Analysis of SMAD1/5 target genes in a sea anemone reveals ZSWIM4-6 as a novel BMP signaling modulator

  1. Paul Knabl
  2. Alexandra Schauer
  3. Autumn P Pomreinke
  4. Bob Zimmermann
  5. Katherine W Rogers
  6. Daniel Čapek
  7. Patrick Müller
  8. Grigory Genikhovich

  • Curated by eLife

    eLife logo

    eLife assessment

    This important work presents a systematic survey of downstream target genes of the BMP pathway during body-axis establishment of the cnidarian Nematostella vectensis. BMP is a well-known developmental regulator, and this work identifies a previously unknown array of downstream targets. Combining genomic approaches and genetic manipulations, the authors present convincing evidence that Zswim4-6 acts as a negative feedback regulator of BMP activity in Nematostella. The authors also test a zebrafish homologue in over-expression assays and show solid evidence that it too dampens BMP signaling activity, leading to the suggestion that zswim4-6 is a conserved regulator of BMP signaling. This work will be of interest to researchers in the fields of both developmental biology and evo-devo.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

BMP signaling has a conserved function in patterning the dorsal-ventral body axis in Bilateria and the directive axis in anthozoan cnidarians. So far, cnidarian studies have focused on the role of different BMP signaling network components in regulating pSMAD1/5 gradient formation. Much less is known about the target genes downstream of BMP signaling. To address this, we generated a genome-wide list of direct pSMAD1/5 target genes in the anthozoan Nematostella vectensis , several of which were conserved in Drosophila and Xenopus . Our ChIP-seq analysis revealed that many of the regulatory molecules with documented bilaterally symmetric expression in Nematostella are directly controlled by BMP signaling. We identified several so far uncharacterized BMP-dependent transcription factors and signaling molecules, whose bilaterally symmetric expression may be indicative of their involvement in secondary axis patterning. One of these molecules is zswim4-6 , which encodes a novel nuclear protein that can modulate the pSMAD1/5 gradient and potentially promote BMP-dependent gene repression.

Article activity feed

  1. Version published to 10.7554/elife.80803 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    This thorough study expands our understanding of BMP signaling, a conserved developmental pathway, involved in processes diverse such as body patterning and neurogenesis. The authors applied multiple, state-of-art strategies to the anthozoan Nematostella vectensis in order to first identify the direct BMP signaling targets - bound by the activated pSMAD1/5 protein - and then dissect the role of a novel pSMAD1/5 gradient modulator, zwim4-6. The list of target genes features multiple developmental regulators, many of which are bilaterally expressed, and which are notably shared between Drosophila and Xenopus. The analysis identified in particular zswim4-6 a novel nuclear modulator of the BMP pathway conserved also in vertebrates. A combination of both loss-of-function (injection of antisense morpholino oligonucleotide, CRISPR/Cas9 knockout, expression of dominant negative) and gain-of-function assays, and of transcriptome sequencing identified that zwim acts as a transcriptional repression of BMP signaling. Functional manipulation of zswim5 in zebrafish shows a conserved role in modulating BMP signaling in a vertebrate.

    The particular strength of the study lies in the careful and thorough analysis performed. This is solid developmental work, where one clear biological question is progressively dissected, with the most appropriate tools. The functional results are further validated by alternative approaches. Data is clearly presented and methods are detailed. I have a couple of comments.

    1. I was intrigued - as the authors - by the fact that the ChiP-Seq did not identify any known BMP ligand bound by pSMAD1/5. Are these genes found in the published ChiP-Seq data of the other species used for the comparative analysis? One hypothesis could be that there is a change in the regulatory interactions and that the initial set-up of the gradient requires indeed a feedback loop, which is then turned off at later gastrula. In this case, immunoprecipitation at early gastrula, prior to the set-up of the pSMAD1/5 gradient, could reveal a different scenario. Alternately, the regulation could be indirect, for example, through RGM, an additional regulator of BMP signaling expressed on the side of lower BMP activity, which is among the targets of the ChiP-Seq. This aspect could be discussed. Additionally, even if this is perhaps outside the scope of this study, I think it would be informative to further assess the effect of ZSWIM manipulation on RGM (and vice versa).

