Compartmentalized oskar degradation in the germ plasm safeguards germline development

Essential revisions:

1) Further characterization of founder granules and oskar mRNA localization to these structures is warranted. Is recruitment of the degradation machinery osk-dependent? Does the oΔn mRNA localized to the germ granules lack localization with these factors? Are there specific elements within oskar mRNA that drive its temporally specific degradation or is recruitment to founder granules sufficient to cause other mRNAs to be degraded. Do the authors have insights into what regions of oskar mRNA are necessary for its toxic effects? Osk transcript degradation in mutants or knockdowns of Dcp1, Me31B and pacman should also be analyzed, if possible.

a) Founder granules do not exist in the absence of oskar mRNA so we cannot directly test whether it is oskar RNA that is recruiting the machinery to founder granules. However, we have now shown experimentally that oΔn RNA in germ granules does not colocalize with DCP1, indicating that oskar RNA per se is not sufficient to recruit degradation machinery. We have added these data (Figure 4H,I.)

b) Currently, oskar is the only known founder granule resident RNA. We previously performed FISH for numerous transcripts that are enriched in the germ plasm but thus far have not identified any other RNA localized to founder granules. oskar localization is unusual in requiring splicing of the first intron to create a stem-loop localization element that acts in an unknown manner with the exon junction complex that is deposited during this splicing event. This localization signal mediates transport of oskar to the posterior during stages 7-10 of oogenesis. A second wave of oskar localization occurs at during stages 11-13, and it is not known what sequences in oskar mediate this localization. The bottom line is that we do not know how to drive RNAs into founder granules so we are not able to test whether recruitment to founder granules is sufficient to cause other mRNAs to be degraded.

c) Related to the point above, oskar is a very complicated RNA, with numerous elements that regulate various aspects of its transport, anchoring, and translational regulation scattered throughout the transcript. Dissection of the regions that control the degradation would take several years of work and is complicated by the need to avoid compromising these other functions – it is thus beyond the scope of this paper.

d) We do not know what regions of oskar are necessary for the toxic effects. C.E.E. (first author)'s thesis committee discouraged her from pursuing this question because they felt that she was unlikely to learn anything relevant to "normal" oskar function or germ cell development. We agreed that it was better to stay focused on upstream events (e.g., regulation of oskar) given limits of time and resources.

e) Unfortunately, is not possible to analyze oskar degradation in Dcp1, Me31B and pacman mutants because oogenesis is arrested. Dcp1 mutants arrest at stage 6, Me31B mutant egg chambers degenerate at or before stage 10, and pacman mutant egg chambers degenerate after stage 9. We tried a partial knockdown of pacman by RNAi and although we observed a small effect, the data were highly variable and not statistically significant (likely due to the partial nature of the knockdown).

2) The role of Aubergine (and potentially piRNAs) in the degradation of osk, presented in Figure 6 is interesting, but not convincing and also lacks experimental support. First, the work cited implicating AUB in the degradation of osk in the bulk cytoplasm is itself not entirely convincing [Barckmann et al., 2015]. Second, and more concerning, the observations mentioned in the Discussion section that AUB localizes to germ granules and not founder granules undermines the data presented in this manuscript that a) DCP1 (and later ME31B and PCM) colocalize to osk founder granules associated with osk degradation; and (b) oΔn mRNA localized to germ granules (which contain AUB) does not undergo degradation. Therefore, the conclusion that AUB is involved in regulating osk degradation is not concrete, and the authors cannot conclude that AUB plays a role in the recruitment of DCP1 to osk in founder granules based on the data presented here. The authors should experimentally address these concerns or consider deleting the Aub section altogether.

We agree that this is a surpising result. We have now performed a rescue experiment by expressing a functional gfp-aub transgene in mnk^– aub^–females. We show that degradation of oskar is indeed restored by expression of gfp-aub. Moreover, we show that DCP1 is recruited to oskar in these embryos. We have added a figure (Figure 8) with these new data.

I had hoped to be able to test whether the unlocalized population of Aubergine that is not in germ granules is the relevant source by expressing GFP-Aub mutants that either cannot be methylated or cannot bind piRNAs and consequently cannot associate with germ granules (obtained from the Aravin lab). Unfortunately, these proteins appear to be toxic and mnk^– aub^–females expressing them produce few eggs, all of which disintegrate rapidly after laying.

We have also toned down our conclusions in the Discussion. In particular, we say no longer state that Aubergine recruits DCP1 to founder granules because it implies a direct role for Aubergine. Rather we have changed the wording to state that recruitment of DCP1 to founder granules requires Aubergine and that Aubergine is required for oskar degradation. We suggest that the cytoplasmic pool of Aubergine could be the source and/or that Aubergine may act indirectly. Finally, we revised our citation of the previous work to say "Aubergine has been proposed to mediate degradation of mRNAs during the MZT…." which we also hope addresses the reviewer's concerns.