The zinc-finger transcription factor Hindsight regulates ovulation competency of Drosophila follicles

Essential revisions:

Although most of the experiments are conducted in a rigorous manner, supporting their conclusions, weaknesses revolve around the fact that the validation of the RNAi-based method is not sufficient to fully support their conclusions. First, the data in Figure 2—figure supplement 1 show that knockdown is quite incomplete. Thus, additional validation that the phenotypes are indeed caused by hnt knockdown is required.

In this study, we have used two independent hnt^RNAi lines, which are targeting non-overlapping regions of hnt mRNA (hnt^RNAi1 targeting nucleotide 4864-5206 and hnt^RNAi2 targeting nucleotide 1787-2077), to demonstrate that both hnt^RNAi lines cause similar ovulation phenotypes. Although both RNAi lines are predicted to have potential off-targets (hnt^RNAi1: CG10597 and CG5119; hnt^RNAi2: CG10228, CG12673, and CG7991), the only overlapping target between these two hnt^RNAi lines is the hnt gene. These data strongly suggest that hnt is responsible for the ovulation defect. We have also tried to use UAS-hnt^EP55 to rescue the defect of hnt^RNAi females. Despite using the Gal4/UAS system to overexpress hnt mRNA, unfortunately Hnt protein was still depleted by RNAi and not restored to normal level, and thus no rescue was observed (we added this result to new Figure 8—figure supplement 1). In contrast, mammalian homolog RREB-1, which will not be targeted by hnt^RNAi, is able to fully rescue the follicle rupture defect caused by hnt knockdown (Figure 8). This data also strongly support our conclusion that rupture defect in RNAi-based method is indeed caused by hnt knockdown, not peculiar RNAi off-target effect. Furthermore, we showed that mature follicles from hnt transheterozygous mutant females with a temperature-sensitive hnt allele also showed defect in OA-induced follicle rupture (Figure 4—figure supplement 1). All these genetic evidence support our conclusion that hnt is the gene required for follicle rupture/ovulation.

Reviewers also pointed out that hnt^RNAi knockdown is quite incomplete. It is true that hnt knockdown is incomplete, which is largely due to the late expression of Gal4 driver (both are in stage-14 egg chambers) and short developmental time (stage 14 only lasts for a little more than two hours and it only takes six hours to develop from stage 10 to stage 14). However, hnt is strongly depleted whenever Gal4 driver has high expression. For example, FC1 Gal4 is efficient to deplete hnt in stage 14B and stage 14C (Figure 3D and Figure 3—figure supplement 1H), while FC2 Gal4 is efficient to deplete hnt in stage 14C (Figure 3H). The incomplete depletion of hnt is also quite consistent with incomplete block of ovulation in hnt^RNAi females. Thus, we strongly believe the ovulation defect is indeed caused by hnt depletion.

The UAS-RREB-1 rescue experiment should be controlled for possible GAL4 / UAS "competition". This key experiment shows that there is a rescue of the hnt RNAi mediated knockdown phenotype by co-expression of UAS-RREB-1::GFP. I believe that it is possible that the rescue effect could be due to reduced efficiency of the RNAi knockdown, possibly through a non-specific effect such as less effective induction of UAS-hnt-RNAi expression due to the presence of the additional UAS responder element of UAS-RREB-1::GFP. (A possible titration effect of having additional UAS sequences available for the GAL4.) To control for this possibility of "GAL4 competition", a control experiment should be performed with a UAS-GFP in place of UAS-RREB-1::GFP. The level of expression of the UAS-GFP should be comparable to the UAS-RREB-1::GFP.

We do not believe that UAS-RREB1::GFP rescue effect is due to a titration effect of Gal4 by having an additional UAS sequence in this rescue experiment, because UAS-Oamb, which also contains an additional UAS sequence, could not rescue the follicle rupture defect (Figure 7C), nor did UAS-hnt^EP55 (Figure 8—figure supplement 1A). In addition, we performed the experiment as reviewers suggested to use a UAS-GFP to do the rescue experiment, and it didn’t show any rescue effect as well (Figure 8—figure supplement 3).

Does hnt-overexpression resemble RREB-1 expression? It would be informative to show if there is a phenotype associated with hnt over-expression (using UAS-hnt, UAS-GFP-Hnt, EP55, or P{GSV1}GS1018) in stage 14 follicle cells. Or if not over-expressed per se, is there a consequence to a prematurely high level of hnt expression, or maintenance of a high level of expression, in stage 14C – and could this address whether the down-regulation of hnt expression in 14C stage is functionally significant? I.e. Is hnt over-expression sufficient for the premature rupture phenotype and/or upregulation of Mmp2 and Oamb? Would such a phenotype resemble UAS-RREB1 expression, for example? These are fairly achievable experiments and their omission from this study seems puzzling to me.

