ABSTRACT
The Hippo pathway plays a crucial role in cell proliferation and differentiation during tumorigenesis, tissue homeostasis and early embryogenesis. Scaffold proteins from the ezrin-radixin-moesin (ERM) family, including neurofibromin 2 (NF2; Merlin), regulate the Hippo pathway through cell polarity. However, the mechanisms underlying Hippo pathway regulation via cell polarity in establishing outer cells remain unclear. In this study, we generated artificial Nf2 mutants in the N-terminal FERM domain (L64P) and examined Hippo pathway activity by assessing the subcellular localization of YAP1 in early embryos expressing these mutant mRNAs. The L64P-Nf2 mutant inhibited NF2 localization around the cell membrane, resulting in YAP1 cytoplasmic translocation in the polar cells. L64P-Nf2 expression also disrupted the apical centralization of both large tumor suppressor 2 (LATS2) and ezrin in the polar cells. Furthermore, Lats2 mutants in the FERM binding domain (L83K) inhibited YAP1 nuclear translocation. These findings demonstrate that NF2 subcellular localization mediates cell polarity establishment involving ezrin centralization. This study provides previously unreported insights into how the orchestration of the cell-surface components, including NF2, LATS2 and ezrin, modulates the Hippo pathway during cell polarization.
INTRODUCTION
Neurofibromin 2 (NF2) (also known as Merlin) is a member of the ezrin, radixin and moesin (ERM) protein family. Targeted deletion of NF2 results in tumorigenesis (Gehlhausen et al., 2015; Giovannini et al., 2000; McClatchey et al., 1998). NF2 and ERM proteins serve as scaffolds that assemble protein complexes at the cell cortex, situated between the cell membrane and cytoskeleton (McClatchey and Fehon, 2009). Among these, NF2 activates the growth inhibitory kinase cascade and the Hippo pathway, leading to the phosphorylation of Yes-associated protein 1 (YAP1), which controls cell proliferation and differentiation (Hamaratoglu et al., 2006; Yin et al., 2013). Thus, NF2 functions as a tumor suppressor and a potent regulator of diverse cellular processes. The Hippo pathway is crucial for the initial cell segregation into the inner cell mass (ICM) and trophectoderm (TE) during preimplantation development (Nishioka et al., 2009). The ICM and TE give rise to the embryo and placenta, respectively; therefore, these cell lineages form essential components of organogenesis.
ICM/TE cell specification is achieved by the differential expression patterns of cell lineage-specific transcription factors (Rossant, 2018). Specifically, the TE-specific transcription factor CDX2 determines cell fate from an undifferentiated state to TE (Strumpf et al., 2005). CDX2 expression is facilitated by another transcription factor, TEAD4 (Yagi et al., 2007). TEAD4 interacts with the transcriptional activator YAP1 to initiate CDX2 expression (Nishioka et al., 2009). This interaction is primarily regulated by YAP1 phosphorylation via the Hippo pathway kinases LATS1 and LATS2 (Lorthongpanich et al., 2013). When YAP1 is not phosphorylated by LATS1 or LATS2, it translocates to the nucleus, where it interacts with TEAD4 and initiates Cdx2 transcription. Thus, Hippo pathway inactivation is essential for achieving a transcriptional profile similar to a TE cell lineage in the outer cells of an embryo (Nishioka et al., 2009). In contrast, the Hippo pathway is activated by the phosphorylation by LATS1 or LATS2 in the inner cells, leading to the formation of ICM cell lineages.
Phosphorylated YAP1 does not translocate into the nucleus, and TEAD4-dependent Cdx2 expression is inhibited, resulting in the maintenance of pluripotency in the inner cells. LATS1- and LATS2-dependent YAP1 phosphorylation in inner cells requires NF2, making it essential for forming a fully functional ICM through Hippo pathway activation (Cockburn et al., 2013). However, the exact molecular mechanisms through which NF2 serves as an upstream component of the Hippo pathway during preimplantation development remain unknown. Thus, exploring the relationship between NF2 and cell polarization is needed to understand how NF2 regulates the Hippo pathway. Cell polarization is a major upstream factor for the Hippo pathway because prospective TE cells need to establish apical-basal polarity to acquire the epithelial characteristics of TE. Apical domain formation requires the organization of both the cytoskeleton and ERM proteins, with TFAP2C and TEAD4 inducing actin regulators from the eight-cell stage in mouse embryos (Zhu et al., 2020). Therefore, the eight-cell stage in mouse embryos is a key period for establishment of polarization. There are no inner cells at this stage, and all blastomeres have non-cell adhesion surfaces. Cell-cell junctions develop after this stage, followed by the close packing of the blastomeres composing an embryo, a phenomenon called compaction. Inner non-polarized cells appear following cell division after compaction. Depending on whether a blastomere becomes part of the ICM or TE, functional differences become more defined and fixed by the blastocyst stage. However, the localization of NF2 from the eight-cell stage to the blastocyst stage warrants further study.
The primary structure of NF2 might elucidate its function during preimplantation development, as it significantly affects the subcellular localization of this protein in cultured somatic cells (Li et al., 2010; Yin et al., 2013). Moreover, the role of NF2 as a tumor suppressor depends on its four-point-one/ezrin/radixin/moesin (FERM) domain (L64 of NF2) and the phosphorylation of S518 on the C-terminal. Point mutations at these amino acid residues can alter Hippo pathway activity (Li et al., 2010, 2014; Yin et al., 2013). However, the molecular role of NF2 in early embryonic development remains unclear.
