The physiological role of Motin family and its dysregulation in tumorigenesis

During the preimplantation stage, mouse embryos establish two cell lineages by the time of early blastocyst formation: the trophectoderm (TE) and the inner cell mass (ICM) [12]. AMOT, AMOTL1, and AMOTL2 behave variously in embryo development. In early mammalian embryo, AMOT is differentially expressed in the progenitors of the pluripotent ICM and the differentiated TE precursors [13, 14]. The appropriate formation of TE and ICM lineages is dependent on the differential activity of the Hippo-signaling pathway between the outer- and inner-cell populations [15]. The inner cells show high levels of Hippo pathway while outer cells of the embryo show low Hippo pathway [13, 14]. These different activities of Hippo pathway was at least partly regulated by AMOTs. At the adherent junctions of the inner cells, AMOT recruits and activates LATS1/2 to promote Hippo pathway [13, 14]. In apical membrane of the outer cells, F-actin inhibits the ability of AMOT to activate Hippo signaling [13, 14]. AMOT phosphorylation disrupts AMOT—F-actin interaction and leads to reduced F-actin stress fibers and focal adhesions. Also, this phosphorylation was reported to suppress endothelial cell migration in vitro and angiogenesis in zebrafish embryos in vivo [16]. The localization of AMOT is regulated by Rho-associated kinase (Rock) to regulate appropriate apical-basolateral polarization in outer cells [17]. In AMOTL2 knockdown (KD) embryos, they exhibit a normal YAP1 distribution and Hippo pathway activation. However, with AMOTL2 KD in AMOT-null mutants, these embryos exhibit strong nuclear localization of YAP1 in the inner cells, suggesting that some Motin family proteins are essential for the activation of Hippo pathway in preimplantation embryos [18]. During the peri-implantation period, AMOT, AMOTL1, and AMOTL2 are differentially expressed in uterine cells [19]. There was evidence showing a transitional expression pattern of AMOT during the embryonic development: from the anterior visceral endoderm (AVE) to the visceral endoderm (VE) subsequently, and the latter was associated with extra-embryonic ectoderm. In subregions of VE, AMOT regulated morphogenetic movements that were required for embryo viability [20]. AMOT-null mutant embryos failed to exclude YAP1 from the nuclei of the inner cells until 32-cell stage [18]. The embryos survived until the postimplantation stages [18].

On the other hand, AMOTL2 is expressed maternally and in restricted cell types zygotically [21]. Silencing AMOTL2 expression impaired convergence and extension movement [21]. In mosaic embryos, cells with AMOTL2 knockdown failed to migrate properly [21]. AMOTL2 partially co-localized with RhoB- or EEA1- positive endosomes and tyrosine kinase c-Src, which in turn regulated the membrane architecture [21]. Thus, AMOTL2 is essential for cell movements in vertebrate embryos [21]. In zebrafish, AMOTL2 regulates embryonic development through inhibiting Wnt/β-catenin signaling [22]. It can interact with and trap β-catenin in the Rab11-positive recycling endosomes to reduce the amount of β-catenin in both the cytosol and nucleus [22].

AMOT was first identified to mediate angiostatin inhibition of endothelial migration and tube formation in vitro [1]. In zebrafish, knocking down AMOT was shown to decrease the number of filopodia of endothelial tip cells [23]. AMOT is also required for endothelial polarization during migration in that AMOT regulates Rac1 activity in endothelial and epithelial cells [23]. When transgenic mice expressed C-terminal deletion mutant AMOT through endothelial cell-specific receptors tyrosine kinase (TIE) promoter, migration of endothelial cells into the neuroectoderm and intersomitic regions was inhibited [24]. AMOT also interacts with tissue factor pathway inhibitor-1 (TFPI-1), which leads to YAP1 activation and further increases the genes expression relating to proliferation and migration [25]. In Madin–Darby canine kidney (MDCK) cells, overexpression of AMOT induces relocalization of polarity proteins and loss of transepithelial electrical resistance [26]. And the interaction between AMOT and Rich1 (RhoGAP interacting with Cdc42-interacting protein four homologues protein 1) is necessary for maintaining TJ integrity [26]. MUPP1, which has an MRE domain and 13 PDZ domains, is located at TJs. The interaction between MUPP1 and AMOT has been identified in epithelial cells by yeast two-hybrid screening [27]. Patj is a close relative of MUPP1. Possessing a similar structure, Patj interacts with all AMOT family members to regulate the formation of TJ and epithelial polarity [27]. Pdlim2, a member of actin-associated LIM proteins subfamily, is expressed exclusively by podocytes in kidney. AMOTL1 interacts with Pdlim2 and governs the dynamics of the actin cytoskeleton in foot processes [28].