We agree with reviewers that these are excellent experiments to carry out. We used hnt^EP55 line to perform the rescue and overexpression experiments as RREB1::GFP line. Unfortunately, hnt^RNAi lines are strong enough to block any overexpressed hnt mRNA through Gal4/UAS system and leave Hnt protein still depleted in rescue experiments, and no rescue effect was observed. The strong depletion of Hnt protein by hnt^RNAi when hnt^EP55 was included was further validated using a flip-out actin-Gal4 system. These data are now included in Figure 8—figure supplement 1. In addition, we also observed slight increase of follicle rupture when hnt is overexpressed using FC1 Gal4, very similar with overexpression of RREB1::GFP with FC1 Gal4. This is consistent with our hypothesis that Hnt functions as competency factor to regulate follicle maturation. The phenotype is weak, which is likely due to the fact that FC1 Gal4 only starts to be expressed in stage 14, when endogenous Hnt already starts expression. It would be ideal to have a Gal4 driver expressed in stage 13 and use such Gal4 driver to overexpress hnt, but unfortunately, we are not aware of the existence of this type of Gal4 line. Furthermore, we tried to maintain high hnt expression in stage-14C using FC2 Gal4, it caused slight reduction of follicle rupture but no egg-laying defect. Thus, we didn’t make any conclusion to avoid over-interpreting the data. We also performed similar experiments with UAS-Hnt::GFP lines from Dr. Howard Lipshitz’s lab and they showed the same result as hnt^EP55 line.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Reviewer #2:

[…] Two questions come to mind:first,is hnt^peb/+ control (~ 50%) showing a dosage effect compared to + /+ control (~80%)?;

We have repeated this experiment multiple times with two different researchers. Both came out the same result that hnt^peb/+; FC2>RFP/+ follicles showed ~50% rupture rate instead of ~80% in wild-type control. Therefore, we think it is possible that hnt^peb/+ causes a dosage effect. This dose effect may not just caused by hnt in the mature follicle cells, but also from other tissues expressing hnt (see response to question 2). Alternatively, hnt^peb may behave as a dominant negative allele since the nature of the temperature sensitivity is unknown.

and second,can the loss-of-function phenotype of hntpeb /hntXE81 and hntpeb /hntEH704a be rescued by FC1 or FC2 GAL4 expression of UAS-RREB-1? This latter experiment would completely and definitively address any concerns regarding effectiveness of RNAi depletion phenotypes and their rescue by UAS-RREB1. Since the rescue of the knockdown is pivotal to a conserved function, I would really like to see inclusion of this experiment. It should only take a couple of fly generations to get the GAL4 + UAS lines into the hntpeb /hntXE81 and hntpeb /hntEH704a backgrounds, and the assay is well established in the lab. I think it will strengthen the rescue results.

We do not believe that FC1 or FC2 driving UAS-RREB-1 expression will rescue the hnt loss-of-function phenotype for the following reasons: Hnt is not only expressed in stage-14 follicle cells, but also expressed in follicle cells of stage 7-stage10B, which is essential for follicle cell differentiation according to my previous study (Sun and Deng, 2007). Although no morphological defect was observed in hnt^peb/XE81and hnt^peb/EH704amature follicles, we cannot exclude that subtle defect in early development may contribute to the follicle rupture defect. That’s why we used RNAi to bypass earlier Hnt function in our initial experiments. In addition, Hnt is also expressed in the midgut enterocytes and secretory cells in the female reproductive tract. The latter has been demonstrated to regulate ovulation as well (Sun and Spradling, 2013). Hnt may also function in the secretory cells in the female reproductive tract to regulate follicle rupture in addition to its role in the mature follicle cells. This is supported by our recent finding that another transcription factor is functioning in both secretory cells and mature follicle cells to regulate follicle rupture (Knapp and Sun, manuscript in preparation; this story is also going to be presented in the Annual Drosophila Research Conference in April 2018). We are interested in investigating the role of Hnt in serectory cells, but this is out of the scope of this study. Therefore, we do not believe this experiment will work. Even this will work, which will take another two months with several generations of crosses, it would not add any new findings to our current conclusion.

Also, is there any evidence that RREB-1 is expressed in the mammalian ovary? Is it possible to find out somehow? (Such as some high throughput data somewhere?) If so, then it would certainly strengthen the conserved function in ovulation. But if it is not? Then…?

I am sure there’re high throughput data available to support that RREB-1 is expressed in mammalian ovary, but many of those data are not conclusive and need further validation. We are interested in this question and our preliminary data had showed that RREB-1 is expressed in follicles in LH-dependent manner. We are in the process of generating RREB-1 conditional knockout mouse to test its role in mammals. This project is out of the scope of this current study.