Cnidarian hair cell development illuminates an ancient role for the class IV POU transcription factor in defining mechanoreceptor identity

  1. Ethan Ozment
  2. Arianna N Tamvacakis
  3. Jianhong Zhou
  4. Pablo Yamild Rosiles-Loeza
  5. Esteban Elías Escobar-Hernandez
  6. Selene L Fernandez-Valverde
  7. Nagayasu Nakanishi

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This manuscript focusses on a little studied, but highly interesting presumptive mechanosensory cell type in cnidarians, the 'hair cell'. The work shows that the POU-IV transcription factor is required for the maturation of this cell type in the sea anemone Nematostella vectensis. Because POU-IV transcription factors also play essential roles in the differentiation of mechanoreceptors in many bilaterian phyla, this suggests an evolutionarily ancient role of POU-IV in regulating mechanosensory identity. This study will hence be of great interest to developmental biologists and evolutionary biologists who are interested in the developmental evolution of neuronal cell types.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1, Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Although specialized mechanosensory cells are found across animal phylogeny, early evolutionary histories of mechanoreceptor development remain enigmatic. Cnidaria (e.g. sea anemones and jellyfishes) is the sister group to well-studied Bilateria (e.g. flies and vertebrates), and has two mechanosensory cell types – a lineage-specific sensory effector known as the cnidocyte, and a classical mechanosensory neuron referred to as the hair cell. While developmental genetics of cnidocytes is increasingly understood, genes essential for cnidarian hair cell development are unknown. Here, we show that the class IV POU homeodomain transcription factor (POU-IV) – an indispensable regulator of mechanosensory cell differentiation in Bilateria and cnidocyte differentiation in Cnidaria – controls hair cell development in the sea anemone cnidarian Nematostella vectensis. N. vectensis POU-IV is postmitotically expressed in tentacular hair cells, and is necessary for development of the apical mechanosensory apparatus, but not of neurites, in hair cells. Moreover, it binds to deeply conserved DNA recognition elements, and turns on a unique set of effector genes – including the transmembrane receptor-encoding gene polycystin 1 – specifically in hair cells. Our results suggest that POU-IV directs differentiation of cnidarian hair cells and cnidocytes via distinct gene regulatory mechanisms, and support an evolutionarily ancient role for POU-IV in defining the mature state of mechanosensory neurons.

Article activity feed

  1. eLife

    Author Response:

    Reviewer #1:

    The paper demonstrates the role of Pou domains for various sensory cells. Using CRISPR to delete the gene, the authors show an incomplete deletion of sensory cells. Further evidence shows problems with the formation of mechanosensory cells. Overall, the presentation is clear but can be expanded by adding the role of bHLH genes (Atoh1 is upstream of Pou4f3). If possible, I suggest expanding the role of TMC as it is the main receptor in mammalian hair cells that connects to the stereocilia. Please note that the cnidarian organization is a central kinocilium surrounded by microvilli, comparable to choanoflagellates. This paper is a great original presentation but it could provide a broader perspective by expanding on the evolution of Pou IV and by adding a discussion of the evolution of bHLH, Myc and TMC in order to provide this broader perspective.

    We thank the reviewer for the suggestions to include and expand the discussion about bHLH (and other) factors that may have roles in cnidarian hair cell development. We agree that identification of upstream factors to POU-IV is of particular importance, and that bHLH genes related to the Atonal family are reasonable candidates based on comparative data from bilaterians. In the revised version of the manuscript, we have therefore added a discussion about the evolution of bHLH factors, and the potential role of atonal-like bHLH genes in controlling POU-IV expression in the context of cnidarian hair cell development. We have also emphasized the relevance of such studies to further illuminating the evolution of animal mechanoreceptor development and its gene regulatory mechanisms.