Seventy five percentage of AMOT knockout mice died between embryonic day E11 and E11.5 [23]. Most of them exhibited severe vascular insufficiency in the intersomitic region and dilated vessels in the brain [23]. In addition, anti-AMOT antibody significantly inhibits endothelial migration and decreases the number of endothelial filopodia of tip cells during retinal angiogenesis [29]. Moreover, AMOT is a pivotal adaptor protein in the intersection between trafficking, cell junctions and cell migration, which plays a role in directional migration and angiogenesis [30]. Pals1-associated TJ protein (Patj), Lin Seven 1 (Pals1), and Par-3 are crucial for cell polarity. Together with Patj and synectin-binding GEF (Syx), AMOT forms a ternary complex. In vivo, AMOT plays an additive role with Syx in directing endothelial sprouts [31]. AMOT-Patj/multi-PDZ-domain protein 1 (Mupp1)-Syx govern the directional migration of capillaries in the embryo signaling through controlling RhoA GTPase activity to the leading front of migrating cells [32]. AMOT expression is up-regulated in dermal mesenchymal stem cells (DMSCs) in psoriasis, suggesting its involvement in the excessive angiogenesis and vasodilation [33].

AMOT distributes on cell surface and co-localizes with tight junction (zonula occludens) protein (ZO-1) in cell–cell contracts in endothelial cells in vitro and in angiogenic blood vessels of the postnatal mouse’s retina in vivo [3]. AMOT-p130 recruits ZO-1 to actin stress fibers, which is responsible for its localization to actin and tight junctions (TJs) [3]. Furthermore, AMOT-p130 coprecipitates with MAGI-1b, a component of endothelial TJs. Therefore, AMOT contributes to the assembly of endothelial cell–cell junctions [3]. AMOT-p80 and AMOT-p130 play coordinating roles in tube formation by affecting cell migration and cell shape respectively [4]. They were suggested to be differentially expressed in specific stages of mouse retina vascularization [34]. During retinal angiogenesis in vivo, AMOT-p80 was found to be expressed in the migratory phase [34]. In contrast, AMOT-p130 was detected during the period of blood vessel stabilization and maturation [34]. They have distinct functions, AMOT-p80 stimulated endothelial cell migration and angiogenesis while AMOT-p130 stabilized and matured vessels [34]. The rate of AMOT-p80 and AMOT-p130 expression served as an indicator for migration or stabilization of endothelial cells: increased AMOT-p80/-p130 ratio directly reflected the enhanced angiogenic ability of skeletal muscle in response to exercise training [35].

Similar to AMOT, AMOTL1 was also indicated to play a role in endothelial migration and TJ formation in vitro [36]. AMOTL1 mainly affected the stability of cell–cell junctions in stalk cells during sprouting angiogenesis in vivo [36]. It increased the velocity of migration and decreased the persistence of migration directionality. AMOTL1 interacted with AMOT-p80 to form a complex and induced the remodeling of actin cytoskeleton [37]. Motin family has also involved in controlling stability and permeability of endothelial cell junction. HECW2 (HECT, C2 and WW domain containing E3 ubiquitin protein ligase 2), the endothelial cell ubiquitin E3 ligase, physically interacts with AMOTL1 and enhances its stability [38]. In normal retinal, AMOTL1 is essential for normal establishment of vascular networks in the post-natal mouse retina. It forms a complex with N-cadherin to mediate endothelial junctional complex [39].

AMOTL2, accordingly, was detected in blood vessel cells, suggesting its critical roles in regulating multiple behaviors of endothelial cells during angiogenesis. In zebrafish transgenic embryos, knockdown of AMOTL2 impaired the intersegmental vessel growth and suppressed proliferation of endothelial cells [40]. AMOTL2 knockdown also inhibited cell proliferation and migration and disrupted cell polarity of cultured human umbilical vein endothelial cells [40]. It was required for MAPK/ERK activation during angiogenesis [40]. AMOTL2 also participates in aortic vessel lumen expansion through linking VE-cadherin to contractile actin fibers, which has been verified in zebrafish, mouse, and endothelial cell culture systems [41].

Role of AMOTs in Hippo-YAP1 pathway is not clearly proven, even still highly controversial. Some laboratories have shown they inhibit cell proliferation, where others demonstrated the promotion of cell proliferation or tumor growth. Hippo pathway functions as an organ size controller, mainly through its ability to regulate cell proliferation and apoptosis [42, 43]. The core kinase cascade in mammals consists of MST1/2 kinases, WW45, LATS1/2 and Mob1. MST1/2 interacts with WW45 and phosphorylates Mob1 and LATS1/2, resulting in their activation [44]. Activated LATS1/2 phosphorylates YAP1/TAZ. YAP1 and TAZ are the nuclear effectors of Hippo pathway and they belong to the paralogous multifunctional co-activators [45]. AMOTs correlate with YAP1 through the binging of their PPXY and LPXY motifs with the WW domains of YAP1 and TAZ [5‐7]. YAP1 is a transcriptional co-activator and a key component in Hippo pathway. Therefore, Motin family proteins have been also identified as novel components in Hippo-YAP1 pathway [6, 46].

