The Drosophila Sp8 transcription factor Buttonhead prevents premature differentiation of intermediate neural progenitors

Thank you for sending your work entitled “The Drosophila Sp8 Transcription Factor Buttonhead Prevents Premature Differentiation of Intermediate Neural Progenitors” for consideration at eLife. Your article has been favorably evaluated by Janet Rossant (Senior editor), Marianne Bronner (Reviewing editor), and 2 reviewers.

The Reviewing editor and the reviewers discussed their comments before we reached this decision, and the Reviewing editor has assembled the following comments to help you prepare a revised submission.

Drosophila type-II neuroblast (NB) lineages produce neural cells indirectly, by generating transiently amplifying Intermediate Neural Progenitors (INPs). INPs play a crucial role in mammalian brain development and understanding how they form and proliferate is therefore a central question in neurobiology. In a previous study, the authors have shown that the Ets transcription factor Pointed P1 (PntP1) is required for the specification of Type-II NB lineages in Drosophila larval brains. PntP1 suppresses Asense (As) expression in NBs, which promotes the formation of INPs. Here, Xie et al. show that the product of the button head gene (btd) works cooperatively with PntP1 to generate functional INPs.

They first show that btd is required for the maturation of INPs: In its absence, INPs are still generated but they undergo premature cell cycle exit and differentiate into Ganglion Mother Cells (GMCs). Xie et al. elegantly demonstrate that this phenotype is due to the ectopic expression of Prospero (Pros) in the nuclei of newly generated INPs. Thus, btd prevents premature differentiation of INPs by repressing early Pros expression. The authors then test whether the btd phenotype might be due to the loss of PntP1. They show that although PntP1 expression decreases in the NBs of btd mutant type-II lineages (leading to ectopic Ase expression in a subset of these NBs), this defect is not at the origin of the loss of mature INPs. It is instead a secondary effect of the ectopic expression of Pros. These results nicely confirm the key role played by btd in the formation of mature INPs. Finally, Xie et al. show that co-expressing btd with PntP1 in type-I NB lineages is sufficient to convert most of them into type-II NB lineages. The major finding reported is that the transcription factor encoding btd gene is required in INPs to prevent their premature differentiation, and that this occurs because btd suppresses the (nuclear) expression of the homeodomain protein Prospero in INP sublineages. These interesting data add to our knowledge of the molecular mechanisms involved in the control of the (limited) proliferative potential of INPs in the Drosophila neural stem cell model. This paper is interesting, well written, and the data are convincing.

Major concerns that need to be addressed:

1) Some of the data presented has been published previously and are not declared as such. The data in Figure 1A, A' in this paper correspond to that Figure 1B, B' in previously published paper by the same senior author (Zhu et al., 2011), with only artwork and magnification changed. The authors should check all figures carefully to ensure that this does not happen and not recycle the same data.

2) The authors should mention clearly that the use of btd-Gal4 to study Btd expression and drive transgenes might not faithfully reproduce the endogenous expression pattern of Btd. There are two reasons for this. First, enhancer-reporter constructs do not necessarily reflect endogenous protein expression. Moreover, if enhancer-reporter constructs do mimic endogenous expression in one tissue, this does not prove that the constructs mimic endogenous expression in other tissues. Clearly, immunolabelling or in situ hybridization would be preferable and should be added to the paper if an antibody is available. If this is not available, the authors should make sure that they explain that they have made all the controls to test whether the Gal4 line reproduces the expression pattern. In particular, the ability to rescue the loss of function phenotype with the human construct is a good argument that the Gal4 line is expressed in the right tissue. Rescue with the fly UAS construct should be added to confirm this.

3) Second, the btd-Gal4 construct used by the authors is an insertion in the btd gene that actually causes a mutation in the gene (Estella et al., 5929). This is not a serious issue since the authors always use it as heterozygous (the authors should confirm this in the text), but this fact should be explained and the authors should explain the controls that explain why this is not an issue.

4) In some quantifications, the authors have a sample size of n=3. They should quantify at least 6 or 7 samples for each experiment to further support their conclusions.

1) Some of the data presented has been published previously and are not declared as such. The data in Figure 1A, A' in this paper correspond to that Figure 1B, B' in previously published paper by the same senior author (Zhu et al., 2011), with only artwork and magnification changed. The authors should check all figures carefully to ensure that this does not happen and not recycle the same data.

We apologize that we reused an image that has been published before without realizing it. We did not intentionally recycle the data, but we really appreciate that the reviewers noticed this problem. We should have looked more carefully to make sure the same image was not used again. We have replaced the reused image in Figure 1A, A’ with new ones with similar quality in the revision. We will also make sure that similar things do not happen again in the future.

2) The authors should mention clearly that the use of btd-Gal4 to study Btd expression and drive transgenes might not faithfully reproduce the endogenous expression pattern of Btd. There are two reasons for this. First, enhancer-reporter constructs do not necessarily reflect endogenous protein expression. Moreover, if enhancer-reporter constructs do mimic endogenous expression in one tissue, this does not prove that the constructs mimic endogenous expression in other tissues. Clearly, immunolabelling or in situ hybridization would be preferable and should be added to the paper if an antibody is available. If this is not available, the authors should make sure that they explain that they have made all the controls to test whether the Gal4 line reproduces the expression pattern. In particular, the ability to rescue the loss of function phenotype with the human construct is a good argument that the Gal4 line is expressed in the right tissue. Rescue with the fly UAS construct should be added to confirm this.