    Reviewer #2:

    Whereas the role of POU-IV for the differentiation of cnidocytes and other neurons of Nematostella has been previously characterized (Tourniere et al., 2020), the present study extends previous reports by specifically addressing the role of POU-IV for the so-called "hair cells" of Nematostella (not to be confused with the hair cells of the vertebrate inner ear and lateral line). These presumably mechanosensory hair cells are identified here as postmitotic neurons, which are ciliated and carry a collar of stereovilli - actin-filled microvilli with a long actin-rich rootlet. Using CRISPR/Cas9 based gene editing, the study shows that transgenic animals, in which the POU-IV gene has been disrupted, become touch insensitive. While hair cells can still be identified in these POU-IV mutants, they lack the stereovillar rootlets suggesting that POU-IV is required for proper hair cells maturation, but dispensable for early steps of hair cell specification and differentiation. The study then uses ChIP-Seq to identify direct target genes of POU-IV in Nematostella and to characterize a POU-IV binding motif, which turned to be evolutionarly highly conserved with POU-IV binding motifs in bilaterians. Comparison of the ChIP-Seq data with published bulk and single-cell transcriptome data indicated that POU-IV activates substantially different sets of effector genes (but no regulatory genes) in hair cells and cnidocytes, and identified polycystin1 as a hair cell-specific direct target of POU-IV. Taken together, this suggests that POU-IV had an evolutionary ancient role as a terminal selector gene for mechanosensory neurons, which predated the split between cnidarian and bilaterian lineages but that its function diverged (e.g. by the acquisition of new target genes) during the evolution of cnidocytes as a novel cell type in cnidarians.

    Combining gene editing with sequencing and with careful morphological and behavioral characterisation of cellular phenotypes, the study provides valuable new insights into the evolution of sensory neurons. POU-IV class transcription factors have previously been implicated in the specification of mechano- and chemosensory neurons in bilaterians. The present study together with the previous study of Tourniere et al. (2020) now suggests an even deeper evolutionary origin of this cell type in the last common ancestor of eumetazoans. The paper is very well written and the results are beautifully documented. The authors are overall cautious and conservative in the conclusions drawn from their findings. However, two points deserve a more critical discussion, first, the question of which sensory modality is mediated by the hair cells (are these dedicated mechanoreceptors or possibly multimodal cells?), and second, the question whether POU-IV serves as transcriptional activator or repressor in cnidocytes.

    We thank the reviewer for an excellent summary of our work and for the questions. In response to the first question, we do not rule out the possibility that hair cells could be multimodal sensory cells, and have added a discussion that raises this possibility in the revision. In response to the second question, POU-IV should be regarded as a leaky repressor in cnidocytes. We have modified the language to clarify this point in the revised version of the manuscript.

    Reviewer #3:

    In this manuscript, Ozmet et al. investigated the developmental genetics of mechanoreceptor cells (hair cells) in the cnidarian model N. vectensis. They used CRISPR-Cas9-mediated mutagenesis to showed that POU-IV homeodomain transcription factor regulates the differentiation of hair cells in this organism. The authors applied behavior assay, EM observations, and various types of fluorescence labeling to show that pou-iv -/- polyps exhibit defects in touch-sensitive behavior, likely due to the failure of forming the complete stereocilliary rootlet structure near the apical side of the hair cells in those mutant polyps. The authors went on to apply ChIP-seq in N. vectensis and showed that the POU-IV-binding motifs are conserved across Cnidaria and Bilateria. They also used this ChIP-seq dataset to screen for possible POU-IV downstream targets and identified one of the candidate genes, PKD1, as a conserved effector gene that has been shown playing important functions in hair cells across different bilaterian animals. Furthermore, by cross-checking their results with the newly published single-cell transcriptome data from N. vectensis adults, the authors identified the putative cell cluster (c79) of mechanosensory hair cells and confirmed that pou-iv and PKD1 are indeed co-expressed in this cell type. This approach also enabled the identification of additional candidate POU-IV downstream targets, and based on the GO term analysis, it appears that many of these genes are involved in ion transport and sensory perception functions. In summary, the authors provide strong evidence to support that POU-IV likely functions as a terminal selector factor of hair cell development in the sea anemone N. vectensis. Comparing their findings with other animals, the authors suggested that POU-IV factor plays a conserved role in regulating mechanoreceptor differentiation across Cnidarians and Bilaterians and that this regulatory mechanism may represent an ancestral trait dated back to their common ancestor.

    This is a detailed study on the role of POU-IV factor during cnidarian mechanoreceptor cell development. In general, the manuscript is well written, most of the data presented are of great quality, and the conclusions of the paper are supported by the data. This study is a significant advancement to our understanding on the evolutionary origin of sensory neurons and the possible genetic mechanisms underlying the diversification of neuronal cell types in animals.