This study aimed to investigate the effects of NF2 with point-mutated RNA on its subcellular localization after the eight-cell stage to explore the significance of NF2 subcellular localization in Hippo pathway regulation. Additionally, this study aimed to examine the effect of LATS2 subcellular localization on Hippo pathway activity in the outer cells and evaluate the interaction between LATS2 and ezrin through the FERM-binding domain (FBD) of LATS2.
RESULTS
The primary structure of NF2 determines its subcellular localization during preimplantation development
NF2 subcellular localization during preimplantation development was investigated through live-cell imaging (Fig. 1A, Movies 1 and 2). Wild-type Nf2 (Nf2WT)-Gfp mRNAs were microinjected into one of the blastomeres of two-cell stage embryos and visualized until the morula stage. Microinjection of Nf2WT-Gfp mRNA did not hinder embryonic development up to the morula stage. Furthermore, NF2 was uniformly expressed around the cell membrane until the early eight-cell stage. After the onset of compaction in the late eight-cell stage, NF2 rapidly decreased around the apical domain and remained at the cell adhesion sites (Fig. 1A). NF2 was gradually excluded from the center of the apical domain in the outer cells (Fig. 1B,C; Movies 1 and 2). Immunofluorescence of NF2 proteins at the morula stage showed a similar localization pattern to the GFP signals of Nf2WT-Gfp mRNA microinjected embryos (Fig. S1). These results demonstrate that NF2 protein localization switches from around the cell membrane to the adhesion sites after the onset of compaction in the late eight-cell stage.
Subcellular localization of NF2 until the morula stage. Images obtained through live-cell imaging (Movies 1 and 2). Five embryos were analyzed. (A) Representative images of the four-cell, early eight-cell, late eight-cell and morula stages are shown, starting from the left. Green indicates NF2WT-GFP. (B,C) Enlargements of the areas outlined in A at the early eight-cell and late eight-cell stages. NF2WT-GFP was gradually excluded from the center of the apical domain, as indicated by the yellow (present) and white (absent) arrowheads. (D) Four types of NF2 proteins fused with GFP: wild type (WT), substitution of leucine with proline at amino acid position 64 within the FERM domain (L64P) and substitutions of serine with alanine (S518A) or glutamic acid (S518E) at amino acid position 518. Confocal microscopy at the two-cell (2C; n=3), 4C (n=3) and morula (n=10) stages after microinjection with any of the Nf2WT-Gfp mRNAs into one of the blastomeres of the two-cell stage embryos (yellow, DNA; cyan, NF2). NF2L64P was uniformly localized throughout the cytoplasm, differing from other types of NF2. Scale bars: 25 μm.
Subcellular localization of NF2 until the morula stage. Images obtained through live-cell imaging (Movies 1 and 2). Five embryos were analyzed. (A) Representative images of the four-cell, early eight-cell, late eight-cell and morula stages are shown, starting from the left. Green indicates NF2WT-GFP. (B,C) Enlargements of the areas outlined in A at the early eight-cell and late eight-cell stages. NF2WT-GFP was gradually excluded from the center of the apical domain, as indicated by the yellow (present) and white (absent) arrowheads. (D) Four types of NF2 proteins fused with GFP: wild type (WT), substitution of leucine with proline at amino acid position 64 within the FERM domain (L64P) and substitutions of serine with alanine (S518A) or glutamic acid (S518E) at amino acid position 518. Confocal microscopy at the two-cell (2C; n=3), 4C (n=3) and morula (n=10) stages after microinjection with any of the Nf2WT-Gfp mRNAs into one of the blastomeres of the two-cell stage embryos (yellow, DNA; cyan, NF2). NF2L64P was uniformly localized throughout the cytoplasm, differing from other types of NF2. Scale bars: 25 μm.
N-terminal FERM domain (L64P) and C-terminal phosphorylation site (S518A and S518E) mutations were further prepared to explore the primary structure of NF2 and its subcellular localization in mouse embryos. The substitution of leucine with proline at amino acid position 64 in the FERM domain of NF2 weakens its function as a tumor suppressor, as demonstrated in a previous study (Li et al., 2010). In the present study, NF2L64P in mouse embryos was uniformly localized throughout the cytoplasm from the two-cell to the morula stage, which was distinct from the localization pattern of NF2WT (Fig. 1E). However, the phosphorylation state of serine at amino acid position 518 alters the three-dimensional conformation of NF2 in cultured cells (Hong et al., 2020; Li et al., 2010, 2015; Yin et al., 2013). Both NF2S518A and NF2S518E, which mimic the dephosphorylation and phosphorylation states, respectively, showed similar localization patterns to those of NF2WT (Fig. 1E). Therefore, NF2L64P in mouse embryos was compared with NF2WT to explore the relationship between NF2 subcellular localization and Hippo pathway activity in subsequent experiments.
Effects of the primary structure of NF2 on YAP1 nuclear localization in preimplantation embryos
The Hippo pathway is inactive in the outer cells of morula-stage embryos. The outer cells of embryos injected with Nf2WT-Gfp mRNAs showed YAP1 nuclear localization (Fig. 2A). Compared with the YAP1 localization pattern in Nf2WT-expressing cells, the difference in Hippo pathway activity was unclear between the inner and outer cells expressing NF2L64P. Some inner cells expressing NF2L64P showed YAP1 nuclear localization, whereas YAP1 fluorescence signals in the nuclei of some outer cells were decreased (Fig. 2A). In embryos injected with Nf2L64P-Gfp mRNAs, the YAP1 N/C ratio in outer cells expressing NF2L64P decreased when compared with that of outer cells without NF2L64P expression (Fig. 2B).
