Stem cell regionalization during olfactory bulb neurogenesis depends on regulatory interactions between Vax1 and Pax6

Reviewer #1:

1) Figure 1 – The combination of ISH and IC is legitimate but the data are not compelling. The overlap in Figure 1G-H are good examples. This is also evident in the qualitative description of overlap. In any event, these data are not very clean. The paper would benefit from a quantitative approach. For example, a double-labeling smFISH expression analysis would be more successful.

We quantified the percentage of double stained cells for the different markers used in Figure 1 D-F. Moreover, in the RMS (Figure 1 G,H) cells migrate partially intermingled, making a graded distribution of Vax1 and Pax6 positive cells, as is obvious in the SVZ, less evident. We added higher magnification images to illustrate the differential expression of Vax1 in the ventral (Figure 1G’) and dorsal (Figure 1G’’) aspects of the RMS. We also added higher magnifications to illustrate the co-expression of DLX2 and Vax1 in the RMS (Figure 1H’ and H’’). The Results section was changed accordingly.

2) Figure 2 – The comparison of GCL tdt+ numbers (Figure 2 B,C ) should be compared to PGL numbers. The current comparison to 2D,E is not a fair comparison (apples to apples…). For claim a specificity effect, they should compare to other cell types in the PGL in the same animals.

New data showing the PGC quantification was added to the manuscript as Figure 2—figure supplement 1B. Please also note that, in addition to the CB+ PGC population now presented in Figure 2E, we already analyzed in the same animals the number of TH⁺ PGC (Figure 4I).

3) Figure 3D – The large scale gradient is convincing but not very novel. To drive this conclusion home, a similar analysis that is done in 3A should be conducted at a single cell level.

In C-F the claim for specificity is weak. They should show that expression of other TFs are not hampered in the GFP cells. It’s possible that the mere over expression of a Vax1 impacts other proteins in a negative way. I am not from the field but it seems to me that a better control is needed (e.g. at the very least a scrambled version of Vax1OE).

Here, the reviewer’s comments/questions are in part a bit unclear to us and we hope we got the main points.

To our knowledge the existence of large-scale gradients in the postnatal SVZ is still a relatively new concept. Designing and performing a single cell analysis experiment that provides reliable information about such gradients is not easy and, to our eyes, exceeds the scope of this study. Use of a scrambled protein as control for gain-of-function experiments in vivo is a rather unusual approach in the field. As far as we know, such a control has so far not been applied. Finally, specific Pax6/Vax1 regulatory interactions have been shown in other brain developmental contexts, like for example the eye, arguing for the specificity of the observed effects here.

4) Figure 4 – in 4B the CR distribution seems to be different – it has a double peak as compared to controls. More appropriate statistics may show this difference.

4I distributions are very similar. The text description of these as being marginally significant is misleading.

Concerning 4b: We agree that there is some heterogeneity in this specific experiment. However, all samples have been treated exactly the same way and analyzed by the appropriate statistical tests. We do not see which variables can be tested to address the reviewers point.

Concerning Figure 4I: We accept the reviewer’s critique and removed any potentially misleading conclusion concerning these data from the Results section.

5) The miR-7 data are robust but remain anecdotal and currently not very natural to add it to the storyline of the paper that remains intact without it.

We added additional data, reinforcing the link between Vax1 and miR-7, by showing the all three miR-7 promoters contain putative Vax1 binding sites. We opted to maintain the data in the manuscript, but tempered our conclusions concerning this aspect. We also removed the mention of miR-7 from the manuscript title.

Reviewer #2:

1) Data regarding the contribution of miR-7 are limited. I think more data are either necessary to show that Vax1 works through mir-7 to regulate neuronal fate or the authors should change their writing and conclusion.

For example, in the Abstract: "We provide evidence that this repression occurs via activation of microRNA miR-7, targeting Pax6 mRNA." I think it is overselling the data presented in this manuscript. Much more data would be required for this conclusion. For example, the mir-7 data could be left out of the Abstract but presented at the end of the Introduction as well as discussed.

We accept the reviewer’s critique and react threefold. First, we provide new data showing that all three miR-7 promoters contain putative Vax1 binding sites. Second, we tempered our conclusions concerning the Vax1-miR-7 link in the Abstract and the main manuscript. Third, we removed the term miR-7 from the title.

