An exon junction complex‐independent function of Barentsz in neuromuscular synapse growth
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (Review Commons)
Abstract
Article activity feed
-
-
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We thank the reviewers for their positive comments on our manuscript. To address their criticisms, we propose to do the following experiments:
__Reviewer 1 (mi____nor comments)____: __
- *In Fig. 1, the authors show that Btz-WT, but not Btz-HD, localizes to the posterior pole of the oocyte. Do the authors see Btz-WT and/or Btz-HD localized to MNs/muscles/glia at the NMJ? * We have had difficulty detecting the expression of our Btz-GFP transgenes at the NMJ. In case this was due to competition with endogenous wild-type Btz, we will repeat the staining in a *btz *mutant background. If the protein is still undetectable, we can include data showing the …
Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We thank the reviewers for their positive comments on our manuscript. To address their criticisms, we propose to do the following experiments:
__Reviewer 1 (mi____nor comments)____: __
- *In Fig. 1, the authors show that Btz-WT, but not Btz-HD, localizes to the posterior pole of the oocyte. Do the authors see Btz-WT and/or Btz-HD localized to MNs/muscles/glia at the NMJ? * We have had difficulty detecting the expression of our Btz-GFP transgenes at the NMJ. In case this was due to competition with endogenous wild-type Btz, we will repeat the staining in a *btz *mutant background. If the protein is still undetectable, we can include data showing the localization of UAS-Btz-GFP when overexpressed in muscles or motor neurons.
*The mitochondrial phenotypes observed in Btz mutants are striking. But it seems possible that there are defects in overall mitochondrial levels in muscle in addition to defects in their localization. Overall, mitochondrial levels seemed reduced in Btz mutants. Is it possible to do a ATP5A immunoblot in Btz mutants to test whether overall mitochondrial levels are altered? *
We will do a Western blot to compare ATP5A levels in btz2/+ and btz2/Df(3R)BSC497 larval carcasses.
*ECM proteins are known to be critical for regulating TGFB signaling. That, taken with the multi-tissue genetic requirement for Btz, suggests that Btz might directly regulate either Ltl or Frac RNA, given that these ECM proteins are likely deposited by multiple cell types. *
We agree that this is a possibility and we will mention it in the Discussion.
Reviewer 2 (major comments):
- *In Figure 1, regarding the validation of rescue constructs: the EJC interaction-defective mutant is based solely on conservation, as all structural/interaction studies cited with Btz bound to EJC have been with human proteins. They use Vasa localization as a readout of EJC-dependent function, but this is indirect and only assesses one aspect of EJC function (localization). Since many of the main conclusions in the paper are predicated on this mutant being EJC-independent, they should validate this with the Drosophila orthologs using immunoprecipitation. They demonstrate the capability of expressing GFP-tagged versions of Casc3 WT and mutant in S2 cells, so this should not be a cumbersome control experiment to include. * We will express tagged Btz-WT and Btz-HD proteins in S2 cells and test whether they can be co-immunoprecipitated with Myc-tagged Drosophila eIF4AIII.
Regarding Figure 3, it could be postulated that the number of boutons would be influenced by the length of axons. Is axon outgrowth accounted for in these experiments? This would influence number of synaptic boutons. Panel F looks very different from panel A in terms of axon length (could this be due to axon outgrowth defect and/or impacted muscle size?) Can quantitation be done also by normalizing to axon length (bouton number/axon length)? Or perhaps this is accounted for in muscle size? If so, this should be explained.
The NMJ grows during development by adding both axonal branches and synaptic boutons, so its size can be measured by counting the number of boutons or branches or measuring branch length. These measures are usually well correlated. In this paper we used bouton number normalized to muscle surface area as our measure of NMJ size, but we did observe corresponding changes in the number and length of branches, as the reviewer points out. We will explain this more clearly in the text.
*In Figure 3 quantification: n's vary between genotypes significantly, and this should be explained (e.g. was there a recovery issue between genotypes or just fewer needed for WT-like?). *
The btz mutant larvae are more difficult to dissect due to muscle fragility, and some crosses in this genetic background may have yielded fewer usable filets than desired. We believe the numbers we obtained are sufficient to show which differences are significant.