Subcellular localization of YAP1 in embryos microinjected with Nf2WT-Gfp or Nf2L64P-Gfp mRNA. (A) Either Nf2WT (left; n=10) or Nf2L64P-Gfp (right; n=10) mRNA was microinjected into one of the blastomeres of two-cell stage embryos. Microinjected embryos were used for YAP1 immunostaining at the morula stage. YAP1 nuclear localization in inner cells expressing NF2L64P (white arrowheads). Decreased YAP1 fluorescence in the nuclei of the outer cells (yellow arrowheads). Scale bar: 25 μm. (B) Quantification of the YAP1 N/C ratio in the outer cells with (n=14 cells) or without (n=18 cells) NF2L64P expression using three embryos microinjected with Nf2L64P-Gfp mRNA.
Subcellular localization of YAP1 in embryos microinjected with Nf2WT-Gfp or Nf2L64P-Gfp mRNA. (A) Either Nf2WT (left; n=10) or Nf2L64P-Gfp (right; n=10) mRNA was microinjected into one of the blastomeres of two-cell stage embryos. Microinjected embryos were used for YAP1 immunostaining at the morula stage. YAP1 nuclear localization in inner cells expressing NF2L64P (white arrowheads). Decreased YAP1 fluorescence in the nuclei of the outer cells (yellow arrowheads). Scale bar: 25 μm. (B) Quantification of the YAP1 N/C ratio in the outer cells with (n=14 cells) or without (n=18 cells) NF2L64P expression using three embryos microinjected with Nf2L64P-Gfp mRNA.
To investigate the cause of YAP1 mislocalization in NF2L64P-expressing cells, we analyzed the localization of angiomotin (Amot), as the NF2-AMOT interaction is crucial for Hippo pathway activation and subsequent YAP1 cytoplasmic localization (Hirate et al., 2013). In cells not expressing NF2L64P, Amot was localized at both the apical domains and adherens junction sites (Fig. 3). Conversely, the fluorescent signals indicating AMOT localization at both the apical domains and adherens junction sites were notably reduced in cells expressing NF2L64P (Fig. 3). These findings suggest that the observed YAP1 mislocalization in outer cells expressing NF2L64P may be due to impaired recruitment of AMOT to both apical domains and adherens junction sites.
Disrupted Amot localization in Nf2L64P-expressing cells. Either Nf2WT (left; n=12) or Nf2L64P-Gfp (right; n=5) mRNA was microinjected into one of the blastomeres of two-cell stage embryos. Microinjected embryos were then used for Amot immunostaining at the morula stage. In the embryo possessing Nf2L64P-Gfp-expressing cells, non-microinjected cells are outlined with a pink square and Nf2L64P-Gfp-expressing cells are outlined with a green square. Each outlined area is enlarged on the right. White and yellow arrowheads indicate the apical and adherence junction sites, respectively. Amot localization in non-microinjected cells was similar to that in the controls but differed from NF2L64P-GFP-expressing cells (green square), in which the Amot fluorescence signal completely disappeared. Scale bar: 25 μm.
Disrupted Amot localization in Nf2L64P-expressing cells. Either Nf2WT (left; n=12) or Nf2L64P-Gfp (right; n=5) mRNA was microinjected into one of the blastomeres of two-cell stage embryos. Microinjected embryos were then used for Amot immunostaining at the morula stage. In the embryo possessing Nf2L64P-Gfp-expressing cells, non-microinjected cells are outlined with a pink square and Nf2L64P-Gfp-expressing cells are outlined with a green square. Each outlined area is enlarged on the right. White and yellow arrowheads indicate the apical and adherence junction sites, respectively. Amot localization in non-microinjected cells was similar to that in the controls but differed from NF2L64P-GFP-expressing cells (green square), in which the Amot fluorescence signal completely disappeared. Scale bar: 25 μm.
Significance and regulation of Lats2 recruitment to the apical domain in Hippo pathway activity
LATS2 directly phosphorylates YAP1, which may be associated with YAP1 mislocalization caused by NF2L64P expression. Microinjection of Lats2WT-mCherry and NF2WT-Gfp mRNAs into the one-cell stage embryos revealed that LATS2 is localized to both the adherens junction and apical domains (Fig. 4A). More intense fluorescence signals for LATS2 were observed in the apical domains (Fig. 4A). Notably, NF2 and LATS2 fluorescence signals were alternately localized in the apical domain of NF2WT-GFP-expressing cells. Additionally, LATS2 recruitment to the apical domains was correlated with YAP1 nuclear localization (Fig. S2A). Furthermore, 72.4% of LATS2WT-expressing cells with LATS2WT localized to the apical domain were concomitant with CDX2 expression (Fig. S2A,B). These results suggest that apical localization of LATS2 is required for Hippo pathway inactivation. Moreover, YAP1 nuclear localization, CDX2 expression and LATS2 recruitment to the apical domains are correlated with NF2 elimination from the apical domain.
