Par protein localization during the early development of Mnemiopsis leidyi suggests different modes of epithelial organization in the metazoa

In bilaterians and cnidarians, epithelial cell-polarity is regulated by the interactions between Par proteins, Wnt/PCP signaling pathway, and cell-cell adhesion. Par proteins are highly conserved across Metazoa, including ctenophores. But strikingly, ctenophore genomes lack components of the Wnt/PCP pathway and cell-cell adhesion complexes raising the question if ctenophore cells are polarized by mechanisms involving Par proteins. Here, by using immunohistochemistry and live-cell imaging of specific mRNAs, we describe for the first time the subcellular localization of selected Par proteins in blastomeres and epithelial cells during the embryogenesis of the ctenophore Mnemiopsis leidyi. We show that these proteins distribute differently compared to what has been described for other animals, even though they segregate in a host-specific fashion when expressed in cnidarian embryos. This differential localization might be related to the emergence of different junctional complexes during metazoan evolution.

Interestingly, ctenophores or comb jellies, whose position at the base of metazoan tree is still under debate (Dunn et al., 2008;Hejnol et al., 2009;Ryan et al., 2013;Moroz et al., 2014;Whelan et al., 2017), (Simion et al., 2017), (Feuda et al., 2017), possess a stereotyped development ( Figure 1B) and do not have the genes that encode the components of the Wnt/PCP pathway in their genomes (Ryan et al., 2013). Thus, the study of the subcellular organization of the Par system components in ctenophores is important to understand the evolution of tissue organization in Metazoa.
This mechanism is deployed in bilaterian cells to establish embryonic and epithelial cell polarity during early development and is critical for axial organization (Salinas-Saavedra et al., 2015;Cha et al., 2011;Munro, 2006;Patalano et al., 2006;Goldstein and Macara, 2007;Weisblat, 2007;Alford et al., 2009;Munro and Bowerman, 2009;Doerflinger et al., 2010;Chan and Nance, 2013;Lang and Munro, 2017;Tepass, 2012;Nance and Zallen, 2011;Weng and Wieschaus, 2017;Zhu et al., 2017;Ragkousi et al., 2017;Schneider and Bowerman, 2003;Macara, 2004;Vinot et al., 2004;Dollar et al., 2005;Ossipova et al., 2005). Components of the Par system are unique to, and highly conserved, across Metazoa, including placozoans, poriferans, and ctenophores (Fahey and Degnan, 2010;Belahbib et al., 2018). But strikingly, ctenophore genomes do not have many of the crucial regulators present in other metazoan genomes (Belahbib et al., 2018;Ganot et al., 2015). For example, none of the components of the Crb complex, a Scribble homolog, or Human and Drosophila SJs, are present (Belahbib et al., 2018;Ganot et al., 2015), and the cytoplasmic domain of cadherin lacks the crucial biding sites to catenins that interact with the actin cytoskeleton (Belahbib et al., 2018). These data raise the question of whether or not ctenophore cells are polarized by mechanisms involving the apicobasal cell polarity mediated by Par proteins. Here, by using antibodies raised to specific ctenophore proteins and confirmed by live-cell imaging of injected fluorescently labeled mRNAs, we describe for the first time the subcellular localization of selected components of the Par system during the development of the ctenophore Mnemiopsis leidyi. Data obtained here challenge the conservation of the apicobasal cell polarity module and raise questions about the epithelial tissue organization as an evolutionary trait of all metazoans.

