, Wolbachia vesicles (arrows) are attached near the minus ends of the astral microtubules (bracket) but not the polar microtubules (arrowhead). (B) A blastoderm stage embryo shows Wolbachia localization near centrosome (arrow). (C) Wolbachia vesicles are perinulclear during interphase (arrows). (D) GM130 and p120 are present in separate vesicles in Wolbachia-free CS embryos during mid-cellularization. They sometimes are present in the two adjacent vesicles (arrowheads). (E,F) In Wolbachia-infected CS embryos during mid-cellularization, p120-containing vesicles are physically separated from Wolbachia vesicles, but are in proximity (arrowheads) (E). Wolbachia vesicles either are juxtaposed to GM130-containing vesicles (arrowheads) or contain GM130 proteins (arrow) (F). (G) In Wolbachia-infected CS embryos, some Wolbachia vesicles are in proximity with Hrs vesicles (arrow), but did not contain Hrs (arrowhead). Portions marked with brackets in E, F, and G are magnified in E', F' and G'. Scale bar: A,C,E,F,G, 10 mm; B, 4.4 mm; this report. To detect Wolbachia, we used three antisera, anti-Vang antisera [14], anti-Stardust (Sdt) antisera [15] and anti-Dlg antisera [16]. These antisera along with DNA markers recognized Wolbachia with higher specificity than anti-Hsp60 (cloneLK2, Sigma) or the Wolbachia-specific antibody [17] (data not shown).

Wolbachia bacteria are present in Golgi-related vesicles
Wolbachia bacteria are present in membrane-bound vesicular structures that are attached to astral microtubules near centrosomes ( [7]: Figures 1A and 1B) and are mostly perinuclear during interphase ( Figure 1C). Since such localization patterns are reminiscent of mammalian Golgi apparatus, we reasoned that Wolbachia bacteria may be present in host Golgi vesicles. To test this possibility, we utilized two Golgi markers, GM130 and p120. GM130 is a tightly associated peripheral cis-Golgi protein that is involved in Golgi ribbon formation as well as mitotic Golgi fragmentation in mammalian cells [18][19][20][21]. p120 is proposed as a fly homolog of rat MG-160, a sialoglycoprotein of the medial Golgi cisternae [22][23][24]. It has been shown that GM130 and p120 are present in the two juxtaposed, but clearly distinct vesicles in fly imaginal discs during the third-instar larval stage (See Figure 1G in [24]), suggesting that the cis-and the medial-Golgi are physically distinguishable in flies. We observed in fly embryos that GM130 and p120 were sometimes present in the juxtaposed vesicles but rarely in the same vesicle, indicating that the cis-and medial Golgi units are spatially separated from each other in both embryos and larvae ( Figure 1D).

Wolbachia vesicles are concentrated near the site of membrane biogenesis
Previous studies have shown that Wolbachia vesicles are concentrated in the cortical layer and also scattered in the entire cytoplasm of newly laid embryos. As the embryo further develops to syncytial blastoderm stage, most Wolbachia vesicles become localized to the cortex along with nuclei and centrosomes that have migrated from the embryo interior to the cortex [7,8]. During the subsequent cellularization stage, we found that Wolbachia vesicles became more narrowly concentrated in the sub-apical region of the cortex (Figures 2A, 2B and 2C): approximately 80% of Wolbachia vesicles (2406/2960) were concentrated in the 5 mm span of apical region. The highest percentage (,32%) of Wolbachia vesicles (955/2960) was at ,3 mm from the apical surface of the cellularizing embryo ( Figures 2D, 2E and 2F). This region with highest percentage of Wolbachia vesicles precisely coincide with the new membrane addition site that is located in between the apical and the basolateral regions, as identified by Lecuit and Wieschaus [27]. They showed that membrane addition occurs only at the sub-apical region of plasma membrane during mid-cellularization. Further, these Wolbachia vesicles were concentrated near the newly forming cell boundary ( Figure 2G). These data raise a possibility that Wolbachia selectively reside in a special group of Golgi-related vesicles that is involved in membrane biogenesis of newly forming epithelial cells.