    Indeed, BMP genes are direct BMP signaling targets in Drosophila (dpp) (Deignan et al., 2016, https://doi.org/10.1371/journal.pgen.1006164) and frog (bmp2, bmp4, bmp5, bmp7) (Stevens et al., 2021, https://doi.org/10.1242/dev.145789). Of all these ligands, only the dorsally expressed Xenopus bmp2 is repressed by BMP signaling, while another dorsally expressed Xenopus BMP gene admp is not among the direct targets. All other BMP genes listed here are expressed in the pMad/pSMAD1/5/8-positive domain and are activated by BMP signaling.

    In Nematostella, we do not find BMP genes among the ChIP-Seq targets, but this is not that surprising considering the dynamics of the bmp2/4, bmp5-8 and chordin expression, as well as the location of the pSMAD1/5-positive cells. In late gastrulae/early planulae, Chordin appears to be shuttling BMP2/4 and BMP5-8 away from their production source and over to the gdf5-like side of the directive axis (Genikhovich et al., 2015; Leclere and Rentsch, 2014). By 4 dpf, chordin expression stops, and BMP2/4 and BMP5-8 start to be both expressed AND signal in the mesenteries. If bmp2/4 and bmp5-8 expression were directly suppressed by pSMAD1/5 (as is the case chordin or rgm expression), this mesenterial expression would not be possible. Therefore, in our opinion, it is most likely that at late gastrula and early planula the regulation of bmp2/4 and bmp5-8 expression by BMP signaling is indirect. We do not have an explanation for why gdf5-like (another BMP gene expressed on the “high pSMAD1/5” side) is not retrieved as a direct BMP target in our ChIP data. Since we do not understand well enough how BMP gene expression is regulated, we do not discuss this at length in the manuscript.

    As the Reviewer suggested, we analyzed the effect of ZSWIM4-6 KD on the expression of rgm. Expectedly, since it is expressed on the “low BMP side”, its expression was strongly expanded (Figure 6 - Figure Supplement 4)

    1. I do not fully understand the rationale behind the choice of performing the comparative assays in zebrafish: as the conservation was initially identified in Xenopus, I would have expected the experiment to be performed in frog. Furthermore, reading the phylogeny (Figure 4A), it is not obvious to me why ZSWIM5 was chosen for the assay (over the other paralog ZSWIM6). Could the Authors comment on this experiment further?

    The comparison was done in zebrafish because we were planning to generate zswim5 mutants, whose analysis is currently in progress. ZSWIM6 is not expressed at the developmental stages we were interested in, while ZSWIM5 was, based on available zebrafish expression data (White et al., 2017):

    Reviewer #2 (Public Review):

    The authors provide a nice resource of putative direct BMP target genes in Nematostella vectensis by performing ChIP-seq with an anti-pSmad1/5 antibody, while also performing bulk RNA-seq with BMP2/4 or GDF5 knockdown embryos. Genes that exhibit pSmad1/5 binding and have changes in transcription levels after BMP signaling loss were further annotated to identify those with conserved BMP response elements (BREs). Further characterization of one of the direct BMP target genes (zswim4-6) was performed by examining how expression changed following BMP receptor or ligand loss of function, as well as how loss or gain of function of zswim4-6 affected development and BMP signaling. The authors concluded that zswim4-6 modulates BMP signaling activity and likely acts as a pSMAD1/5 dependent co-repressor. However, the mechanism by which zswim4-6 affects the BMP gradient or interacts with pSMAD1/5 to repress target genes is not clear. The authors test the activity of a zswim4-6 homologue in zebrafish (zswim5) by over-expressing mRNA and find that pSMAD1/5/9 labeling is reduced and that embryos have a phenotype suggesting loss of BMP signaling, and conclude that zswim4-6 is a conserved regulator of BMP signaling. This conclusion needs further support to confirm BMP loss of function phenotypes in zswim5 over-expression embryos.

    Major comments

    1. The BMP direct target comparison was performed between Nematostella, Drosophila, and Xenopus, but not with existing data from zebrafish (Greenfeld 2021, Plos Biol). Given the functional analysis with zebrafish later in the paper it would be nice to see if there are conserved direct target genes in zebrafish, and in particular, is zswim5 (or other zswim genes) are direct targets. Since conservation of zswim4-6 as a direct BMP target between Nematostella and Xenopus seemed to be part of the rationale for further functional analysis, it would also be nice to know if this is a conserved target in zebrafish.