Results
MlPar-6 gets localized to the apical cortex of cells during early M. leidyi development We characterized the subcellular localization of the MlPar-6 protein during early M. leidyi development by using our specific MlPar-6 antibody ( Figure 2 and Figure 2-figure supplements 1-6). Although MlPar-6 immunoreactivity can be detected in the periphery of the entire cell, in all of over 100 specimens examined, its expression appears to be polarized to the animal cortex (determined by the position of the zygotic nucleus; Figure 2A and  Figure 2D) undergoing epibolic movements, syncytial endoderm, and mesenchymal 'mesoderm' (quotation marks its debatable homology). However, this protein remains polarized in 'static' ectodermal cells remaining at the animal pole (blastopore) and vegetal pole (4-7 hpf; Figure 2-figure supplements 3F-J and 4). By the end of gastrulation (8-9 hpf; Figure 2E), MlPar-6 becomes localized asymmetrically to the apical cortex of the ectodermal epidermal cells and the future ectodermal pharyngeal cells that start folding inside the blastopore ( Figure 2E and Interestingly, we do not observe a clear cortical localization in later cydippid stages, and the antibody signal is weaker after 10 hpf in juveniles ( Figure 2F). Contrary to expectations, at these later stages, MlPar-6 is cytosolic and does not localize in the cortex of epidermal cells, and a few epithelial and mesenchymal cells showed nuclear localization ( Figure 2F). Thereafter, MlPar-6 remains cytosolic in all scored stages up to 24 hpf (Figure 2-figure supplement 6). Cytosolic and nuclear localization of Par-6 has been reported in other organisms when the polarizing roles of this protein are inactive (Mizuno et al., 2003;Johansson et al., 2000;Cline and Nelson, 2007). Thus, our data suggest that MlPar-6 does not play a role in cell polarity during juvenile cydippid stages. These patterns of apical localization seem not to be affected by the cell cycle (

MlPar-1 remains cytoplasmic during early M. leidyi development
In bilaterians and cnidarians, the apical localization of MlPar-6 induces the phosphorylation of MlPar-1, displacing this protein to basolateral cortical regions (Ohno et al., 2015;Salinas-Saavedra et al., 2015;Ragkousi et al., 2017;. Using our specific MlPar-1 antibody, we characterized the subcellular localization of the MlPar-1 protein during the early M. leidyi development ( Figure 3 and all its supplements). Even though MlPar-1 appears to be localized in the cortex at the cell-contact regions of early blastomeres and gastrula stages ( Figure 3D-E), this antibody signal was not clear enough to be discriminated from the cytosolic distribution, possibly due to edge effects. Nevertheless, and strikingly, MlPar-1 remains as punctate aggregations distributed uniformly in the cytosol, and in some cases, co-distributes with chromosomes during mitosis ( Figure 3 and Figure 3-figure supplement 2). We did not observe asymmetric localization of MlPar-1 in the cell cortex of M. leidyi embryos at any of the stages described above for MlPar-6. localizes to the apical cortex of the ectodermal cells (Ecto) but is absent from endodermal (Endo) and 'mesodermal' ('Meso') cells. White arrowhead indicates MlPar-6 protein in regions of cell-contact. Yellow arrowheads indicate the absence of cortical localization. (E) Until 9 hpf, MlPar-6 protein localizes to the apical cortex of the ectoderm (white arrows) and pharynx (white arrowhead) but it is not cortically localized after 10 hpf (F; Yellow arrowheads indicate nuclear localization). Images are maximum projections from a z-stack confocal series. The 8 cell stage corresponds to a single optical section. Orientation axes are depicted in the Figure                These results were also supported in vivo when we overexpressed the mRNA encoding for MlPar-1 fused to mCherry (MlPar-1-mCherry) into M. leidyi embryos by microinjection (Figure 3-figure supplement 3). Similar to MlPar-6-mVenus mRNA overexpression, the MlPar-1-mCherry translated protein was observed after 4 hr post injection into the uncleaved egg. Our in vivo observations on living embryos confirm the localization pattern described above by using MlPar-1 antibody at gastrula stages. MlPar-1-mCherry localizes uniformly and form aggregates in the cytosol during gastrulation (4-5 hpf; Figure 3-figure supplement 3D-E and Video 1). This localization pattern remains throughout all recorded stages until cydippid juvenile stages where MlPar-1-mCherry remains cytosolic in all cells but is highly concentrated in the tentacle apparatus and underneath the endodermal canals (24 hpf; Figure 3- MlPar-6 and MlPar-1 Proteins can localize like host proteins localize in a heterologous system To discount the possibility that the observations recorded in vivo for both MlPar-6-mVenus and MlPar-1-mCherry proteins are caused by a low-quality mRNA or lack of structural conservation, we overexpressed each ctenophore mRNA into embryos of the cnidarian Nematostella vectensis and followed their localization by in vivo imaging (Figure 4). In N. vectensis embryos, MlPar-6-mVenus and MlPar-1-mCherry symmetrically distribute during early cleavage stages ( Figure 4A and C) and both proteins localize asymmetrically only after blastula formation ( Figure 4B and D). In these experiments, both MlPar-6-mVenus and MlPar-1-mCherry translated proteins display the same pattern as the previously described endogenous N.   vectensis Par-6 and Par-1 proteins (Salinas-Saavedra et al., 2015). These data suggest that the protein structure of ctenophore MlPar-6 and MlPar-1 contains the necessary information to localize as other bilaterians proteins do.