*In Figure 4 panels B and F (mutants), there appears to be reduced axon outgrowth (see point above). This should be taken into account when expressing bouton number. *
As explained in our response to point 2, axon length and bouton number are correlated measures of synapse size and vary together in this figure as expected.
*The RNA-seq data (Figure 5) has a potential issue in that they used larvae with a balancer chromosome (Df), which yields a 50% reduction in any genes on that chromosome. They acknowledge this and removed these genes from the analysis, but the concern remains that this still might be a confounding variable (for example, if reduction in any of these genes might disrupt a signaling pathway). We do not think that the RNA-seq needs to be repeated, but we propose that the authors validate these targets using qPCR in their MN-specific btz knockdown system (this way, they can also include magoh and eif4aIII knockdowns for comparison). *
Because only one btz allele was available, we used transheterozygotes with a deficiency for the region to avoid homozygosing other mutations that might be present on the btz2 chromosome. As a consequence, we did observe reduced expression of genes located within the deficiency (which covers a small region, not an entire chromosome), and it is possible that this might contribute to the phenotype. However, we have seen a similar reduction in NMJ size in *btz2 *homozygotes. We do not think that motor neuron-specific btz knockdown is a useful genotype to validate the RNA-Seq results because *ltl *and frac levels do not change significantly in the CNS, only in muscle, and knockdown only in motor neurons would be unlikely to change daw levels measured in the whole CNS. Knocking down mago or *eIF4AIII *in muscle is lethal before the third larval instar stage, preventing us from comparing their effects on gene expression to those of btz. However, we will do qRT-PCR to measure daw, ltl and frac mRNA levels in btz2 homozygous mutant muscles.
__Reviewer 2 (minor comments): __
- *Some statements made in the introduction that are not entirely accurate: **
"A fourth core subunit, known as Barentsz (Btz), Cancer susceptibility candidate gene 3 (CASC3), or Metastatic lymph node 51 (MLN51), associates with the complex following the completion of splicing, and is required for the effects of the EJC on translation, NMD and mRNA localization (Chazal et al., 2013; Palacios et al., 2004; Shibuya et al., 2006; van Eeden et al., 2001)."
A recent study indicates that Casc3 is not required for EJC-dependent NMD targets in human cells, but rather enhances NMD on a subset of targets (Gerbracht et al. 2020 NAR). Perhaps "is required" should be changed to "plays a role in cytoplasmic EJC-mediated processes, such as...". It has also been shown that EJC core can assemble without Casc3 (e.g. Ballut et al 2005 NSMB, Gehring et al 2009 PLoS Biol). Previous work from the authors show that Casc3 (Btz) is not necessary for EJC function in pre-mRNA splicing (Roignant et al, 2010 Cell). Further, there exists a population of Casc3 lacking EJCs in human cells (Mabin et al 2018 Cell Reports). Collectively, all this evidence points to Casc3 not being a core EJC subunit. *
We will change the text so that we do not refer to Btz/Casc3 as a core subunit.
- "In the mouse brain, haploinsufficiency for Magoh, Rbm8a or Eif4a3 causes severe microcephaly, but complete loss of Casc3 has a much milder effect that can be attributed to developmental delay (Mao et al., 2017; Mao et al., 2016; Mao et al., 2015; Silver et al., 2010)."
From Mao et al 2017: complete loss and hypomorphic mutants were embryonic and perinatally lethal (contrary to what the authors are stating here), while compound mutants and heterozygotes exhibited neurodevelopmental delay. By "milder effects" the authors could also be referring to brain size being proportional to body size in the complete loss homozygotes; either way, this should be clarified. *
By “milder effects” we meant the effect on brain size. We will clarify this in the revised text.
*Fly-specific nomenclature could be made more accessible to a broader audience, as the full readership will likely not have expertise in Drosophila genetics. For example, w118, btz2 labels used in figures are not explained anywhere in the manuscript. While the authors do a good job of describing various mutants in a more accessible fashion in the results section, the genotype labels in figures can be better explained in the legends. *
We apologize for this and will clarify the genotype labels in the figure legends.