    Thank you for the suggestion. In the paper by Greenfeld et al., 2021, zebrafish zswim5 was downregulated approximately 2.4x in the bmp7 mutant at 6 hpf, while zswim6 was barely expressed and not affected at this stage. We added this information to the text of the manuscript. Expression of several other zebrafish zswim genes was also affected in the bmp7 mutant, but these genes do not appear relevant for our study since their corresponding orthologs are not identified as pSMAD1/5 ChIP-Seq targets in Nematostella. Notably, zebrafish zzswim5 is not clearly differentially expressed in BMP or Chd overexpression conditions (See Supplementary file 1 in Rogers et al. 2020). Importantly, in the paper, we wanted to compare ChiP-Seq data with ChIP-Seq data, however, unfortunately, no ChIP-Seq data for pSMAD1/5/8 is currently available for zebrafish, thus precluding comparisons.

    Related to this, in the discussion it is mentioned that zswim4/6 is also a direct BMP target in mouse hair follicle cells, but it wasn't obvious from looking at the supplemental data in that paper where this was drawn from.

    Please see Supplementary Table 1, second Excel sheet labeled “Mx ChIP_Seq” in Genander et al., 2014, https://doi.org/10.1016/j.stem.2014.09.009. Zswim4 has a single pSMAD1 peak associated with it, Zswim6 has two.

    1. The loss of zswim4-6 function via MO injection results in changes to pSmad1/5 staining, including a reduction in intensity in the endoderm and gain of intensity in the ectoderm, while over-expression results in a loss of intensity in the ectoderm and no apparent change in the endoderm. While this is interesting, it is not clear how zswim4-6 is functioning to modify BMP signaling, and how this might explain differential effects in ectoderm vs. endoderm. Is the assumption that the mechanism involves repression of chordin? And if so one could test the double knockdown of zswim4-6 and chordin and look for the rescue of pSad1/5 levels or morphological phenotype.

    We do not think that the mechanism of the ZSWIM4-6 action is via repression of Chordin. As loss of chordin leads to the loss of pSMAD1/5 in Nematostella (Genikhovich et al., 2015), the proposed experiment is, unfortunately, not feasible to test this hypothesis. Currently, we see two distinct effects of the modulation of zswim4-6 expression. First, it affects the pSMAD1/5 gradient, possibly by destabilizing nuclear SMAD1/5, as has been proposed by Wang et al., 2022 for the vertebrate Zswim4. This is in line with our results shown on Fig. 6C-F’ and Fig. 6-Figure supplement 3. In our opinion, the reaction of the genes expressed on the “high BMP” side of the directive axis to the overexpression or KD of ZSWIM4-6 (Fig. 6I-K’, 6N-P’) can be explained by these changes in the pSMAD1/5 signaling intensity. Secondly, zswim4-6 appears to promote pSMAD1/5-mediated gene repression. This is in line with the reaction of the genes expressed on the “low BMP” side of the directive axis (Fig. 6G-H’, 6L-M’, Fig. 6-Figure Supplement 4). These genes are repressed by BMP signaling, but they expand their expression upon zswim4-6 KD in spite of the increased pSMAD1/5. Our ChiP experiment (Fig. 6Q) supports this view.

    1. Several experiments are done to determine how zswim4-6 expression responds to the loss of function of different BMP ligands and receptors, with the conclusion being that swim4-6 is a BMP2/4 target but not a GDF5 target, with a lot of the discussion dedicated to this as well. However, the authors show a binary response to the loss of BMP2/4 function, where zswim4-6 is expressed normally until pSmad1/5 levels drop low enough, at which point expression is lost. Since the authors also show that GDF5 morphants do not have as strong a reduction in pSmad1/5 levels compared to BMP2/4 morphants, perhaps GDF5 plays a positive but redundant role in swim4-6 expression. To test this possibility the authors could inject suboptimal doses of BMP2/4 MO with GDF5 MO and look for synergy in the loss of zswim4-6 expression.