    Strengths: The authors applied multiple approaches to examine the developmental process of hair cells in N. vectensis and analyze the molecular genetic functions of POU-IV factor during this process. The generation of gene-specific KO animals with CRISPR-Cas9 mediated mutagenesis in N. vectensis and the characterization of the sensory ability of those mutant animals with behavior assays provide compelling data to show that POU-IV factor is involved in the final maturation of mechanoreceptor hair cells. The ChIP-seq data generated by this study further enabled the authors to analyze the POU-IV factor binding sequences across animals, and the data also help to identify candidate downstream targets of POU-IV factor in N. vectensis system. Because POU-IV factor is likely involved in the development of multiple cell types in N. vectensis (as shown by previous publications and this study), this dataset would be highly valuable in the future for analyzing the differentiation process of different neuronal cell types in N. vectensis. In fact, by comparing with the recently available scRNA-seq resources, the authors have demonstrated the usefulness of this dataset and pointed out several interesting future research directions. Because N. vectensis is one of the few experimentally tractable systems within Cnidarian, which represents the sister group of bilatarian animals, experimental data from N. vectensis would give important mechanistic insights to infer the possible developmental characters in the common ancestor of Cnidarians and Bilaterians.

    We appreciate the reviewer’s excellent summary of our work.

    Weakness:

    1. The specificity of the POU-IV antibody staining. It appears that the signals of the POU-IV immune-staining are distributed quite extensively, especially near the basal part of the epithelia ectoderm (Figure 2A-L). And in their Western blot, the authors also noticed an extra band that might represent another protein in the N. vectensis sample that cross-reacted with their anti-POU-IV antibody. Although the authors provided controlled experiment showing that the immunostaining signals disappeared after they pre-absorbed the antibody with the POU-IV antigen (Figure 2 - supplement 2), this result can only demonstrate that indeed their antibody reacts specifically with this antigen. This cannot rule out the possibility that other N. vectensis protein(s) may possess peptide motifs similar to this antigen region and can be recognized by their antibody. Therefore, it would be nice if the authors can do double staining using in situ hybridization with pou-iv anti-sense riboprobe and immunostaining with their POU-IV antibody, to examine whether these two different methods would give overlapping results, so that they can be more confident about the specificity of their POU-IV antibody staining.

    We previously carried out the suggested double labeling experiments combining in situ hybridization with the pou-iv riboprobe and immunostaining with the anti-POU-IV. However, the signal to noise ratio of anti-POU-IV antibody staining decreased substantially when immunostaining was combined with in situ hybridization. We have therefore chosen, in this case, to compare the results of single labeling experiments for Pou-iv in situ and antibody staining side-by-side (Figure 2 – Figure supplement 1) to provide evidence that the pattern of POU-IV antibody staining in developing tentacles indeed recapitulates that of pou-iv mRNA expression.

    1. The electron microscopy data (Figure 5J-L) are not as clear as one would expect showing the differences in rootlet structure between the wildtype and mutant polyps, given that the phalloidin staining results (Figure 5B, E, H) show quite noticeable reduction in the mutant polyp tip. It is very hard to see the stereovillar rootlets (rlst) in Figure 5J, and thus it is very difficult to assess whether these structures are indeed affected in Figure 5K and 5L. In addition, the rootlet structure of the apical cilium in Figure 5K and L (presumably underneath "ci") appears to be less prominent compared to that shown in Figure 1I (labeled as "rlci"). I am not sure whether this is due to differences of the section angle, or whether it really reflects some differences between the wildtype and mutant.

    We had similar concerns with respect to the clarity of stereociliary rootlets of the F2 wildtype sibling animal shown in Figure 5J, and therefore included in the original submission additional electron microscopy evidence of stereociliary rootlets in hair cells of a wildtype sibling in Figure 5 – Figure supplement 2. Regarding the ciliary rootlet structure, we assume that the apparent lack of the structure in Figure 5K and L is due to the section angle not capturing the structure, as ciliary rootlets can be observed in other TEM sections of hair-cell-like cells in pou-iv mutant polyps. For clarification, we have added a supplementary figure (Figure 5 – Figure supplement 3) of a presumptive hair cell of a pou-iv mutant showing a ciliary rootlet.