Regulation of Lats2 recruitment and Ezrin localization to the apical domains by Nf2 in early embryos. (A) The Nf2WT-Gfp/Lats2-mCherry mRNA (left; n=9) or Nf2L64P-Gfp/Lats2-mCherry mRNA (right; n=12) was microinjected into one-cell stage embryos. NF2 and LATS2 fluorescence signals were alternately localized in the apical domain of Nf2WT-Gfp-expressing cells. In addition, LATS2 localization to the apical domains was extremely prohibited in the Nf2L64P-Gfp-expressing cells. (B) The effect of NF2 localization on LATS2 localization in the apical domains. Three types of LATS2 localization patterns (type I, absence in the apical domains; type II, broad and thin localization in the apical domains; type III, narrow and thick localization in the apical domains). The percentage of cells with each localization type in Nf2WT-Gfp/Lats2-mCherry (n=11; 87 cells) or Nf2L64P-Gfp/Lats2-mCherry mRNA (n=11; 101 cells) embryos. (C) The Nf2WT-Gfp/Ezrin-mCherry mRNA (left; n=4) or Nf2L64P-Gfp/Ezrin-mCherry mRNA (right; n=5) was microinjected into one-cell stage embryos. The areas shown in more detail at the bottom of the panel are outlined in the merged images above. NF2 and Ezrin proteins in Nf2WT-Gfp/Ezrin-mCherry embryo (left) were localized to the apical domains in a mutually exclusive manner. Meanwhile, Ezrin localization in the Nf2L64P-Gfp/Ezrin-mCherry embryo (right) was thinly localized across the apical domain. Plot profiles of the white dashed lines in the enlarged images are shown in embryos microinjected with the Nf2WT-Gfp/Ezrin-mCherry (left) or Nf2L64P-Gfp/Ezrin-mCherry (right) mRNAs, respectively. Scale bars: 30 μm.
Regulation of Lats2 recruitment and Ezrin localization to the apical domains by Nf2 in early embryos. (A) The Nf2WT-Gfp/Lats2-mCherry mRNA (left; n=9) or Nf2L64P-Gfp/Lats2-mCherry mRNA (right; n=12) was microinjected into one-cell stage embryos. NF2 and LATS2 fluorescence signals were alternately localized in the apical domain of Nf2WT-Gfp-expressing cells. In addition, LATS2 localization to the apical domains was extremely prohibited in the Nf2L64P-Gfp-expressing cells. (B) The effect of NF2 localization on LATS2 localization in the apical domains. Three types of LATS2 localization patterns (type I, absence in the apical domains; type II, broad and thin localization in the apical domains; type III, narrow and thick localization in the apical domains). The percentage of cells with each localization type in Nf2WT-Gfp/Lats2-mCherry (n=11; 87 cells) or Nf2L64P-Gfp/Lats2-mCherry mRNA (n=11; 101 cells) embryos. (C) The Nf2WT-Gfp/Ezrin-mCherry mRNA (left; n=4) or Nf2L64P-Gfp/Ezrin-mCherry mRNA (right; n=5) was microinjected into one-cell stage embryos. The areas shown in more detail at the bottom of the panel are outlined in the merged images above. NF2 and Ezrin proteins in Nf2WT-Gfp/Ezrin-mCherry embryo (left) were localized to the apical domains in a mutually exclusive manner. Meanwhile, Ezrin localization in the Nf2L64P-Gfp/Ezrin-mCherry embryo (right) was thinly localized across the apical domain. Plot profiles of the white dashed lines in the enlarged images are shown in embryos microinjected with the Nf2WT-Gfp/Ezrin-mCherry (left) or Nf2L64P-Gfp/Ezrin-mCherry (right) mRNAs, respectively. Scale bars: 30 μm.
LATS2 recruitment to the apical domain was disrupted in NF2L64P-expressing cells (Fig. 4A). The centralization resulting from LATS2 recruitment to the apical domain was represented by narrow and thick fluorescent signals of LATS2 (Fig. 4A,B). Furthermore, 59.8% of the cells expressing Nf2WT-Gfp and LatsWT-mCherry showed narrow and thick LATS2 signals (Fig. 4B; type III), whereas only 4.0% of those expressing Nf2L64P-Gfp and LatsWT-mCherry showed such localization (Fig. 4B). Given that the apical domain formation is closely connected with cell polarity establishment, we investigated ezrin localization by microinjecting the Ezrin-mCherry mRNAs into embryos. The co-microinjection of the Ezrin-mCherry and Nf2WT-Gfp mRNAs revealed that these two proteins were localized to the apical domains in a mutually exclusive manner, showing centralized ezrin localization (Fig. 4C). In contrast, the embryos microinjected with the Ezrin-mCherry and Nf2L64P-Gfp mRNAs represented monotonous ezrin localization in the apical domains (Fig. 4C). These results suggest that inhibition of NF2 localization disrupts LATS2 recruitment and centralization of ezrin in the apical domains, indicating that LATS2 recruitment might depend on ezrin centralization, which is regulated by NF2 subcellular localization in the apical domains.
A FERM domain-binding domain mutation in LATS2
A FBD mutation (L83K) in LATS2 was prepared to investigate how LATS2 is recruited to the apical domains in early embryos. Substitution of leucine with lysine at amino acid position 83 of LATS2 by reference to the FBD of LATS1 (Li et al., 2015) inhibited LATS2 recruitment to the apical domains (Fig. 5A). LATS2L83K localization to the adhesion junction site was reinforced by the scaffold protein NF2 (Fig. 5A,B). Furthermore, the YAP1 N/C ratio in outer cells expressing LATS2L83K decreased when compared with that of outer cells without LATS2L83K expression (Fig. 5B,C). These results indicate that LATS2 recruitment to the apical domains relies on its FERM domain-binding ability. This suggests that the interaction between LATS2 and ERM proteins, including ezrin, through the FERM domain is crucial for LATS2 localization to the apical domains and for subsequent inactivation of the Hippo pathway in mouse embryos.