*Fig 2 L-N panels might warrant more explanation. Can the mitochondria be counted here? Is there also a difference in volume/morphology that could be quantitated? In Figure 2N, muscle fibers are more densely packed in mutant vs. control; can this be explained? *
We are hesitant to quantify mitochondria or comment on muscle fiber packing based on the EM images, because only one individual of each genotype was examined. We prefer to simply use these images to provide a higher resolution view of the change in mitochondrial distribution that we observed and quantified using light microscopy. However, we do plan to do a Western blot to determine whether there are changes in the number of mitochondria in *btz *mutants (see Reviewer 1 point 2).
*In Fig 2, to draw parallels between panels A-K and L-N, it might also be helpful to use the red/yellow arrow system on panel A for comparison. *
This is a good suggestion that we will follow.
*In Figure 3, it might be helpful for a general audience to include zoomed-in picture of boutons (as in Fig 5B), as some panels appear to have less defined bouton shape. *
We do observe that boutons tend to be less well separated from each other in btz mutants, and will include zoomed-in pictures to document this.
*Is the bouton size different in the mutant in Figure 3? Can this be quantified? *
We do not think that there is a significant difference in bouton size in btz mutants, but we will measure this and include a quantification.
*Fold changes are modest and not very apparent in staining (we acknowledge that this could be due to early developmental time point). Images could better point out differences in WT vs. mutant that are not readily apparent to those outside the fly neurodevelopment audience. *
Because of the inherent variability in synapse shape, it can be difficult to appreciate changes in bouton number from a single image. However, our quantifications show that the changes are consistent and significant.
*Fig 4 NMJs are shown on different scale (more zoomed in) than in Figure 3, and differences are bit easier to see at this scale. Presenting Fig 3 on this scale might help the reader with visualizing the differences in WT versus mutant. *
We will crop the images in Figure 3 so as to show them at the same scale as in Figure 4.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
Summary
Ho et al. describes the developmental functions of the Drosophila Casc3 ortholog, Barentsz (Btz) using in vivo loss-of-function and rescue experiments in Drosophila larvae. In this study, the authors find that loss of Casc3 contributes to neuromuscular defects in the larval fly. Utilizing transgenics of WT and EJC interaction-defective mutants, they demonstrate that Btz has both EJC-dependent and independent functions in the larval neuromuscular junction, wherein muscle defects are EJC dependent and synaptic defects are EJC-independent. Using RNA-seq, they find that upregulated mRNAs include those that belong to the Activin …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
Summary
Ho et al. describes the developmental functions of the Drosophila Casc3 ortholog, Barentsz (Btz) using in vivo loss-of-function and rescue experiments in Drosophila larvae. In this study, the authors find that loss of Casc3 contributes to neuromuscular defects in the larval fly. Utilizing transgenics of WT and EJC interaction-defective mutants, they demonstrate that Btz has both EJC-dependent and independent functions in the larval neuromuscular junction, wherein muscle defects are EJC dependent and synaptic defects are EJC-independent. Using RNA-seq, they find that upregulated mRNAs include those that belong to the Activin signaling pathway. They go on to find that the neuromuscular defects in Btz mutants can be attributed to dysregulation of Activin signaling, and are rescued with loss of the Activin ligand, Dawdle (Daw).
Major Comments
Overall, the paper presents well-controlled experiments that support the main conclusions. We propose achievable validation experiments that we believe will strengthen the conclusions of the paper. There is some concern that the magnitude of the effects are overstated, or could be made more apparent to a broader audience (i.e. those in the mRNA regulation field beyond Drosophila geneticists).
• In Figure 1, regarding the validation of rescue constructs: the EJC interaction-defective mutant is based solely on conservation, as all structural/interaction studies cited with Btz bound to EJC have been with human proteins. They use Vasa localization as a readout of EJC-dependent function, but this is indirect and only assesses one aspect of EJC function (localization). Since many of the main conclusions in the paper are predicated on this mutant being EJC-independent, they should validate this with the Drosophila orthologs using immunoprecipitation. They demonstrate the capability of expressing GFP-tagged versions of Casc3 WT and mutant in S2 cells, so this should not be a cumbersome control experiment to include.
• Regarding Figure 3, it could be postulated that the number of boutons would be influenced by the length of axons. Is axon outgrowth accounted for in these experiments? This would influence number of synaptic boutons. Panel F looks very different from panel A in terms of axon length (could this be due to axon outgrowth defect and/or impacted muscle size?) Can quantitation be done also by normalizing to axon length (bouton number/axon length)? Or perhaps this is accounted for in muscle size? If so, this should be explained.