  2. Version published to 10.7554/elife.74336 on eLife
  3. eLife

    Evaluation Summary:

    This manuscript focusses on a little studied, but highly interesting presumptive mechanosensory cell type in cnidarians, the 'hair cell'. The work shows that the POU-IV transcription factor is required for the maturation of this cell type in the sea anemone Nematostella vectensis. Because POU-IV transcription factors also play essential roles in the differentiation of mechanoreceptors in many bilaterian phyla, this suggests an evolutionarily ancient role of POU-IV in regulating mechanosensory identity. This study will hence be of great interest to developmental biologists and evolutionary biologists who are interested in the developmental evolution of neuronal cell types.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1, Reviewer #2 and Reviewer #3 agreed to share their name with the authors.)

  4. eLife

    Reviewer #1 (Public review)

    The paper demonstrates the role of Pou domains for various sensory cells. Using CRISPR to delete the gene, the authors show an incomplete deletion of sensory cells. Further evidence shows problems with the formation of mechanosensory cells.
    Overall, the presentation is clear but can be expanded by adding the role of bHLH genes (Atoh1 is upstream of Pou4f3). If possible, I suggest expanding the role of TMC as it is the main receptor in mammalian hair cells that connects to the stereocilia. Please note that the cnidarian organization is a central kinocilium surrounded by microvilli, comparable to choanoflagellates. This paper is a great original presentation but it could provide a broader perspective by expanding on the evolution of Pou IV and by adding a discussion of the evolution of bHLH, Myc and TMC in order to provide this broader perspective.

  5. eLife

    Reviewer #2 (Public Review):

    Whereas the role of POU-IV for the differentiation of cnidocytes and other neurons of Nematostella has been previously characterized (Tourniere et al., 2020), the present study extends previous reports by specifically addressing the role of POU-IV for the so-called "hair cells" of Nematostella (not to be confused with the hair cells of the vertebrate inner ear and lateral line). These presumably mechanosensory hair cells are identified here as postmitotic neurons, which are ciliated and carry a collar of stereovilli - actin-filled microvilli with a long actin-rich rootlet. Using CRISPR/Cas9 based gene editing, the study shows that transgenic animals, in which the POU-IV gene has been disrupted, become touch insensitive. While hair cells can still be identified in these POU-IV mutants, they lack the stereovillar rootlets suggesting that POU-IV is required for proper hair cells maturation, but dispensable for early steps of hair cell specification and differentiation. The study then uses ChIP-Seq to identify direct target genes of POU-IV in Nematostella and to characterize a POU-IV binding motif, which turned to be evolutionarly highly conserved with POU-IV binding motifs in bilaterians. Comparison of the ChIP-Seq data with published bulk and single-cell transcriptome data indicated that POU-IV activates substantially different sets of effector genes (but no regulatory genes) in hair cells and cnidocytes, and identified polycystin1 as a hair cell-specific direct target of POU-IV. Taken together, this suggests that POU-IV had an evolutionary ancient role as a terminal selector gene for mechanosensory neurons, which predated the split between cnidarian and bilaterian lineages but that its function diverged (e.g. by the acquisition of new target genes) during the evolution of cnidocytes as a novel cell type in cnidarians.

    Combining gene editing with sequencing and with careful morphological and behavioral characterisation of cellular phenotypes, the study provides valuable new insights into the evolution of sensory neurons. POU-IV class transcription factors have previously been implicated in the specification of mechano- and chemosensory neurons in bilaterians. The present study together with the previous study of Tourniere et al. (2020) now suggests an even deeper evolutionary origin of this cell type in the last common ancestor of eumetazoans. The paper is very well written and the results are beautifully documented. The authors are overall cautious and conservative in the conclusions drawn from their findings. However, two points deserve a more critical discussion, first, the question of which sensory modality is mediated by the hair cells (are these dedicated mechanoreceptors or possibly multimodal cells?), and second, the question whether POU-IV serves as transcriptional activator or repressor in cnidocytes.

  6. eLife

    Reviewer #3 (Public Review):

