Main

The endosymbiotic bacteria (fluorescent staining in Fig. 1) are maternally inherited and sensitive to the antibiotic tetracycline. Analysis of their DNA after amplification by polymerase chain reaction (PCR) revealed an ftsZ sequence3 that was identical to that of Wolbachia strain A. This sequence was the same in both the Canton S wild-type Drosophila stock and the laboratory stock4 that we used as a source of endosymbionts for all the other stocks.

Figure 1: Colonization of Drosophila germ cells by Wolbachia.
figure 1

a, Wolbachia clustered on asters around proliferating nuclei and concentrated in the germ plasm on the posterior dorsal side of a wild-type Drosophila embryo (single optical section). b, Germ-plasm colonization in an unfertilized egg from a heterozygous oskar5/+ mother. c, Unfertilized egg lacking germ-cell determinants from an oskar5/oskar5 mother. Scale bar, 50 μm. Serial 2-μm optical sections of Drosophila eggs were fixed in methanol and stained for DNA with H33258, collected on a Bio-Rad MRC 1024ES confocal microscope with multi-photon excitation, using a 40×0.85NA (air) lens set at a zoom of 1.5 and a pixel size of 0.388 μm, and then converted to a single two-dimensional image by maximum brightness projection.

Full size image

These bacteria are found throughout the syncytial embryos of infected strains of Drosophila, concentrating in the region of the cortical actin cytoskeleton and accumulating with the mitochondrial motor protein KLP67A on the microtubules of mitotic asters4,5,6. These mechanisms might ensure transmission of the bacteria, as would colonization of the posterior pole, the site of germ-cell formation. This has been reported for Nasonia wasp embryos, which also show cortical localization7,8, but not for Drosophila.

To identify the host genes required for posterior localization of Wolbachia in the absence of asters, we counted the bacteria in four regions of unfertilized eggs (Table 1) from wild-type females and from females bearing mutations in gurken or oskar, two maternally acting genes that determine the site of germ-cell formation. The posterior pole region P1 had a mean bacterial density significantly different from those of areas A1, A2 and P2 in eggs from wild-type and heterozygous mothers. The bacteria were not localized in eggs from homozygous mutant mothers. The posterior localization of Wolbachia therefore requires not only the functional anteroposterior axis of polarized microtubules specified by Gurken signalling9, but also the germ plasm determined by Oskar2.

Table 1 Quantification of Wolbachia in regions of the Drosophila egg
Full size table

The eggs of our Canton S strain had a density of 3.04±1.04×106 bacteria per mm3. We used published data7 to calculate the density in the Nasonia vitripennis egg as 0.51±0.13×106 bacteria per mm3. In contrast to the tight posterior localization of Wolbachia in the Nasoniaembryo, most bacteria in Drosophila colonize somatic cells. The fate of these bacteria is unknown, but they may infect the germ line later in development.

Alternatively, the symbiosis may be less well established in Drosophila. We considered that transferring Wolbachia from Nasonia to Drosophila would show whether the germ-cell signals of the two species are recognized by a conserved bacterial mechanism. We found that early germline association seems to be symbiotic: Wolbachia infection was not a significant variable influencing the number of germ cells or the fidelity of their migration to the gonad in the Canton S strain (results not shown).

Posterior localization of Wolbachia was also abolished in eggs of Drosophila that are mutant for the vasa or tudor genes, which encode germplasm components localized by Oskar (data not quantified). Functional germ-cell determinants are localized in the mature oocyte (stage 14)10, but we could find no oocytes at stages 1–14 that had posterior localization of Wolbachia, in contrast to 80% of unfertilized eggs 0–3 hours after egg deposition (results not shown).

If germ plasm is merely a rich food source that stimulates local proliferation, why are Wolbachia not seen there earlier in oogenesis? Although we have not excluded the possibility of rapid local proliferation after egg deposition, we think it is more likely that the bacteria seen in ovarian nurse cells and oocytes relocate in response to a component of pole plasm that is translated only after egg activation11.