    Thanks for this great suggestion! We performed this experiment (Fig. 5H’’-L) and indeed, a suboptimal dose of BMP2/4MO + GDF5lMO results in a complete radialization of the embryo and abolished zswim4–6, similar to the effect of a high dose of BMP2/4. This result suggests that rather than being a ligand-specific signaling function, GDF5-like signaling alone still provides sufficiently high pSmad1/5 levels to activate zswim4-6 expression to apparent wildtype levels, demonstrating the sensitivity of this gene to even very low amounts of BMP signaling.

    1. The zswim4-6 morphant embryos show increased expression of zswim4-6 mRNA, which is said to indicate that zswim4-6 negatively regulates its own expression. However in zebrafish translation blocking MOs can sometimes stabilize target transcripts, causing an artifact that can be mistakenly assumed to be increased transcription (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7162184/). Some additional controls here would be warranted for making this conclusion.

    Thanks for raising this important experimental consideration. To-date, we do not have any evidence for MO-mediated transcript stabilization in Nematostella, and we have not found such data in the literature on models other than zebrafish. mRNA stabilization by the MO also seemed unlikely because we were unable to KD zswim4-6 using several independent shRNAs - an effect we frequently observe with genes, whose activity negatively regulates their own expression. However, to test the possibility that zswim4-6MO binding stabilizes zswim4-6 mRNA, we injected mRNA containing the zswim4-6MO recognition sequence followed by the mCherry coding sequence (zswim4-6MO-mCherrry) with either zswim4-6MO or control MO. We could clearly detect mCherry fluorescence at 1 dpf if control MO was co-injected with the mRNA, but not if zswim4-6MO was coninjected with the mRNA. At 2 dpf (the stage at which we showed upregulation of zswim4-6 upon zswim4-6MO injection on Fig. 6I-I’), zswim4-6MO-mCherrry mRNA was undetectable by in situ hybridization with our standard FITC-labeled mCherry probe independent of whether zswim4-6MO-mCherrry mRNA was co-injected with the control MO or ZSWIM4-6MO, while hybridization with the FITC-labeled FoxA probe worked perfectly.

    Author response image 1.

    We are currently offering two alternative hypothesis for the observed increase in zswim4-6 levels in the paper rather than stating explicitly that ZSWIM4-6 negatively regulates its own expression: “The KD of zswim4-6 translation resulted in a strong upregulation of zswim4-6 transcription, especially in the ectoderm, suggesting that ZSWIM4-6 might either act as its own transcriptional repressor or that zswim4-6 transcription reacts to the increased ectodermal pSMAD1/5 (Fig. 6I-I’).” Given the sensitivity of zswim4-6 to even the weakest pSMAD1/5 signal (zswim4/6 is expressed upon GDF5-like KD, which drastically reduces pSMAD1/5 signaling intensity (see Fig. 1 and 2 in Genikhovich et al., 2015, http://doi.org/10.1016/j.celrep.2015.02.035 and Fig. 6-Figure supplement 3 of this paper), the latter option (that it reacts to the increased ectodermal pSMAD1/5) is, in our opinion, clearly the more probable one.

    1. Zswim4-6 is proposed to be a co-repressor of pSmad1/5 targets based on the occupancy of zswim4-6 at the chordin BRE (which is normally repressed by BMP signaling) and lack of occupancy at the gremlin BRE (normally activated by BMP signaling). This is a promising preliminary result but is based only on the analysis of two genes. Since the authors identified BREs in other direct target genes, examining more genes would better support the model.

    We suggest that ZSWIM4-6 may be a co-repressor of pSMAD1/5 targets because it is a nuclear protein (Fig. 4G), whose knockdown results in the expansion of the ectodermal expression of several genes repressed by pSMAD1/5 in spite of the expansion of pSMAD1/5 itself (Fig. 6G-H’, 6L-M’, Fig. 6-Figure Supplement 4). Our limited ChIP analysis supports this idea by showing that ZSWIM4-6 is bound to the pSMAD1/5 site of chordin (repressed by pSMAD1/5) but not on gremlin (activated by pSMAD1/5). We agree that adding the analysis of more targets in order to challenge our hypothesis would be good. However, given technical limitations (having to inject many thousands of eggs with the EF1a::ZSWIM4-6-GFP plasmid in order to get enough nuclei to extract sufficient immunoprecipitated chromatin for qPCR on 3 genes (chordin, gremlin, GAPDH) for each biological replicate, it is currently unfortunately not feasible to test more genes. It will be of great interest for follow up studies to generate a knock-in line with tagged zswim4-6 to analyze target binding on a genome-wide scale. We stress in the discussion that currently the power of our conclusion is low.