    In this manuscript, Ozmet et al. investigated the developmental genetics of mechanoreceptor cells (hair cells) in the cnidarian model N. vectensis. They used CRISPR-Cas9-mediated mutagenesis to showed that POU-IV homeodomain transcription factor regulates the differentiation of hair cells in this organism. The authors applied behavior assay, EM observations, and various types of fluorescence labeling to show that pou-iv -/- polyps exhibit defects in touch-sensitive behavior, likely due to the failure of forming the complete stereocilliary rootlet structure near the apical side of the hair cells in those mutant polyps. The authors went on to apply ChIP-seq in N. vectensis and showed that the POU-IV-binding motifs are conserved across Cnidaria and Bilateria. They also used this ChIP-seq dataset to screen for possible POU-IV downstream targets and identified one of the candidate genes, PKD1, as a conserved effector gene that has been shown playing important functions in hair cells across different bilaterian animals. Furthermore, by cross-checking their results with the newly published single-cell transcriptome data from N. vectensis adults, the authors identified the putative cell cluster (c79) of mechanosensory hair cells and confirmed that pou-iv and PKD1 are indeed co-expressed in this cell type. This approach also enabled the identification of additional candidate POU-IV downstream targets, and based on the GO term analysis, it appears that many of these genes are involved in ion transport and sensory perception functions. In summary, the authors provide strong evidence to support that POU-IV likely functions as a terminal selector factor of hair cell development in the sea anemone N. vectensis. Comparing their findings with other animals, the authors suggested that POU-IV factor plays a conserved role in regulating mechanoreceptor differentiation across Cnidarians and Bilaterians and that this regulatory mechanism may represent an ancestral trait dated back to their common ancestor.

    This is a detailed study on the role of POU-IV factor during cnidarian mechanoreceptor cell development. In general, the manuscript is well written, most of the data presented are of great quality, and the conclusions of the paper are supported by the data. This study is a significant advancement to our understanding on the evolutionary origin of sensory neurons and the possible genetic mechanisms underlying the diversification of neuronal cell types in animals.

    Strengths:
    The authors applied multiple approaches to examine the developmental process of hair cells in N. vectensis and analyze the molecular genetic functions of POU-IV factor during this process. The generation of gene-specific KO animals with CRISPR-Cas9 mediated mutagenesis in N. vectensis and the characterization of the sensory ability of those mutant animals with behavior assays provide compelling data to show that POU-IV factor is involved in the final maturation of mechanoreceptor hair cells. The ChIP-seq data generated by this study further enabled the authors to analyze the POU-IV factor binding sequences across animals, and the data also help to identify candidate downstream targets of POU-IV factor in N. vectensis system. Because POU-IV factor is likely involved in the development of multiple cell types in N. vectensis (as shown by previous publications and this study), this dataset would be highly valuable in the future for analyzing the differentiation process of different neuronal cell types in N. vectensis. In fact, by comparing with the recently available scRNA-seq resources, the authors have demonstrated the usefulness of this dataset and pointed out several interesting future research directions. Because N. vectensis is one of the few experimentally tractable systems within Cnidarian, which represents the sister group of bilatarian animals, experimental data from N. vectensis would give important mechanistic insights to infer the possible developmental characters in the common ancestor of Cnidarians and Bilaterians.

    Weakness:
    1. The specificity of the POU-IV antibody staining. It appears that the signals of the POU-IV immune-staining are distributed quite extensively, especially near the basal part of the epithelia ectoderm (Figure 2A-L). And in their Western blot, the authors also noticed an extra band that might represent another protein in the N. vectensis sample that cross-reacted with their anti-POU-IV antibody. Although the authors provided controlled experiment showing that the immunostaining signals disappeared after they pre-absorbed the antibody with the POU-IV antigen (Figure 2 - supplement 2), this result can only demonstrate that indeed their antibody reacts specifically with this antigen. This cannot rule out the possibility that other N. vectensis protein(s) may possess peptide motifs similar to this antigen region and can be recognized by their antibody. Therefore, it would be nice if the authors can do double staining using in situ hybridization with pou-iv anti-sense riboprobe and immunostaining with their POU-IV antibody, to examine whether these two different methods would give overlapping results, so that they can be more confident about the specificity of their POU-IV antibody staining.
    2. The electron microscopy data (Figure 5J-L) are not as clear as one would expect showing the differences in rootlet structure between the wildtype and mutant polyps, given that the phalloidin staining results (Figure 5B, E, H) show quite noticeable reduction in the mutant polyp tip. It is very hard to see the stereovillar rootlets (rlst) in Figure 5J, and thus it is very difficult to assess whether these structures are indeed affected in Figure 5K and 5L. In addition, the rootlet structure of the apical cilium in Figure 5K and L (presumably underneath "ci") appears to be less prominent compared to that shown in Figure 1I (labeled as "rlci"). I am not sure whether this is due to differences of the section angle, or whether it really reflects some differences between the wildtype and mutant.

  7. Version published to 10.1101/2021.10.12.464036v1 on bioRxiv