Regulation of LATS2 recruitment to the apical domains via FERM domain-binding domain in LATS2. The FERM-binding domain (FBD) is required for interaction with ERM proteins such as Ezrin. FBD-mutated Lats2 (Lats2L83K)-mCherry mRNA was prepared. (A) Nf2WT-Gfp/Lats2L83K-mCherry mRNA was microinjected into one-cell stage embryos (n=13). LATS2L83K could localize to the adherence junctional sites but not to the apical domains, owing to reinforcement by the scaffold protein NF2. (B) Immunostaining results of YAP1 proteins whose blastomeres were microinjected with Lats2L83K-mCherry mRNA at the two-cell stage of the embryos. LATS2L83K localization to the adherence junctional sites was not observed because NF2 was not overexpressed, as shown in A. YAP1 showed nuclear localization in the cells without Lats2L83K-mCherry expression (red arrowheads). However, cells with or without YAP1 nuclear localization were observed (white arrowheads). (C) Quantification of the YAP1 N/C ratio in the outer cells with (27 cells) or without (40 cells) LATS2L83K expression using five embryos microinjected with Lats2L83KP-Gfp mRNA. Scale bars: 30 μm.
Regulation of LATS2 recruitment to the apical domains via FERM domain-binding domain in LATS2. The FERM-binding domain (FBD) is required for interaction with ERM proteins such as Ezrin. FBD-mutated Lats2 (Lats2L83K)-mCherry mRNA was prepared. (A) Nf2WT-Gfp/Lats2L83K-mCherry mRNA was microinjected into one-cell stage embryos (n=13). LATS2L83K could localize to the adherence junctional sites but not to the apical domains, owing to reinforcement by the scaffold protein NF2. (B) Immunostaining results of YAP1 proteins whose blastomeres were microinjected with Lats2L83K-mCherry mRNA at the two-cell stage of the embryos. LATS2L83K localization to the adherence junctional sites was not observed because NF2 was not overexpressed, as shown in A. YAP1 showed nuclear localization in the cells without Lats2L83K-mCherry expression (red arrowheads). However, cells with or without YAP1 nuclear localization were observed (white arrowheads). (C) Quantification of the YAP1 N/C ratio in the outer cells with (27 cells) or without (40 cells) LATS2L83K expression using five embryos microinjected with Lats2L83KP-Gfp mRNA. Scale bars: 30 μm.
DISCUSSION
The Hippo pathway in non-polar cells is usually active in order to maintain an undifferentiated state. This facilitates LATS2-dependent YAP1 phosphorylation, leading to the inhibition of TE-specific transcription. LATS2-dependent YAP1 phosphorylation requires the formation of the AMOT-LATS-NF2 complex, as evidenced by coimmunoprecipitation experiments using HEK293T cells (Hirate et al., 2013). However, the mechanisms through which the subcellular localization of these molecules affects Hippo pathway inactivation in outer cells require further elucidation. In this study, NF2L64P expression detached both AMOT and LATS2 from the membrane of non-polar cells in early embryos, demonstrating that NF2 functions as a scaffold for Hippo pathway components during subcellular localization (Fig. 6). Our results indicate that the L64P missense mutation within the FERM domain is sufficient to disable the function of NF2 as a scaffold for Hippo pathway activation at the membrane of non-polar cells.
Schematic showing the roles of NF2 subcellular localization in the outer cells of mouse embryos. NF2 is a scaffold for Hippo pathway components, including phosphorylated AMOT (pAMOT) and LATS2, that facilitates YAP1 translocation into the nucleus of inner cells. In contrast, NF2 does not contribute to the scaffold for Amot and LATS2, but remains in the adherent junctional regions. The remaining NF2 may enable the localization of Amot and LATS2 to the apical domains because disruption of NF2 localization disturbs localization around the cell membrane. The mechanism through which LATS2 localizes to the apical domains in the outer cells is unclear. However, it could depend on ERM proteins, including EZRIN, because of the FERM domain-binding domain in LATS2 proteins.
Schematic showing the roles of NF2 subcellular localization in the outer cells of mouse embryos. NF2 is a scaffold for Hippo pathway components, including phosphorylated AMOT (pAMOT) and LATS2, that facilitates YAP1 translocation into the nucleus of inner cells. In contrast, NF2 does not contribute to the scaffold for Amot and LATS2, but remains in the adherent junctional regions. The remaining NF2 may enable the localization of Amot and LATS2 to the apical domains because disruption of NF2 localization disturbs localization around the cell membrane. The mechanism through which LATS2 localizes to the apical domains in the outer cells is unclear. However, it could depend on ERM proteins, including EZRIN, because of the FERM domain-binding domain in LATS2 proteins.
The role of NF2 in polar cells is less well-defined compared with its role in non-polar cells. The Hippo pathway is typically inactive in polar cells, necessitating the sequestration of AMOT to the apical domains. We observed recruitment of LATS2 to the apical domain (Fig. S2). However, the significance and regulation of LATS2 recruitment within the Hippo pathway remain unclear. Our study demonstrates that NF2 localization to apical domains is unlikely to be associated with LATS2 recruitment, as NF2 levels decrease in the apical domain after compaction and remain concentrated at adherens junction sites (Fig. 1). Moreover, disruption of NF2 localization at adherens junction sites prevents both AMOT and LATS2 from localizing to the apical domains (Figs 3 and 4A), leading to YAP1 mislocalization in polar cells (Fig. 5). Therefore, we hypothesize that another component interacting with NF2 regulates LATS2 recruitment to the apical domains (Fig. 6).