Discussion
Par protein asymmetry is established early but not maintained during M. leidyi embryogenesis The asymmetric localization of the Par/aPKC complex has been used as an indicator of apical-basal cell polarity in a set of animals, including bilaterians ( ) and embryonic polarity is controlled by the Wnt signaling system (Kumburegama et al., 2011;Wikramanayake et al., 2003;Lee et al., 2007;Martindale and Hejnol, 2009;Martindale and Lee, 2013). In spite of these differences, once epithelial tissues form and epithelial cell-polarity is established in both bilaterian and cnidarian species, the asymmetric localization of Par proteins become highly polarized and is maintained through development. In those cases, Par-mediated apicobasal cell polarity is responsible for the maturation and maintenance of cell-cell adhesion in epithelial tissue (Ohno et al., 2015;. We have suggested that the polarizing activity of the Par system was already present in epithelial cells of the MRCA between Bilateria and Cnidaria (Salinas-Saavedra and Martindale, 2018; Salinas-Saavedra and Martindale, 2018) and could be extended to all Metazoa, where these proteins are present (including ctenophores, sponges, and placozoans Fahey and Degnan, 2010; Belahbib et al., 2018). However, our current data suggest a different scenario for ctenophores where the Par protein polarization observed during earlier stages (characterized by the apical and cortical localization of MlPar-6; Figure 2) is not maintained when ctenophore juvenile epithelial tissues form after nine hpf. Epithelial cells of later cydippid stages do not display an asymmetric localization of MlPar-6 ( Figure 2-figure supplement 6). Furthermore, the subcellular localization of MlPar-1 does not display a clear localization during any of the observed developmental stages (Figure 3 and all its supplements). Instead, punctate aggregates distribute symmetrically in the cytosol. MlPar-1 and mCherry aggregates may be consequence of the highly protein availability in the cytosol that is not captured to the cell cortex.
The components of the ctenophore MlPar/aPKC complex (MlPar-3/MlaPKC/MlPar-6 and MlCdc42) are highly conserved and contain all the domains present in other metazoans (Figure 1figure supplements 1-2; Fahey and Degnan, 2010;Belahbib et al., 2018). Similarly, the primary structure of MlPar-1 protein (a Serine/threonine-protein kinase) is highly conserved and contains all the domains (with the same amino acid length) required for its proper functioning in other metazoans (Figure 1-figure supplement 3; Fahey and Degnan, 2010;Belahbib et al., 2018), and localizes to the lateral cortex when expressed in cnidarian embryos (Figure 4). Regardless, these proteins do not asymmetrically localize to the cortex of M. leidyi juvenile epithelium. Interestingly, the punctuate aggregates of MlPar-1-mCherry are highly dynamic and move throughout the entire cytosol ( Figure 3-figure supplement 3), suggesting a potential association with cytoskeletal components (see Video 1) as MlPar-1 conserve these motifs.
Recent studies have shown that ctenophores do not have homologs for any of the Crb complex components (Belahbib et al., 2018), required for the proper stabilization of the CCC and Par/aPKC complex in other studied taxa (Ohno et al., 2015;Harris and Peifer, 2004;Tepass, 2012;Chalmers et al., 2005;Hayase et al., 2013;Whitney et al., 2016). The lack of MlPar-6 ( Figure 2) polarization during later stages is totally congruent with these observations, indicating that Par proteins in ctenophores do not have the necessary interactions to stabilize apico-basal cell polarity in their cells as in other animals. In addition, ctenophore species do not have the molecular components to form SJs and lack a Scribble homolog (Belahbib et al., 2018;Ganot et al., 2015). This could explain the cytosolic localization of MlPar-1 during the observed stages (Benton and St Johnston, 2003;Iden and Collard, 2008;Humbert et al., 2015;Bilder et al., 2000;Vaccari et al., 2005), (Bonello et al., 2019).