    1. The rationale for further examination of zswim4-6 function in Nematostella was based in part on it being a conserved direct BMP target in Nematostella and Xenopus. The analysis of zebrafish zswim5 function however does not examine whether zswim5 is a BMP target gene (direct or indirect). BMP inhibition followed by an in situ hybridization for zswim5 would establish whether its expression is activated downstream of BMP.

    In the paper by Greenfeld et al., 2021, zebrafish zswim5 was downregulated approximately 2.4x in the bmp7 mutant at 6 hpf. However, this gene was not among the 57 genes, which were considered to be direct BMP targets because their expression was affected by bmp7 mRNA injection into cycloheximide-treated bmp7 mutants (Greenfeld et al., 2021). We added this information to the text of the manuscript.

    1. Although there is a reduction in pSmad1/5/9 staining in zebrafish injected with zswim5 mRNA, it is difficult to tell whether the resulting morphological phenotypes closely resemble zebrafish with BMP pathway mutations (such as bmp2b). More analysis is warranted here to determine whether stereotypical BMP loss of function phenotypes are observed, such as dorsalization of the mesoderm and loss of ventral tail fin.

    We agree, and we have tuned down all zebrafish arguments. Analyses of zswim5 mutants are currently ongoing.

  3. eLife

    eLife assessment

    This important work presents a systematic survey of downstream target genes of the BMP pathway during body-axis establishment of the cnidarian Nematostella vectensis. BMP is a well-known developmental regulator, and this work identifies a previously unknown array of downstream targets. Combining genomic approaches and genetic manipulations, the authors present convincing evidence that Zswim4-6 acts as a negative feedback regulator of BMP activity in Nematostella. The authors also test a zebrafish homologue in over-expression assays and show solid evidence that it too dampens BMP signaling activity, leading to the suggestion that zswim4-6 is a conserved regulator of BMP signaling. This work will be of interest to researchers in the fields of both developmental biology and evo-devo.

  4. eLife

    Reviewer #1 (Public Review):

    This thorough study expands our understanding of BMP signaling, a conserved developmental pathway, involved in processes diverse such as body patterning and neurogenesis. The authors applied multiple, state-of-art strategies to the anthozoan Nematostella vectensis in order to first identify the direct BMP signaling targets - bound by the activated pSMAD1/5 protein - and then dissect the role of a novel pSMAD1/5 gradient modulator, zwim4-6. The list of target genes features multiple developmental regulators, many of which are bilaterally expressed, and which are notably shared between Drosophila and Xenopus. The analysis identified in particular zswim4-6 a novel nuclear modulator of the BMP pathway conserved also in vertebrates. A combination of both loss-of-function (injection of antisense morpholino oligonucleotide, CRISPR/Cas9 knockout, expression of dominant negative) and gain-of-function assays, and of transcriptome sequencing identified that zwim acts as a transcriptional repression of BMP signaling. Functional manipulation of zswim5 in zebrafish shows a conserved role in modulating BMP signaling in a vertebrate.
    The particular strength of the study lies in the careful and thorough analysis performed. This is solid developmental work, where one clear biological question is progressively dissected, with the most appropriate tools. The functional results are further validated by alternative approaches. Data is clearly presented and methods are detailed.

    I have a couple of comments.