At the end of cellularization stage, a minor fraction of Wolbachia vesicles was found near the region between the membrane front and the growing lateral membrane (arrows in Figure 2H). This region corresponds to the basal adherens junction, whose integrity is essential for the growth of the plasma membrane [28,29]. Because Wolbachia vesicles were not found near the basal junction during the mid-cellularization when the extensive membrane biogenesis occurs ( Figure 2F), localization of Wolbachia vesicles near the basal adherens junction may be a unique feature of established epithelial cells. We therefore examined whether the epithelial cells in wing imaginal discs, another example of established epithelial cells, also have Wolbachia vesicles near basal adherens junction. As shown in Figure 2I, majority of Wolbachia vesicles was enriched sub-apically at the region ,2-6 mm from the apical surface of wing epithelial cells, but a minor fraction of them was also found near the basal adherens junction. These data show the similarity between the embryo and larval epithelial cells in terms of Wolbachia localization, and suggest that similar Golgirelated vesicles are present near the membrane addition site and the basal adherens junction. Further studies are required to reveal the role of these vesicles in the two different membrane sites.

Wolbachia-containing vesicles are mislocalized in several polarity mutant embryos
If Wolbachia are indeed present in a special group of Golgi vesicles participating in membrane growth, Wolbachia could be used as a marker for these Golgi vesicles. We have previously reported that Dlg and its partner Vang are involved in new membrane growth, in addition to their well-studied functions in apical-basal cell polarity and planar cell polarity (PCP) [14,[30][31][32][33][34]. Thus, we reasoned that Golgi vesicles involved in membrane growth might be mislocalized in dlg and Vang mutant embryos. To test this, we examined the localization pattern of Wolbachia in these mutant embryos. The temperature sensitive dlg HF321 embryos obtained from the homozygous Wolbachia-infected dlg HF321 parents were cultured at the restrictive temperature (25uC) in order to obtain partial loss of function dlg phenotype. Vang stbm-153 and Vang stbm-7-6 embryos were obtained from the crosses between homozygous mutant adults. Unlike other Vang null mutant such as Vang stbm-6 , both Vang stbm-153 and Vang stbm-7-6 are hypomorphs that can produce homozygous embryos with varying degrees of hatching rate: at 25uC, 82% of Vang stbm-153 embryos and 20% of Vang stbm-7-6 can hatch when 99% of wild type embryos can hatch [31]. Vang stbm-153 allele contains a frameshift mutation that can generate a truncated protein of 205 amino acid residues [31]. The molecular lesion of Vang stbm-7-6 has not been identified (Flybase). Interestingly, Vang stbm-153 flies can be maintained as a healthy homozygous stock, suggesting that they may contain a partially functional Vang protein. Cross-sections of these Vang embryos revealed that Wolbachia vesicles were frequently located below the cortex in these dlg HF321 , Vang stbm-153 and Vang stbm-7-6 embryos, unlike the Wolbachia vesicles in wild type embryos that were strictly present in the sub-apical region of the cortex (Figures 3A-3C).
The Vang gene is shown to genetically interact with other PCP genes such as dsh and fz, and Vang protein also physically interacts with Dsh and Fz [32,35,36]. Potential involvement of Vang in Golgi vesicle localization raised a possibility to us that other PCP D,E',F',G', 2.5 mm.We then examined the pattern of GM130 and p120 in Wolbachia-infected CS embryos. The Wolbachia vesicles rarely contained p120 protein (,1%; 2/228), but 46% of Wolbachia vesicles located close to the p120 vesicles (106/228) ( Figure 1E and 1E'). In contrast, 17% of Wolbachia vesicles contained GM130 protein (40/238), and 76% were juxtaposed to GM130-containing vesicles (180/238) ( Figure 1F and 1F'). These data suggest that Wolbachia bacteria reside in a type of Golgi vesicles that are closely related to cis-Golgi. We also found that Wolbachia were not present in endosomes, using an antibody against Hepatocyte growth factor-regulated tyrosine kinase substrate (Hrs) that is present in endosomes [25] ( Figure 1G). In conjunction with the previous report that Wolbachia are absent in mitochondria [26], we concluded that Wolbachia are present in a group of cis-Golgi related vesicles. doi:10.1371/journal.pone.0022703.g001 proteins may also play a role in apical localization of Golgi vesicles in cellularizing embryos. To test this notion, dsh 1 and fz 1 embryos infected with Wolbachia were examined. As shown in Figures 3D  and 3E, a significant fraction of Wolbachia vesicles in dsh 1 and fz 1 embryos was also found in embryo interior during cellularization, which was further confirmed by quantitative analysis (Figures 3A'-3E'). Mutations in non-PCP genes such as syntaxin did not result in mislocalization of Wolbachia vesicles ( Figure S1). It appeared that the PCP proteins may play previously unidentified roles in localization of Golgi vesicles important for membrane biogenesis.