The role of NF2 in non-polar inner cells was fully addressed in a previous study (Hirate et al., 2013) elucidating that NF2 is required for activation of the Hippo pathway through interaction with LATS protein kinases. Difference in the LATS2WT-mCherry localization between the embryos injected with only Lats2WT-mCherry or with Lats2WT-mCherry/Nf2WT-Gfp mRNAs might support this model. The embryos microinjected with single Lats2WT-mCherry mRNA showed localization of LATS2WT-mCherry to the apical domains (Fig. S2), whereas LATS2WT-mCherry in the embryos co-microinjected with Lats2WT-mCherry and Nf2WT-Gfp mRNAs localized to both the apical domains and the adherens junction sites (Fig. 4A). Interestingly, LATS2WT-mCherry in the embryos co-microinjected with Lats2WT-mCherry and Nf2L64P-Gfp mRNAs was clearly reduced at both the apical domains and the adherens junction sites (Fig. 4A). This observation indicated that localization of NF2 to the adherens junction sites was required for accurate LATS2 localization, as the NF2L64P-GFP did not localize to the adherens junction sites (Fig. 1D). Thus, these results might suggest that Nf2 overexpression by mRNA microinjection increased the scaffold for LATS2 and promoted LATS2 localization to the adherens junction sites, supporting the molecular model suggested previously (Hirate et al., 2013).
AMOT-LATS2-NF2 complex formation in the adherens junction sites of non-polar cells requires phosphorylation of AMOT by LATS2 at S176. In addition, phosphorylation of AMOT is required for interaction with LATS2 (Hirate et al., 2013). However, non-phosphorylated AMOT, which does not interact with LATS2, is recruited to the apical domain of polar cells, as is LATS2 (Fig. S2). This demonstrates that the interactive partner(s) of LATS2 is not non-phosphorylated AMOT in polar cells (Fig. 6).
In this study, we evaluated ERM proteins possessing FERM domain other than NF2 because LATS2 possesses an FBD that forms a complex with the FERM domain (Li et al., 2015). Ezrin is one of the most representative cell polarization markers. It accumulates in the apical domain after compaction during the eight-cell stage (Zhu et al., 2020). We hypothesized that the ERM proteins, including ezrin, localized to apical domains in polar cells sequester LATS2 away from the Hippo pathway complex at adherence junctional sites. Consistent with this hypothesis, the L83K missense mutation within the FBD of LATS2 disrupted LATS2 recruitment to the apical domains (Fig. 5). Additionally, the introduction of Lats2L83K mRNAs inhibited YAP1 nuclear localization and Hippo pathway inactivation in polar cells (Fig. 5B). These results demonstrate that LATS2 recruitment to the apical domains is achieved through the interaction between the LATS2 FBD and FERM domains of ERM proteins, including ezrin, in polarized embryonic cells (Fig. 6). Furthermore, ezrin centralization in the apical domains was disrupted by NF2L64P expression (Fig. 4C), suggesting that NF2 showing exclusive localization with ezrin (Fig. 4C) is associated with the establishment of ezrin centralization. NF2L64P expression also inhibited LATS2 recruitment to the apical domains (Fig. 4A), followed by YAP1 mislocalization in polarized cells (Fig. 2B). These findings indicate that NF2 subcellular localization regulates ezrin centralization and induces LATS2 recruitment to the apical domains, which is required for Hippo pathway inactivation in polarized cells (Fig. 6).
Both proteins commonly belong to the ERM family and possess the N-terminal FERM domain, enabling their participation in scaffolding and signaling events at the cell cortex. This study revealed exclusive localization of NF2 and ezrin to the apical domains. However, the kinetic mechanism(s) governing this exclusive localization was not addressed. Structural features of NF2 might be associated with this mechanism(s), as NF2 lacks the C-terminal F-actin-binding domain present in other ERM proteins despite its striking structural similarities to them (Bretscher et al., 2002). Therefore, further studies on the interactions between NF2, ezrin and F-actin in early embryos could elucidate the mechanism underlying the exclusive localization of NF2 and ezrin in cell polarization and subsequent cell signaling, including the Hippo pathway. We also observed that both NF2S518A and NF2S518E mutations did not affect the subcellular localization of NF2. However, these results do not exclude the possibility that these mutations affect the regulation of the Hippo signaling pathway. To precisely understand the roles of NF2 in the Hippo signaling pathway during preimplantation development, further examination of YAP1 and CDX2 subcellular localization in these phosphorylation mutants is necessary. These mutations mediate Hippo signaling by altering the three-dimensional conformation of NF2 in cultured cells (Hong et al., 2020; Li et al., 2010, 2015; Yin et al., 2013). Furthermore, the association between LATS2 and ERM proteins has not been fully proven, and further studies will be needed.
Concluding remarks
Cell polarity, which results from the asymmetric distribution of cellular structures, molecules and functions, is a fundamental process during preimplantation development. This process is highly conserved and stereotyped across species. Regulation of Hippo signaling in early embryos is contingent upon cell polarization. However, the relationship between cell polarization and Hippo signaling inactivation in apical cells remains unclear. This study demonstrates that the spatial-temporal coordination of NF2, LATS2 and ezrin during cell polarization plays crucial role in the inactivation of the Hippo signaling pathway. These findings suggest a molecular mechanism underlying regulation of the Hippo pathway during early embryonic development through cell polarity.