    1. I was intrigued - as the authors - by the fact that the ChiP-Seq did not identify any known BMP ligand bound by pSMAD1/5. Are these genes found in the published ChiP-Seq data of the other species used for the comparative analysis? One hypothesis could be that there is a change in the regulatory interactions and that the initial set-up of the gradient requires indeed a feedback loop, which is then turned off at later gastrula. In this case, immunoprecipitation at early gastrula, prior to the set-up of the pSMAD1/5 gradient, could reveal a different scenario. Alternately, the regulation could be indirect, for example, through RGM, an additional regulator of BMP signaling expressed on the side of lower BMP activity, which is among the targets of the ChiP-Seq. This aspect could be discussed. Additionally, even if this is perhaps outside the scope of this study, I think it would be informative to further assess the effect of ZSWIM manipulation on RGM (and vice versa).
    2. I do not fully understand the rationale behind the choice of performing the comparative assays in zebrafish: as the conservation was initially identified in Xenopus, I would have expected the experiment to be performed in frog. Furthermore, reading the phylogeny (Figure 4A), it is not obvious to me why ZSWIM5 was chosen for the assay (over the other paralog ZSWIM6). Could the Authors comment on this experiment further?
  5. eLife

    Reviewer #2 (Public Review):

    The authors provide a nice resource of putative direct BMP target genes in Nematostella vectensis by performing ChIP-seq with an anti-pSmad1/5 antibody, while also performing bulk RNA-seq with BMP2/4 or GDF5 knockdown embryos. Genes that exhibit pSmad1/5 binding and have changes in transcription levels after BMP signaling loss were further annotated to identify those with conserved BMP response elements (BREs). Further characterization of one of the direct BMP target genes (zswim4-6) was performed by examining how expression changed following BMP receptor or ligand loss of function, as well as how loss or gain of function of zswim4-6 affected development and BMP signaling. The authors concluded that zswim4-6 modulates BMP signaling activity and likely acts as a pSMAD1/5 dependent co-repressor. However, the mechanism by which zswim4-6 affects the BMP gradient or interacts with pSMAD1/5 to repress target genes is not clear. The authors test the activity of a zswim4-6 homologue in zebrafish (zswim5) by over-expressing mRNA and find that pSMAD1/5/9 labeling is reduced and that embryos have a phenotype suggesting loss of BMP signaling, and conclude that zswim4-6 is a conserved regulator of BMP signaling. This conclusion needs further support to confirm BMP loss of function phenotypes in zswim5 over-expression embryos.

    Major comments

    1. The BMP direct target comparison was performed between Nematostella, Drosophila, and Xenopus, but not with existing data from zebrafish (Greenfeld 2021, Plos Biol). Given the functional analysis with zebrafish later in the paper it would be nice to see if there are conserved direct target genes in zebrafish, and in particular, is zswim5 (or other zswim genes) are direct targets. Since conservation of zswim4-6 as a direct BMP target between Nematostella and Xenopus seemed to be part of the rationale for further functional analysis, it would also be nice to know if this is a conserved target in zebrafish.

    Related to this, in the discussion it is mentioned that zswim4/6 is also a direct BMP target in mouse hair follicle cells, but it wasn't obvious from looking at the supplemental data in that paper where this was drawn from.

    2. The loss of zswim4-6 function via MO injection results in changes to pSmad1/5 staining, including a reduction in intensity in the endoderm and gain of intensity in the ectoderm, while over-expression results in a loss of intensity in the ectoderm and no apparent change in the endoderm. While this is interesting, it is not clear how zswim4-6 is functioning to modify BMP signaling, and how this might explain differential effects in ectoderm vs. endoderm. Is the assumption that the mechanism involves repression of chordin? And if so one could test the double knockdown of zswim4-6 and chordin and look for the rescue of pSad1/5 levels or morphological phenotype.

    3. Several experiments are done to determine how zswim4-6 expression responds to the loss of function of different BMP ligands and receptors, with the conclusion being that swim4-6 is a BMP2/4 target but not a GDF5 target, with a lot of the discussion dedicated to this as well. However, the authors show a binary response to the loss of BMP2/4 function, where zswim4-6 is expressed normally until pSmad1/5 levels drop low enough, at which point expression is lost. Since the authors also show that GDF5 morphants do not have as strong a reduction in pSmad1/5 levels compared to BMP2/4 morphants, perhaps GDF5 plays a positive but redundant role in swim4-6 expression. To test this possibility the authors could inject suboptimal doses of BMP2/4 MO with GDF5 MO and look for synergy in the loss of zswim4-6 expression.