Scribble and Vang are enriched in Golgi vesicles
Control of Wolbachia vesicle localization by PCP proteins suggests that some of these PCP proteins might actually function in the Golgi vesicles. We previously reported that Vang and GM130 frequently colocalize to the same vesicles in both fly embryos and human TE85 cells [14]. We extended this study to Fz and Dsh proteins, but these proteins were either expressed at a very low level or devoid of any distinct patterns, thereby making it difficult to draw any conclusion. We then turned our attention to Scrib, because Vang and Scrib not only genetically interact but also show direct physical interaction [37][38][39]. Fly Scrib is essential for establishment of apico-basal polarity and cooperates with Vang in PCP establishment [39,40]. Scrib is involved in many cellular functions related to PCP genes such as hair cell orientation and convergent extention in mammals [38,40,41]. Interestingly, Wolbachia vesicles in embryonic neuroblast cells are shown to concentrate near the apical membrane where Scrib is enriched [42]. We found that, in addition to the apical membrane of the cellularizing embryo, Scrib was present in cytoplasmic vesicles that either contained Wolbachia or tightly surrounded by Wolbachia vesicles (Figures 4A-A0).
Presence of Scrib in vesicular structures prompted us to examine whether Scrib and Vang colocalize in the same vesicles in the sub-apical region of Wolbachia-free embryos. Any structure recognized by anti-Vang antibody should be considered to contain endogenous Vang protein in Wolbachia-free embryos, based on the specificity of anti-Vang antibody ( Figure S2) [14]. As shown in Figure 4B, Scrib and Vang were sometimes present in the juxtaposed vesicles in the sub-apical region during mid-cellularization. More significant colocalization was observed in the vesicles in the embryo interior: all large-sized Vang-containing vesicles also contained Scrib ( Figure 4C).
We then examined the relationship between the Vangcontaining vesicles and GM130 in more detail. In newly laid embryos, Vang-containing vesicles were small and rarely contained GM130 ( Figure 5A). In contrast, the number and the size of Vang-containing vesicles in the embryo interior noticeably increased during the following mid-cellularization stage when the extensive membrane growth occurs. Furthermore, almost all medium to large-sized Vang-containing vesicles in the embryo interior also contained GM130 ( Figure 5B). Another finding was that the number of these large Vang vesicles decreased significantly at the end of cellularization when there is no further membrane growth ( Figure 5C). This suggests that these Vangcontaining vesicles may be an intermediate form of Golgi vesicles that is prerequisite for the final Golgi vesicles involved in membrane biogenesis during cellularization.
We have shown that Wolbachia are either present in the GM130containing vesicles or in the vesicles juxtaposed to GM130containing vesicles in the sub-apical region of embryos during midcellularization ( Figures 1F and 1F'). Since anti-Vang antisera have cross-reactivity to Wolbachia, it is not possible to check whether Vang is also present in Wolbachia vesicles. If Vang is indeed present embryo cortex at the end of cellularization. Wolbachia vesicles are localized not only in the sub-apical region (bracket) but also in the basal region (arrows, bracket with asterisk). (I) Wolbachia vesicles are visualized by serial tangential sections of epithelial cells in a wing disc. Density of Wolbachia vesicles is highest around 2-6 mm from the apical surface (bracket), and a minor fraction of Wolbachia is present near the basal position (bracket with an asterisk). Dlg was visualized as a membrane marker (green). Wolbachia, stained with both DAPI (blue) and anti-Vang antibody (red), appear as pink. Scale bar: A-C, 10 mm; D,E,H,I, 14 mm; G, 3.7 mm. doi:10.1371/journal.pone.0022703.g002 in Golgi vesicles that are harbored by Wolbachia, we reasoned that Vang should be present in the same vesicles with GM130 at the sub-apical region in the cellularizing embryo. As shown in Figures 5D and 5D', we found that all large GM130-containing vesicles either contained Vang or juxtaposed to Vang-containing vesicles in the Wolbachia-free embryo. This strongly suggests that the Golgi vesicles containing both GM130 and Vang may harbor Wolbachia or may be present in the vesicles juxtaposed Wolbachia vesicles. Thus, there is a possibility that Vang, Scrib and GM130 are present in the same Golgi vesicles, although further study is necessary to provide direct evidence. We then tested whether these Vang and GM130-containing vesicles are affected in mutant embryos that are defective in membrane biogenesis. Unlike in wild type embryos, such medium to large sized vesicles containing Vang were not detected in the embryo interior of Vang stbm-153 , Vang stbm-153 /Vang stbm-7-6 , and dlg HF321 embryos except only in small vesicles (Figures 6 and Figure S3). This suggests that both Vang and Dlg may be involved in generation or maturation of these large Golgi vesicles.