Merlin functions upstream of the core Hippo pathway kinases LATS1/2 and MST1/2. An AMOT-dependent complex comprised of AMOT, YAP and Merlin has been described [47]. Merlin is a kind of TJ associated protein and interacts directly with AMOT through their mutual coiled-coil domains. In epithelial cells, AMOT retains Merlin at mature TJs. In addition, Merlin functions through AMOT and Rich1 to inhibit Rac1 and Ras-MAPK signaling. Interaction of AMOT and Merlin regulates Rac signaling via Cdc42/Rac1 GAP and Rich1. Therefore, AMOT is considered as a potential player in Merlin-related cancers [6, 48]. An elongated α-helix-rich region of AMOT (residues 404–728 of AMOT-p130) robustly bound to the entire C-terminal part of Merlin to form a stable complex [49]. This interaction releases the auto-inhibition and promotes Merlin’s binding to LATS1/2 to activate Hippo pathway [49].

AMOT was first identified as a partner of YAP1 via large screening of multi-protein complexes that were assembled on de-ubiquitinating enzymes [50]. In MDCK cells, AMOT inhibits YAP1-induced transformation and loss of cell contact inhibition. Knockdown AMOTL2 induces MDCK cell transformation in a YAP1/TAZ dependent manner [6]. AMOTL1, YAP1 and ZO-2 form a tripartite complex to regulate the function of YAP1 in HEK293 cells: AMOTL1 inhibits proapoptotic function of YAP1, while ZO-2 enhance it [51]. AMOTL1 was also reported to be stabilized by being directly phosphorylated by AMPK, which contributing to YAP1 inhibition [52]. On the other hand, AMOTL2 interacts with YAP1 and the Wnt/β-catenin effector Lef1 to control tissue size [53]. Moreover, mono-ubiquitination in turn mediates the function of AMOTL2 itself [54]. Moreover, Hippo-refractory-TAZ mutant (S89A) was reported to be negatively regulated by AMOT and AMOTL1. These results suggest a mechanism of Hippo pathway-independent restriction of TAZ and YAP1 by AMOTs, indicating that the function of AMOTs compensate for the absence of LATS1/2 kinases [55]. As for mammalian cells, AMOT could also regulate YAP1 activity through competitively interacting with endosomal integral membrane protein endotubin (EDTB) [56]. Overexpression of AMOT and AMOTL1 leads to the cytoplasmic retention of TAZ and downregulation of CTGF and Cyr61 expression. Cytoplasmic YAP1 recruits c-Abl to protect AMOTL1 against Nedd4.2-mediated degradation. Therefore, cytoplasmic YAP1 involves in the protecting of AMOTL1 [57].