MATERIALS AND METHODS
Ethical approval
All animal experiments were approved by the Regulatory Committee for the Care and Use of Animals of Hokkaido University and were performed in accordance with the National University Corporation Hokkaido University Regulations on Animal Experimentation.
Embryo preparation
Female ICR mice were superovulated 48 h apart through the intraperitoneal injection of 7.5 IU equine chorionic gonadotropin (eCG; ASKA Pharmaceutical) and 7.5 IU human chorionic gonadotropin (hCG; ASKA Pharmaceutical). Sperm were collected from the cauda epididymis of male ICR mice, suspended in a 200 μl drop of human tubal fluid medium (HTF) (Quinn et al., 1985) in paraffin oil, and pre-incubated for 90 min in a 5% CO2 atmosphere at 37°C. Oocytes at metaphase II were collected from murine oviducts 16 h after hCG administration and transferred to a 100 μl drop of HTF containing 0.5-1×106 sperm/ml. The embryos were washed with M2 medium (Whittingham, 1971) 4-6 h after insemination to remove cumulus cells. They were subsequently transferred to a drop of M16 medium (Brackett and Oliphant, 1975) for in vitro culture.
Plasmid construction
Plasmids were constructed as previously described (Komatsu et al., 2021), with modifications. The templates of Nf2- and Gfp-coding sequences were prepared from the pBabeNf2 wild-type (Addgene, 14116) and CS-CA-Gfp (RIKEN, RDB05964) vectors, respectively, using polymerase chain reaction (PCR). The primer sets used for PCR were as follows: Nf2 forward, 5′-ACGGAATTCGCCACCATGGC-CGGAGCCATCGCTTC-3′; Nf2 reverse, 5′-TGCTCACCATAATGCAGATAGGTCTTCTGCCTTG-3′; Gfp forward, 5′- TATCTGCATTATGGTGAGCAAGGGCGAGG-3′; Gfp reverse, 5′- ACGGGATCCCTAGTAGGATCTGAGTCCGGAC-3′. The wild-type Nf2 gene (Nf2WT) was translationally fused to the Gfp gene at the 3′ or 5′ ends using PCR (Horton et al., 2013). PCR hybridizes and subsequently amplifies two partially complementary DNA sequences to obtain a fusion gene. The Nf2WT-Gfp gene was subcloned into the pGEM-3Zf(+) vector (Promega). Additionally, the templates of Lats2- and mCherry-coding sequences were prepared from, respectively, pcDNA3.1-pA83-Lats2 (RIKEN, RDB12200) and H2B-mCherry vectors constructed in a previous study (Wakai et al., 2014). The primer sets used for PCR were as follows: Lats2 forward, 5′-ACGGAATTCGCCACCATGAGGCCAAAGACTTTTCC-3′; Lats2 reverse, 5′-CTTGCTCACCATCACGTACACCGGCTGGCAGC-3′; mCherry forward, 5′-GGTGTACGTGATGGTGAGCAAGGGCGAGGA-3′; mCherry reverse, 5′-ACGGGATCCTCTAGACTACTTGTACAGCTCGTC-3′. The Lats2 gene was also fused to the mCherry gene using PCR, and the Lats2WT-mCherry gene was then subcloned into the pGEM-3Zf(+) vector. The templates of Ezrin-coding sequences were prepared from total RNAs derived from ICR mouse kidney as previously described (Zhu et al., 2020). The primer sets used for PCR were as follows: ezrin forward, 5′-TTGGAGTGAGATCAGGAACATC-3′; ezrin reverse, 5′-AAGGAGGACGAGGTAGAAGAGTG-3′. After adding the EcoRI recognition site and Kozak sequences through PCR, ezrin was fused to mCherry using the following primer sets: ezrin forward, 5′-CTTGCTCACCATCATGGCCTCGAACTCGTCAATG-3′; ezrin reverse, 5′-GAGGCCATGATGGTGAGCAAGGGCGAG-3′; mCherry forward, 5′-GAGGCCATGATGGTGAGCAAGGGCGAG-3′; mCherry reverse, 5′-ACGAAGCTTCTACTTGTACAGCTCGTCC-3′.
To create three types of Nf2 point mutants (L64P, S518A and S518E), the 5′ side fragments (from the 5′ end of Nf2 to the mutation point) and 3′ side fragments (from the mutation point to the 3′ end of Gfp) of the Nf2WT-Gfp gene were amplified using mismatch primers with each point mutation. The mismatch primer sets were as follows: Nf2L64P forward, 5′-TCTTTGGACCGCAGTATACAATCAAGGACACG-3′; Nf2L64P reverse, 5′-GATTGTATACTGCGGTCCAAAGAACCAGGTTTC-3′; Nf2S518A forward, 5′-GAAGCGACTTGCTATGGAGATAGAGAAAGAA-3′; Nf2S518A reverse, 5′-CTATCTCCATAGCAAGTCGCTTCATGTCCGTATC-3′; Nf2S518E forward, 5′-GAAGCGACTTGAAATGGAGATAGAGAAAGAA-3′; Nf2S518E reverse, 5′-CTATCTCCATTTCAAGTCGCTTCATGTCCGTATC-3′. Each fragment was fused using PCR. The three mutants, Nf2L64P-Gfp, Nf2S518A-Gfp and Nf2S518E-Gfp, were subcloned into the pGEM-3Zf(+) vector and NfWTs-Gfp gene for in vitro transcription and microinjection.