Discussion
Drosophila Golgi system is similar to mammalian Golgi system in its structure and function, and clearly displays several cisternae per stack that are polarized with cis and trans faces [43]. Although these fly Golgi vesicles are functionally diverse and can be distinguished by differences in glycosylation, they are detected as scattered dotty structures by most fly Golgi markers with confocal microscopy because their cisternae are not interconnected, and their size is less than half than that of mammalian counterpart [24,43]. Here we report that Wolbachia bacteria specifically reside in a special group of Golgi-related vesicles that may be functionally linked to membrane biogenesis. This makes Wolbachia an attractable marker for detecting functional fly Golgi vesicles. Wolbachia being in Golgi is also consistent with the previous report that Wolbachia are present in cytoplasmic vacuoles that are associated with astral microtubules and whose outmost membrane is host origin ( [7], Figure 1A).
We found a special relationship between Wolbachia and astral microtubules: all Wolbachia vesicles localized near the minus-ends of microtubules but not the plus-ends of microtubules ( Figure 1A). As the Golgi apparatus of mammalian cells is also shown to localize to the minus ends of microtubules, association between Golgi and the minus ends of microtubules may be a universal phenomenon [44,45]. Same conclusion could be drawn from analysis of literatures on the localization patterns of Wolbachia and microtubules in developing fly oocytes. During mid-oogenesis when microtubules play an essential role for axis formation, microtubule density decreases at the posterior region but increases at the anterior region of the oocyte [46]. At this time, minus-ends of microtubules are precisely concentrated at the anterior pole of oocyte, where Wolbachia vesicles are concentrated [26]. After this stage, both Wolbachia and minus-ends of microtubules become dispersed throughout entire oocyte [26]. These indicate that Wolbachia and minus-ends of microtubules colocalize during this critical stage, and polarization of Golgi vesicles may be important for delivering the axis determinants from the nurse cells to the right regions in the developing oocyte. Another link between Wolbachia and the minus-ends of microtubules is the mislocalization of Wolbachia in dynein mutants; anterior enrichment of Wolbachia in developing oocytes is disrupted when the dynein gene coding for a microtubule minus-end directed motor is mutated. In contrast, mutations in the kinesin gene that codes for a microtubule plus-end directed motor, do not cause any changes in Wolbachia localization in Drosophila oocyte [26,47,48].
An important clue for identifying the function of these Golgi vesicles came from the comparison of Wolbachia localization in wild type and various polarity mutant embryos during cellularization. Wolbachia vesicles were greatly enriched near the new membrane addition site in wild-type embryos, implying their potential involvement in membrane biogenesis ( Figure 2F). Fusion of Golgi vesicles onto the pre-determined membrane addition site leads to the addition of both membrane lipids and associated proteins to the right domain of the plasma membrane, which ensures not only the membrane growth but also the establishment of cell and tissue polarity. Mislocalization of Wolbachia vesicles in embryos defective in polarity genes such as dlg, Vang, fz or dsh, thus indicates that these polarity genes may somehow be involved in localizing these Golgi vesicles (Figure 3).
Involvement of polarity proteins in localizing Golgi vesicles has been recently reported. Dsh is shown to control association of membrane-bound vesicles and Sec8, a vesicle-trafficking protein, in order for apical docking of basal bodies in ciliated epithelial cells [49]. Vangl2, a mammalian homolog of Vang, is also selectively sorted into COPII vesicles by Sec24b, and Vangl2 looptail point mutant proteins fail to sort into COPII vesicles and are trapped in the ER [50,51]. Sec24b is a cargo-sorting member of the core complex that is important for formation of ER-to-Golgi transport vesicle COPII, and also genetically interacts with Scrib [51]. Furthermore, the knock-out mice mutated in Vangl2, scrib, or sec24b gene all show almost identical neural tube defects in addition to polarity defects [50][51][52][53]. Therefore, similar to Dsh, Vangl2 and Scrib may play a direct role in formation or localization of COPII vesicles instead of being just a cargo protein, and Vang and Scribble together with Sec24b may be involved in this process.