2) Figure 1: please show analysis of the in situ/immune data.

As requested, we included quantifications of the in situ/immune staining’s. The data are now presented in the Results section.

3) Better images for Figure 3C would be helpful to actually see the decrease in Pax6 intensity.

We added new images illustrating the downregulation of Pax6 protein after Vax1 electroporation. The new data are now presented as Figure 3D.

4) The sentence: "However, we stably observed a tendency for an increase over independent electroporation experiments (3 experiments for each condition with a total of 12 animals/ condition, Figure 4I)." should be removed. There is no effect.

We agree and removed the phrase from the Results section of manuscript.

Reviewer #3:

1) In the loss-of-function experiment in Figure 2, it would be helpful if the authors could show how efficiently Vax1 expression is suppressed in the targeted tdTomato positive cells　by using in situ hybridization.

To address the efficiency of CRE mediated recombination we expressed the recombinase by in vivo electroporation in control and Vax1/Ai14 double mutants and isolated transfected tdTomato-positive cells by FACS. The quantitative RT-PCR results, now shown in Figure 2—figure supplement 1, clearly show a massive reduction in Vax1 mRNA levels after CRE expression, demonstrating efficient recombination.

This approach was necessary as in light of the relatively small number of transfected cells after in vivo electroporation – and facing the cellular complexity of the postnatal SVZ- in situ hybridization is not suited to give a quantitative notion of gene-knockout.

2) Also in the gain-of-function experiment in Figure 3, the levels of Vax1 derived from the electroporated vector DNA should be shown for the GFP-positive cells, by in situ hybridization.

As for the knockout, we combined electroporation, here of an overexpression plasmid, with FACS and qRT-PCR. We found the expected large increase in Vax1 mRNA levels in the GFP positive fraction. This control is shown in Figure 5B.

3) In this manuscript, the authors analyzed regulatory interactions among Vax1, Pax6, and miR-7 only in neonates. Since interneurons are constantly regenerated during adulthood, it would be helpful, if the authors could demonstrate the data showing that the same regulatory cross-talk can be seen in the adult neurogenesis of interneurons.

In previous work we demonstrated that the dorsoventral Pax6 and the ventrodorsally oriented miR-7 gradients are maintained in adult stages (de Chevigny et al., 2012b, Supplemental Figures 1 and 4). Here, we show now, as the new

Figure 3—figure supplement 1, that the postnatal ventrodorsally oriented Vax1 gradient is also maintained in the adult, allowing to speculate that the mechanism remains active. This new data on adult Vax1 expression is now mentioned and discussed in the manuscript.

It would of course be tempting to test the validity of the Vax1/Pax6/miR-7 interactions in adults. However, in vivo electroporation in adults is far less efficient than in the postnatal, leading only to low amounts of transfected cells in the OB. As our analyses depend on precise quantification of large numbers of cells in gain- and loss of-function experiments, meaningful data cannot be expected from such approaches.

Viral transduction, that is normally used in adults, cannot be targeted to specific subregions of the SVZ. It appears conceivable to design complex transgenic approaches to target specific SVZ compartments with CRE, but such experiments, if possible, would represent a long-term project on its own that exceeds what can be done here. We hope this is acceptable.

4) What is the mechanism of Vax1-mediated miR-7 enhancement? Do they directly interact with each other? Is there any binding site for Vax1 within the promoter region of the miR-7 gene?

Following the reviewer’s suggestion, we investigated the presence of Vax1 binding sites in the miR-7 regulatory regions. Indeed, we found such sites in all three miR-7 promoters, in agreement with a potential direct regulation. This new data are now presented in Figure 5D.

5) Cross-regulatory interactions for Vax1, Pax6, and miR-7 are nicely shown in this study. It would be helpful if the authors could discuss this regulation strategy in the context of neurogenesis during brain development in order for the paper to attract a wider research audience other than those investigating transcription factors.

To place our results in a wider context, we discuss now the OLIG2/Irx3/miR-17-3p regulatory network, that is used to define progenitor domains in the developing spinal cord.