Evolution of cell polarity and epithelial structure in metazoa
Given the genomic conservation of cell-polarity components in the Bilateria and Cnidaria, we propose to classify their epithelium as 'Par-dependent' to include its mechanistic regulatory properties. That is, the structural properties of a 'Par-dependent'epithelium are the result of conserved interactions between subcellular pathways that polarize epithelial cells. Thus, when we seek to understand the origins of the epithelial nature of one particular tissue, we are trying to understand the synapomorphies (shared derived characters) of the mechanisms underlying the origin of that particular tissue. Under this definition, a 'Par-dependent epithelium' may have a single origin in Metazoa, but, different mechanisms might have co-opted to generate similar epithelial morphologies (Figure 4figure supplement 1). Ctenophore epithelia, along with other recent works in N. vectensis endomesoderm (Salinas-Saavedra et al., 2015; and Drosophila midgut (Chen et al., 2018), suggest this possibility. In all these cases, epithelial cells are highly polarized    Salinas-Saavedra and Martindale. eLife 2020;9:e54927. DOI: https://doi.org/10.7554/eLife.54927 along the apical-basal axis, but this polarization does not depend on Par proteins. Therefore, these cells are not able to organize a 'Par-dependent epithelium' (mechanistic definition) but still polarized epithelial morphologies.
Genomic studies also suggest that ctenophore species lack the molecular interactions necessaries to form the apical cell polarity and junctions observed in Cnidaria + Bilateria. Intriguingly, ctenophore genomes do not have the Wnt signaling pathway components (Ryan et al., 2013;Moroz et al., 2014;Pang et al., 2010) that control the activity of Par proteins in bilaterian and cnidarian embryos (components that are also present in poriferan and placozoan genomes Belahbib et al., 2018). For example, in bilaterians the Wnt/PCP signaling pathway antagonizes the action of the Par/aPKC complex (Cha et al., 2011;Besson et al., 2015;Aigouy and Le Bivic, 2016;Humbert et al., 2015;Humbert et al., 2006;Seifert and Mlodzik, 2007), so this may explain the lack of polarization in ctenophore tissue. Furthermore, ctenophore species do not have the full set of cell-cell adhesion proteins (Belahbib et al., 2018;Ryan et al., 2013;Ganot et al., 2015) as we know them in other metazoans, including Placozoans and Poriferans (Magie and Martindale, 2008;Belahbib et al., 2018). The cadherin of ctenophores does not have the cytoplasmic domains required to bind any of the catenins of the CCC (e.g. p120, alpha-and ß-catenin) (Belahbib et al., 2018). This implies that neither the actin nor microtubule cytoskeleton can be linked to ctenophore cadherin through the CCC, as seen essential in other metazoans to stabilize pre-existent Par proteins polarity. This suggests that there are additional mechanisms that integrate the cytoskeleton of ctenophore cells with their cell-cell adhesion system.
In conclusion, regardless the phylogenetic position of the Ctenophora, the conservation of an organized 'Par-dependent epithelium' cannot be extended to all Eumetazoa. Ctenophore cells do not have other essential components to organize the polarizing function of the Par system as in other studied metazoans. Despite the high structural conservation of Par proteins across Metazoa, we have shown that ctenophore cells do not deploy and/or stabilize the asymmetrical localization of Par-6 and Par-1 proteins. Thus, ctenophore tissues organize their epithelium in a different way than the classical definition seen in bilaterians. In agreement with genomic studies, our results question what molecular properties defined the ancestral roots of a metazoan epithelium, and whether similar epithelial morphologies (e.g., epidermis and mesoderm) could be developed by independent or modifications of existing cellular and molecular interactions (including cell adhesion systems). Unless the lack of Par protein localization in M. leidyi is a secondary loss, the absence of these pathways in ctenophores implies that a new set of interactions emerged at least in the Cnidaria+Bilateria ancestor ( Figure 4-figure supplement 1), and that, could have regulated the way by which the Par system polarizes embryonic and epithelial cells. While bioinformatic studies are critical to understand the molecular composition, we need further research to understand how these molecules actually interact with one another to organize cellular behavior (e.g., integrin-collagen, basal-apical interactions) in a broader phylogenetical sample, including Porifera and Placozoa.

Western blot
Western blots were carried out as described (Salinas-Saavedra et al., 2015; using adult epithelial tissue lysates dissected by hand in order to discard larger amount of mesoglea. Antibody concentrations for Western blot were 1:1000 for all antibodies tested.