Similarly, the mismatch primer sets for creating the Lats2 point mutant (L83 K) were as follows: Lats2L83K forward, 5′-GGGAAATCCGATATTCCAAGCTGCCTTTTGCCAAC-3′; Lats2L83K reverse, 5′-GTTGGCAAAAGGCAGCTTGGAATATCGGATTTCCC-3′.
Microinjection of capped RNAs
Capped RNAs from the inserted vectors were transcribed in vitro using the mMESSAGE mMACHINE T7 Ultra Kit (Thermo Fisher Scientific) and a poly A tail was added to the 3′ end using a Poly(A) Tailing Kit (Thermo Fisher Scientific). Subsequently, 200 ng/µl of purified RNA diluted with injection buffer [10 mM Tris-HCl (pH 7.4) and 0.1 mM EDTA] was microinjected into one-cell stage embryos or one of the blastomeres of two-cell stage embryos using a FemtoJet injection device (Eppendorf).
Immunostaining and live-cell imaging analyses
The primary antibodies used include mouse anti-YAP1 (1:100, H00010413-M01; Abnova), mouse anti-AMOT (1:100, Hirate et al., 2013), rabbit anti-CDX2 (1:200, ab76541; Abcam) and mouse anti-NF2/Merlin (1:100, AF1G4; Abcam). The following secondary antibodies were used: Alexa Fluor 555 goat anti-rabbit IgG (A21428, polyclonal, 1:400; Invitrogen), Alexa Fluor 488 goat anti-mouse IgG (A11001, polyclonal, 1:400; Invitrogen), Alexa Fluor 488 goat anti-rabbit IgG Cross-Adsorbed (A11008, 1:400; Invitrogen) and Alexa Fluor 555 goat anti-mouse IgG (A21422, polyclonal, 1:400; Invitrogen).
Embryos were fixed with 4% (w/v) paraformaldehyde (PFA; Wako Pure Chemical Industries) in PBS for 60 min and permeabilized for 20 min with 0.2% (v/v) Triton X-100 in PBS. The embryos were then blocked for 5 min with a blocking buffer: PBS containing 2.0% (w/v) bovine serum albumin (Sigma-Aldrich). Thereafter, the embryos were incubated in blocking buffer containing the primary antibody for 8 h at room temperature (20-22°C). After washing three times in PBS containing 0.2% (w/v) bovine serum albumin (Sigma-Aldrich) for 2 min, the embryos were incubated for 30 min at room temperature with a secondary antibody diluted to 1:400 in washing buffer. For only NF2 immunostaining, embryos were fixed in glyoxal (Sigma-Aldrich) rather than paraformaldehyde, as previously reported (Konno et al., 2023) and permeabilized for 60 min. After washing in 0.02% PVA in PBS for 15 min, the embryos were blocked for 60 min with Blocking One (1:5; Nacalai Tesque) diluted in 0.01% (v/v) Tween 20 in PBS. Subsequently, the embryos were incubated in Can Get signal solution B (Toyobo) containing the primary antibody for 8 h at 4°C. After washing in 0.2% PVA in PBS for 15 min, the embryos were incubated for 30 min at room temperature with a secondary antibody diluted to 1:400 in PBS containing 2.0% (w/v) bovine serum albumin. Nuclei were counterstained with 25 mg/ml Hoechst 33342 (Sigma-Aldrich) in 0.2% (w/v) polyvinyl alcohol in PBS. Fluorescence signals were then visualized using a TCS SP5 confocal laser-scanning microscope (Leica). Two-cell stage embryos injected with purified Nf2WT-Gfp mRNA were used for live-cell imaging analysis. The images were captured every 10 min in ten z-sections per time point using a Confocal Scanner Unit (CSU-W1; Yokogawa Electric Corporation).
For analysis of the YAP1 fluorescence intensity, the average fluorescence intensity in the nucleus and immediately outside the nucleus, i.e. the cytoplasm, was measured with ImageJ program. The YAP1 nuclear/cytoplasm (N/C) ratio was determined in each blastomere based on those average fluorescence intensity. Plot profiles along the apical domains of the embryos microinjected with Nf2WT-Gfp/Ezrin-mCherry or Nf2L64P-Gfp/Ezrin-mCherry were also measured using Plot Profile Tool in ImageJ program.
Statistical analyses
All statistical analyses in this study were performed using GraphPad Prism10 software. Presented error bars (SD), statistical analyses and P values were calculated using an Was this test paired or unpaired? Student's t-test.
Acknowledgements
We thank Dr Hiroshi Sasaki of Osaka University for providing the anti-Amot antibody. We are also grateful to the members of the Laboratory of Animal Genetics and Reproduction at the Hokkaido University for their contributions to animal care.
Footnotes
Author contributions
Conceptualization: M.K.; Methodology: T.W.; Validation: N.G., Y.I., T.W., M.K.; Formal analysis: N.G., M.K.; Investigation: N.G., Y.I., S.S., M.S., T.W.; Resources: H.B., M.T., M.K.; Data curation: N.G., Y.I., M.K.; Writing - original draft: M.K.; Writing - review & editing: N.G., Y.I., S.S., H.B., M.K.; Visualization: N.G.; Supervision: M.K.; Project administration: M.K.; Funding acquisition: M.K.
Funding
This work was partially funded by Grants-in-Aid for Scientific Research (B) from the Japan Society for the Promotion of Science (21H02336 and 24K01902 to M.K.).
Data availability
All relevant data can be found within the article and its supplementary information.
References
Competing interests
The authors declare no competing or financial interests.