We agree with the reviewers that a GAL4 enhancer trap line like btd-GAL4 does not always reflect endogenous expression patterns of the affected gene and it is essential to verify the btd-GAL4 expression by immunostaining or in situ hybridization. However, we tried to generate btd antibodies using two different approaches (using a synthesized peptide and GST-btd fusion proteins), but had no luck. We also tried very hard to detect btd mRNAs by in situ hybridization. We got very nice in situ hybridization signals in the larval optic lobe (which is consistent with the expression of btd-GAL4 in the optic lobe) and ventral imaginal discs, but we were not able to detect strong signals in the central brain or in the ventral nerve cord. Therefore, in order to test whether the btd-GAL4 line reproduces the endogenous expression pattern, we tried to rescue the loss of function phenotype of Btd with the expression of mouse Sp8 construct (sorry we did not have the human Sp8 construct) or fly btd construct driven by btd-GAL4 as reviewers suggested. Ideally this rescue should be done in btd mutant clones. However, we were not able to do such rescue experiments for the following two reasons:

(A) Although the GAL4 insertion in the btd-GAL4 line causes a lethal mutation of btd, we did not observe any obvious phenotypes in btd-GAL4 homozygous mutant type II NB clones, indicating that the GAL4 insertion does not affect the expression of btd in type II NB lineages (please see the new figure: Figure 6–figure supplement 1), which is possible given that the GAL4 is inserted at 753bp upstream of the transcription start site of btd and the insertion of GAL4 may not disrupt the enhancer elements that drive the expression of btd in type II NB lineages. The embryonic lethality of the btd-GAL4 line could be due to the loss of Btd in other tissue or cells. Therefore, we could not do the rescue by using btd-GAL4 to drive the expression of UAS-mSp8 or UAS-btd in btd-GAL4 homozygous mutant type II NB clones. In any event, we were able to rescue the lethality of btd-GAL4 line by driving the expression of UAS-btd and type II NB lineages developed normally in those rescued btd-GAL4 mutant larvae (data not shown, but we would be happy to provide images if needed).

(B) Since btd-GAL4 is inserted in the promoter of btd, we could not do the rescue by using btd-GAL4 to drive the expression of UAS-mSp8 or UAS-btd in type II NB clones homozygous mutant for other btd mutant alleles, such as btd^XG81.

Therefore, in order to rescue the loss of function phenotype of Btd with the expression of mouse Sp8 or fly Btd driven by btd-GAL4, we tried to rescue the Btd RNAi knockdown phenotypes in type II NB lineages by the expression of UAS-mSp8 or UAS-btd driven by btd-GAL4. In a new supplemental figure (Figure 6–figure supplement 2), we showed that Btd RNAi knockdown driven by btd-GAL4 completely eliminated mature INPs in nearly all type II NB lineages, which is much stronger than that of Btd RNAi knockdown driven by pntP1-GAL4 (this could also indicate that the expression of btd-GAL4 is likely more close to the endogenous expression pattern of Btd than pntP1-GAL4 so that the knockdown of Btd driven by btd-GAL4 is more efficient pntP1-GAL4, although we cannot rule out the possibility that the higher RNAi knockdown efficiency is due to the difference of the expression levels of btd-GAL4 and pntP1-GAL4). This Btd RNAi knockdown phenotype could be fully rescued by the expression of UAS-mSp8 driven by the same btd-GAL4. However, the expression of UAS-btd driven by btd-GAL4 could only partially rescue the Btd RNAi knockdown phenotype. The incomplete rescue by UAS-btd is most likely because the UAS-btd construct contains the target sequence of UAS-btd RNAi. In any event, the rescue of the Btd RNAi knockdown phenotypes by the expression of UAS-mSP8 and UAS-btd driven by btd-GAL4, together with strong btd mutant phenotypes in type II NB lineages (and the failure of PntP1 to induced INP-like cells in btd mutant type I NB lineages), strongly argue that btd-GAL4 reflects the endogenous expression pattern of btd.

3) Second, the btd-Gal4 construct used by the authors is an insertion in the btd gene that actually causes a mutation in the gene (Estella et al., 5929). This is not a serious issue since the authors always use it as heterozygous (the authors should confirm this in the text), but this fact should be explained and the authors should explain the controls that explain why this is not an issue.

Thanks for reviewers’ suggestions. We now mentioned in the text that the insertion of GAL4 transgene in the btd-GAL4 enhancer trap line causes a lethal mutation of btd and we only used btd-GAL4 heterozygous female larvae for phenotypic analyses. We also explained in the text that the generation of INPs are not affected in btd-GAL4 heterozygous mutant type II NB lineages as shown in Figure 6A-A’ or in btd-GAL4 homozygous mutant type II NB clones as mentioned above (Figure 6–figure supplement 1). Therefore, btd-GAL4 heterozygous mutant background would not affect the generation of INP-like cells and thus it should not cause any problems when the btd-GAL4 line was used for our research.

4) In some quantifications, the authors have a sample size of n=3. They should quantify at least 6 or 7 samples for each experiment to further support their conclusions.

We have increased the sample size as the reviewers suggested and did the statistical analyses again based on new sample sizes. We also indicated in the figure legends the samples sizes are the number of brain lobes or the number of type II or type I NB lineages/clones.