• In Figure 3 quantification: n's vary between genotypes significantly, and this should be explained (e.g. was there a recovery issue between genotypes or just fewer needed for WT-like?).
• In Figure 4 panels B and F (mutants), there appears to be reduced axon outgrowth (see point above). This should be taken into account when expressing bouton number.
• The RNA-seq data (Figure 5) has a potential issue in that they used larvae with a balancer chromosome (Df), which yields a 50% reduction in any genes on that chromosome. They acknowledge this and removed these genes from the analysis, but the concern remains that this still might be a confounding variable (for example, if reduction in any of these genes might disrupt a signaling pathway). We do not think that the RNA-seq needs to be repeated, but we propose that the authors validate these targets using qPCR in their MN-specific btz knockdown system (this way, they can also include magoh and eif4aIII knockdowns for comparison).
Minor comments
Some statements made in the introduction that are not entirely accurate:
• "A fourth core subunit, known as Barentsz (Btz), Cancer susceptibility candidate gene 3 (CASC3), or Metastatic lymph node 51 (MLN51), associates with the complex following the completion of splicing, and is required for the effects of the EJC on translation, NMD and mRNA localization (Chazal et al., 2013; Palacios et al., 2004; Shibuya et al., 2006; van Eeden et al., 2001)."
A recent study indicates that Casc3 is not required for EJC-dependent NMD targets in human cells, but rather enhances NMD on a subset of targets (Gerbracht et al. 2020 NAR). Perhaps "is required" should be changed to "plays a role in cytoplasmic EJC-mediated processes, such as...". It has also been shown that EJC core can assemble without Casc3 (e.g. Ballut et al 2005 NSMB, Gehring et al 2009 PLoS Biol). Previous work from the authors show that Casc3 (Btz) is not necessary for EJC function in pre-mRNA splicing (Roignant et al, 2010 Cell). Further, there exists a population of Casc3 lacking EJCs in human cells (Mabin et al 2018 Cell Reports). Collectively, all this evidence points to Casc3 not being a core EJC subunit.
• "In the mouse brain, haploinsufficiency for Magoh, Rbm8a or Eif4a3 causes severe microcephaly, but complete loss of Casc3 has a much milder effect that can be attributed to developmental delay (Mao et al., 2017; Mao et al., 2016; Mao et al., 2015; Silver et al., 2010)."
From Mao et al 2017: complete loss and hypomorphic mutants were embryonic and perinatally lethal (contrary to what the authors are stating here), while compound mutants and heterozygotes exhibited neurodevelopmental delay. By "milder effects" the authors could also be referring to brain size being proportional to body size in the complete loss homozygotes; either way, this should be clarified.
General minor comments:
• Fly-specific nomenclature could be made more accessible to a broader audience, as the full readership will likely not have expertise in Drosophila genetics. For example, w118, btz2 labels used in figures are not explained anywhere in the manuscript. While the authors do a good job of describing various mutants in a more accessible fashion in the results section, the genotype labels in figures can be better explained in the legends.
• Fig 2 L-N panels might warrant more explanation. Can the mitochondria be counted here? Is there also a difference in volume/morphology that could be quantitated? In Figure 2N, muscle fibers are more densely packed in mutant vs. control; can this be explained?
• In Fig 2, to draw parallels between panels A-K and L-N, it might also be helpful to use the red/yellow arrow system on panel A for comparison.
• In Figure 3, it might be helpful for a general audience to include zoomed-in picture of boutons (as in Fig 5B), as some panels appear to have less defined bouton shape.
• Is the bouton size different in the mutant in Figure 3? Can this be quantified?
• Fold changes are modest and not very apparent in staining (we acknowledge that this could be due to early developmental time point). Images could better point out differences in WT vs. mutant that are not readily apparent to those outside the fly neurodevelopment audience.
• Fig 4 NMJs are shown on different scale (more zoomed in) than in Figure 3, and differences are bit easier to see at this scale. Presenting Fig 3 on this scale might help the reader with visualizing the differences in WT versus mutant.