    4. The zswim4-6 morphant embryos show increased expression of zswim4-6 mRNA, which is said to indicate that zswim4-6 negatively regulates its own expression. However in zebrafish translation blocking MOs can sometimes stabilize target transcripts, causing an artifact that can be mistakenly assumed to be increased transcription (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7162184/). Some additional controls here would be warranted for making this conclusion.

    5. Zswim4-6 is proposed to be a co-repressor of pSmad1/5 targets based on the occupancy of zswim4-6 at the chordin BRE (which is normally repressed by BMP signaling) and lack of occupancy at the gremlin BRE (normally activated by BMP signaling). This is a promising preliminary result but is based only on the analysis of two genes. Since the authors identified BREs in other direct target genes, examining more genes would better support the model.

    6. The rationale for further examination of zswim4-6 function in Nematostella was based in part on it being a conserved direct BMP target in Nematostella and Xenopus. The analysis of zebrafish zswim5 function however does not examine whether zswim5 is a BMP target gene (direct or indirect). BMP inhibition followed by an in situ hybridization for zswim5 would establish whether its expression is activated downstream of BMP.

    7. Although there is a reduction in pSmad1/5/9 staining in zebrafish injected with zswim5 mRNA, it is difficult to tell whether the resulting morphological phenotypes closely resemble zebrafish with BMP pathway mutations (such as bmp2b). More analysis is warranted here to determine whether stereotypical BMP loss of function phenotypes are observed, such as dorsalization of the mesoderm and loss of ventral tail fin.

  6. eLife

    Reviewer #3 (Public Review):

    To identify direct targets of BMP signal in Nematostella, the authors performed ChIP-seq using an antibody against phosphorylated SMAD1/5 (pSMAD1/5) at late gastrula and late planula stages. In accordance with the highly dynamic BMP activity detected using immunofluorescence, pSMAD1/5 binding profiles change drastically as the larvae develop, with only a fraction of target genes shared between these two time points. The authors then followed up with RNA-seq in control versus BMP2/4 KD embryos and identified significant expression changes in many transcription factors and signaling molecules, including the Gbx-Hox genes, which are known to regulate endoderm patterning. These results, in conjunction with a thorough validation using in situ hybridization, strongly support the authors' claim that the BMP signal in Nematostella directly controls a small set of second-tier targets which in turn execute the morphogenic functions.

    Next, the authors explored the conservation of BMP downstream targets by intersecting their candidate list with two published datasets from Drosophila (2-3hpf) and Xenopus (NF20 stage). Results from such an analysis should be taken with a grain of salt, as the developmental time points and biological context examined here are not necessarily comparable. Furthermore, whole genome duplication in vertebrates means multiple copies of transcription factors and signaling molecules belonging to the same family exist in Xenopus, making a homology-based comparison difficult. A handful of shared targets were identified between different species, although no statics were provided to support the significance of such an observation.

    The authors then focused on Zswim4-6, one of the identified BMP targets with a high pSMAD1/5 enrichment level, and dissected its regulatory properties on BMP activity. Using complimentary knockdown and overexpression experiments, the authors rigorously demonstrated that Zswim4-6 is crucial to maintaining the proper pSMAD1/5 gradient at the late gastrula stage. By ectopically overexpressing a GFP tagged form of Zswim4-6, the authors performed low input ChIP-qPCR and confirmed that Zswim4-6 selectively binds to a regulatory region of a BMP-repressed gene, suggesting it may function as a co-repressor for certain BMP targets.
    Lastly, the authors examined the effect of Zswim5, a bilaterian homolog of Zswim4-6, during zebrafish D-V axis establishment. Overexpression of Zswim5 leads to a dampened pSMAD1/5 gradient and dorsalization of the fish larvae, hinting at the possibility that Zswim5 may function as a BMP modulator in zebrafish as well.

    Overall, despite certain caveats, the experimental evidence supports the claims from the authors that Zswim4-6 is directly activated by and reciprocally modulates the BMP activity in Nematostella. The work presented here opens exciting possibilities to further dissect the gene regulatory networks downstream of the cnidarian BMP signaling pathway and expands our knowledge on the evolution of a bilaterally symmetric body plan.

  7. Version published to 10.1101/2022.06.03.494682v2 on bioRxiv
  8. Version published to 10.1101/2022.06.03.494682v1 on bioRxiv