To properly localize the Golgi vesicles involved in membrane biogenesis, all the players that are involved in multiple Golgi maturation steps should sequentially act. Thus, mislocalization of Wolbachia vesicles may indicate that the Golgi vesicles are not fully matured, and consequently not capable of fusing to the plasma membrane. If some of the polarity proteins are the players in these Golgi maturation processes, the Wolbachia vesicles, indicative of the matured Golgi vesicles, would be mislocalized in mutants of the polarity genes. We found that cells in both dlg and Vang mutant embryos frequently show lack of membrane, supporting this idea [14]. Our data that the large vesicles containing both Vang and GM130 were not detected in a Vang mutant, also support the idea that Vang may be essential for the maturation of Golgi vesicles ( Figure 6).
One of the well-studied examples of PCP is the hair polarity in the fly wing [54]. Hair formation is restricted to the distal part of each wing cell by the core PCP proteins, Fz, Dsh, Vang, and Prickle (Pk) [32,[55][56][57][58]. These core PCP proteins are also involved in the PCP of photoreceptor cells and embryonic denticles [31,35,[59][60][61]. In case of the wing hairs, apical localizations of Vang and Pk in the proximal membrane and Dsh and Fz in the distal membrane in each wing cell during pupal stage are crucial for the positioning of a single distal hair. It is still largely unknown how the selective localization of these PCP proteins is achieved, but at least Fz protein appears to be delivered along the apical microtubules to the distal membrane of the hair cells [62]. The authors found that intracellular vesicles containing Fz-GFP marker preferentially move along the distally oriented apical microtubules and join the distal membrane [62]. One can imagine that the vesicles containing Vang or Pk may be preferentially delivered to the proximal membrane along the apical microtubules. Based on our data and others, we propose that these PCP proteins may play major roles in apical positioning of Golgi vesicles in either proximal or distal region of the wing cell, whose precise position is essential for cell and tissue polarity. When any one of these PCP proteins is not fully functional, proteins essential for PCP function may not be delivered to the proper location at the membrane, and consequently, both apical-basal and proximal-distal polarity would be disrupted. Taken together, these PCP proteins may not be just passively transported to the destined location at the membrane, but rather actively regulate the apical localization and the delivery of the distinct groups of Golgi vesicles.

Fly strains
The original CS strain containing Wolbachia pipientis, the same CS strain cured by tetracycline (250 mg/ml food) treatment [13,63] as well as Wolbachia-infected polarity mutants were used for this study. To infect the flies with the same type of Wolbachia, all the flies were first treated with tetracycline for three generations to cure any resident Wolbachia. The male flies from treated population were then crossed with the females of the Wolbachiainfected balancers. The siblings that had become infected with Wolbachia were mated to generate a Wolbachia-infected line.

Immunocytochemistry
Embryos were collected at either room temperature or 25uC, and fixed in 4% formaldehyde (methanol-free) by heptane method. We found that Wolbachia staining was quite strong with the anti-Vang and anti-Sdt antisera, and somewhat less with the anti-Dlg antisera [17]. Therefore, strong signals from antibody staining that also contain DNA were considered as Wolbachia in Wolbachia-infected embryos. In Wolbachia-free embryos, these antisera should recognize only their own endogenous proteins. For DNA staining, embryos after secondary antibody incubation were incubated with ribonuclease and then stained with propidium iodide. Alternatively, To-Pro-3 (Molecular Probes) or DAPI were used to stain nuclei.

Quantitative analysis of Wolbachia vesicles
Wolbachia vesicles were visualized with propidium iodide as well as two antisera, anti-Stardust (rabbit) and anti-Vang (mouse) antisera. The vesicular structures detected with all three markers were counted as Wolbachia vesicles. In order to obtain quantitative data, series of confocal sections taken along the apical-basal axis were processed with NIH Image J program. Since the density of Wolbachia varied from embryos to embryos, we presented data obtained from a representative embryo in Figure 3, instead of averaging the number of Wolbachia from different embryos along the apical-basal axis. Cross-sections of at least 20 embryos were examined for each mutation, 3 representative embryos were chosen for serial tangential sections, and one of them was presented in Figure 3. Figure S1 Wolbachia are not mislocalized in sys ts3 embryos. Wolbachia in a sys ts3 embryo are apically localized (arrows). Wolbachia are detected with anti-Vang antisera and propidium iodide.