Significance
Overall, this paper contributes conceptually to understanding EJC-mediated mRNA regulation during development. The contribution here is incremental, but meaningful in terms of defining the scope of regulation by the EJC and its peripheral factors in various contexts. These findings will likely be of interest to the fields of RNA metabolism and neurodevelopment. It also adds to the existing work suggesting Casc3 may have additional functions outside of the EJC (e.g. Mao et al. 2017 RNA, Baguet et al 2007 J Cell Sci, Cougot et al. 2014 J Cell Sci); while these previous studies have suggested Casc3 roles in development and mRNA localization/granule formation that are different from the EJC core proteins, this study more directly tests an EJC-independent role in mRNA regulation of specific targets. Further addressing the molecular basis of this regulation will be outside the scope of this article but will be of interest to the field.
We are molecular biologists who study NMD and are thus equipped to address the EJC-related molecular functions and impact on the transcriptome. We do not have expertise in Drosophila genetics or neurobiology, and thus cannot critically evaluate the specific genetic approaches used or anatomy presented to the full extent. We have, however, pointed out areas that need elaboration regarding the genetic approaches and/or presentation of data that may be unfamiliar to a broader audience (i.e. the RNA metabolism field).
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The Ho et al. manuscript defines developmental functions for Barentsz (Btz), a core subunit of the EJC. While other EJC components, such as eIF4AIII, have been shown to have EJC-independent functions, it has not been clear whether Btz also acted independently of this multi-protein complex. The authors make use of two Btz genomic constructs, a wild-type transgene (Btz-WT) and a transgene carrying mutations in the two eIF4AIII-interacting residues (Btz-HD) to rigorously whether or not Btz has any functions independent of the EJC. Interestingly, they show that while Btz-HD does not rescue Btz functions in the ovary or the muscle, it …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
The Ho et al. manuscript defines developmental functions for Barentsz (Btz), a core subunit of the EJC. While other EJC components, such as eIF4AIII, have been shown to have EJC-independent functions, it has not been clear whether Btz also acted independently of this multi-protein complex. The authors make use of two Btz genomic constructs, a wild-type transgene (Btz-WT) and a transgene carrying mutations in the two eIF4AIII-interacting residues (Btz-HD) to rigorously whether or not Btz has any functions independent of the EJC. Interestingly, they show that while Btz-HD does not rescue Btz functions in the ovary or the muscle, it does rescue Btz functions at the larval NMJ. They back up the conclusion that Btz activity at the NMJ is independent of the EJC by showing that the growth phenotype observed in Btz mutants is not shared by mutants in other EJC components. How does Btz regulate NMJ development? The authors performed an RNAseq experiment and found that several components of an Activin/TGFB pathway. Strikingly, they find that Activin overexpression rescues the NM phenotype in Btz mutants, consistent with its identification in the RNAseq analysis.
This is a very logical and well-constructed paper. The results are well-controlled and convincing. Overall, the manuscript was a delight to read and makes an important contribution to dissecting the function of RNA-binding/associated proteins in neuronal development. I have only a few comments that could be considered prior to publication.
Minor comments:
- In Fig. 1, the authors show that Btz-WT, but not Btz-HD, localizes to the posterior pole of the oocyte. Do the authors see Btz-WT and/or Btz-HD localized to MNs/muscles/glia at the NMJ?
- The mitochondrial phenotypes observed in Btz mutants are striking. But it seems possible that there are defects in overall mitochondrial levels in muscle in addition to defects in their localization. Overall, mitochondrial levels seemed reduced in Btz mutants. Is it possible to do a ATP5A immunoblot in Btz mutants to test whether overall mitochondrial levels are altered?
- ECM proteins are known to be critical for regulating TGFB signaling. That, taken with the multi-tissue genetic requirement for Btz, suggests that Btz might directly regulate either Ltl or Frac RNA, given that these ECM proteins are likely deposited by multiple cell types.
Significance
This paper establishes novel functions for the EJC complex protein Btz, and also delineates which functions depend on the EJC and which are independent. This is significant because there is intense interest in how post transcriptional regulation contributes to neuronal development. The paper fits with a body of literature dissecting neuronal functions for EJC proteins. It represents an important addition to this body of work.
The audience will be molecular neuroscientists, especially those with interests in novel genetic regulatory mechanisms.
My expertise is in developmental genetics and molecular neurobiology.
-