Immunohistochemistry
All immunohistochemistry experiments were carried out using the previous protocol for M. leidyi (Salinas-Saavedra and . The primary antibodies and concentrations used were: mouse anti-alpha tubulin (1:500; Sigma-Aldrich, Inc Cat.# T9026. RRID:AB_477593). Secondary antibodies are listed in the Key Resources table. Rabbit anti-MlPar-6, and rabbit anti-MlPar-1 antibodies were custom made high affinity-purified peptide antibodies that commercially generated by Bethyl labs, Inc (Montgomery, TX, USA). Affinity-purified M. leidyi anti-Par-6 (anti-MlPar-6) and anti-Par-1 (anti-MlPar-1) peptide antibodies were raised against a selected amino acid region of the MlPar-6 protein (MTYPDDSNGGSGR) and MlPar-1 protein (KDIAVNIANELRL), respectively. Blast searches against the M. leidyi genome sequences showed that the amino acid sequences were not present in any predicted M. leidyi proteins other than the expected protein. Both antibodies are specific to M. leidyi proteins (Figure 2-figure supplement 2) and were diluted 1:100.

mRNA microinjections
The coding region for each gene of interest was PCR-amplified using cDNA from M. leidyi embryos and cloned into pSPE3-mVenus or pSPE3-mCherry using the Gateway system (Roure et al., 2007).
To confirm the presence of the transcripts during M. leidyi development, we cloned each gene at 2 hpf and 48 hpf. N. vectensis eggs were injected directly after fertilization as previously described (Salinas-Saavedra et al., 2015;DuBuc et al., 2014;Layden et al., 2013) with the mRNA encoding one or more proteins fused in frame with reporter fluorescent protein (N-terminal tag) using an optimized final concentration of 300 ng/ml for each gene. Fluorescent dextran was also co-injected to visualize the embryos. Live embryos were kept at room temperature and visualized after the mRNA of the FP was translated into protein (4-5 hr). Live embryos were mounted in 1x sea water for visualization. Images were documented at different stages. We injected and recorded at least 20 embryos for each injected protein and confocal imaged each specimen at different stages for detailed analysis of phenotypes in vivo. We repeated each experiment at least five times obtaining similar results for each case. The fluorescent dextran and primers for the cloned genes are listed in Key resources table.

Imaging of M. leidyi embryos
Images of live and fixed embryos were taken using a confocal Zeiss LSM 710 microscope using a Zeiss C-Apochromat 40x water immersion objective (N.A. 1.20). Pinhole settings varied between 1.2-1.4 A.U. according to the experiment. The same settings were used for each individual experiment to compare control and experimental conditions. Z-stack images were processed using Imaris 7.6.4 (Bitplane Inc) software for three-dimensional reconstructions and FIJI for single slice and videos. Final figures were assembled using Adobe Illustrator and Adobe Photoshop. Par proteins display a general cytosolic localization when their polarizing activity is inactive. This signal was diminished by modifying contrast and brightness of the images in order to enlighten their cortical localization (active state in cell-polarity and stronger antibody signal) as it has shown in other organisms. All RAW images are available upon request.

Fluorescent intensity measurements and statistical analyses
Images of fixed embryos were measured using FIJI plot profile tool using the RAW source data. Fluorescent intensity was measured along the animal-vegetal axis for 1 and 2 cell stages and along the apico-basal axis for the other later stages. The data obtained were then normalized by the maximum value of each X and Y axes. X axis corresponds to the distance from basal (0) to apical (1) cortex. Y axis corresponds to fluorescence intensity. The normalized data were plotted and the numerical values can be found in figure supplement-data source files. For later stages than 8 cells, we took measurements of two cells located in perpendicular axes of the embryo where the apicobasal axis was clearly detectable. These measurements correspond to cells going through interphase and metaphase. Statistical analyses were executed using GraphPad prism software. To do this, we compared the 10% most basal positions with the 10% most apical positions for each stage. We plotted this data and differences were assessed by comparing medians using Mann-Whitney U test.
Similarly, fluorescent intensity during cell cycle (Figure 2-figure supplement 11) was measured along the apical cortex. The data obtained were then normalized by the maximum value of each X and Y axes. X axis corresponds to the arbitrary distance (0 to 1) along the apical cortex where the middle point corresponds to the cell-cell contact region or cleavage furrow. Y axis corresponds to fluorescence intensity. The normalized data were plotted and the numerical values can be found in Figure 2-figure supplement 11-source data 1.