Tumor suppressive functions of AMOTs have been identified in different cancer types. Overexpression of YAP1 impairs the nuclear retention of SRSF1 (serine/arginine-rich splicing factor 1) and itself via AMOT [58]. Therefore, AMOT mediates translocation of SRSF1 through YAP1 in liver cancer cells [58]. In clinical lung cancer specimens, AMOT expression was found to be significantly decreased, which indicated its tumor suppressive role. In vivo, AMOT knockdown increased the growth and spread of Lewis lung carcinoma [59]. Besides, tankyrase inhibitor sensitized cancer cells to endothelial growth factor receptor (EGFR) inhibition through stabilizing AMOT and inhibiting YAP1 signaling in lung cancer cells [60]. Tankyrases inhibitors stabilize AMOT family proteins, disabling the oncogenic function of YAP1 [61, 62]. Through controlling the stability of the AMOT family proteins, E3 ligase ITCH, and the LATS kinases, DUB3 has the potential to act as a tumor suppressor by limiting YAP/TAZ activity [63]. LPTY and PPEY domains are required for full activity of AMOT. Unlike AMOT-p130, due to the lack of these motifs, AMOT-p80 fails to bind with YAP1. Thereby, it is AMOT-p130, instead of AMOT-p80, that promotes YAP1 phosphorylation and inhibits YAP1 transcriptional activity [8]. In addition, AMOT undergoes proteasomal degradation. Three members of Nedd4 (neural-precursor-cell-expressed developmentally down-regulated)-like ubiquitin E3 ligases, Nedd4, Nedd4-2 and Itch, mediate poly-ubiquitination of AMOT-p130 [8]. Nedd4-like ubiquitin E3 ligases competes with YAP1 to bind AMOT-p130 and subsequently targets AMOT-p130 for ubiquitin-dependent degradation [8]. Ubiquitin ligase atrophin-1 interacting protein 4 (AIP4) is a member of the Nedd4 family. AMOT-p130 is ubiquitinated by AIP4 at residue Lys-481 [64]. AMOT-p130 activates AIP4, resulting in the catalysis of ubiquitination of AIP4, AMOT-p130 and YAP1, meaning that AMOT-p130 works interdependently with AIP4 to reduce the stability of YAP1 and inhibit YAP1-dependent transcription as well as cell growth [64]. Hence, AMOT-p130 and AIP4 cooperatively mediate anti-growth partly through the inhibition of YAP1 [64]. AMOTs could also serve as direct activators of LATS2. AMOTL1, AMOTL2 and AMOT-p130 bind with LATS2 and YAP1 to promote LATS2 phosphorylating YAP1 [65]. Post-translational modifications of AMOT family proteins are critical for their context-dependent functions during carcinogenesis. The N-terminal domain of AMOT contains a consensus motif (HxRxxS) for phosphorylation by LATS1/2 [18]. The phosphorylation of AMOT at Serine 176 shifts localization of this complex to the plasma membrane, where it correlates with the tight-junction proteins Pals1/PATJ and E-cadherin. Therefore, phosphorylation of AMOT S175 is a critical post-translational modification that suppresses YAP1’s ability of promoting cell proliferation and tumorigenesis [47]. The phosphorylation of AMOT-p130 by LATS is a key feature that enables it to inhibit YAP1-dependent signaling and cell growth [66]. Phosphorylation of AMOT by LATS1/2 inhibits F-actin binding, cell migration and angiogenesis [16]. AMOT S175 is crucial for the actin-binding activity. Wild-type and non-phosphorylated AMOT (AMOT-S175A) interact with actin filaments, whereas phospho-mimic AMOT (AMOT-S175D) fails to be localized with actin [67]. S175A promotes cell proliferation, while AMOT and S175D inhibits it [67]. Motin family members are required for YAP1 cytoplasmic relocalization in response to the actin cytoskeleton perturbation [68]. F-actin and YAP1 compete to bind with AMOT-p130, accounting for how F-actin inhibits AMOT-p130-mediated cytoplasmic retention of YAP1 [68]. Furthermore, LATS synergizes with F-actin perturbations by phosphorylating free AMOT-p130 to keep it from interacting with F-actin [68]. Mutants’ defective of AMOTs for F-actin binding shows an enhanced ability to retain YAP1 in the cytoplasm. LATS2-phosphorylated AMOT-p130 inhibits the association of AMOT-p130 with F-actin, and AMOT-p130 binding to F-actin inhibited LATS phosphorylation. In short, F-actin and YAP1 competes to bind with AMOT-p130 [68]. Down-regulation of AMOTL2 promotes epithelial-mesenchymal transition (EMT), which was also observed when YAP1 is overexpressed [69].

In some specific cancer types, oncogenic functions of AMOT have also been revealed. RNF146 (an E3 ubiquitin ligase that recognizes ADP-ribosylated substrates) and tankyrase stabilize the Crumbs complex through downregulation of AMOT proteins at the apical membrane [70]. In breast cancer, AMOT was observed to be highly expressed in cancer tissues comparing with adjacent non-cancerous tissues. Decreased AMOT inhibited cell proliferation and invasion. In addition, the expression of YAP1 and LATS1 was suppressed, especially the expression of YAP1, which was significantly decreased [71]. In renal cell carcinoma (RCC), AMOT maintained nuclear localization of YAP so as to increase cell proliferation [72]. Atypical protein kinase C1 (PKC1) was suggested to be an oncogene in lung and ovarian cancer. Knockdown PKC1 enhanced the binding between YAP1 and AMOT, which retained YAP1 in the cytoplasm. PKC1 directly phosphorylated AMOT at Thr750, whose phosphorylation decreases YAP1 binding. PKC1-AMOT-YAP1 signaling axis enhanced the growth of ovarian serous carcinoma tumor [73]. AMOT-p80 functions as a tumor promoter by enhancing PCa cell proliferation. AMOT-p80 promotes nuclear localization of YAP1 through the Hippo pathway, which resulting in an overexpression of YAP1 targeting protein BMP4 [74]. Oncogenic functions of AMOT-p130 have also been justified. Yi et al. [75] suggested that AMOT-p130 was required for YAP1-mediated hepatic epithelial cell proliferation and tumorigenesis. Mice with a liver-specific AMOT knockout presented reduced hepatic ‘oval cell’ proliferation and tumorigenesis in response to toxin-induced injury [75]. These researchers also revealed that AMOT-p130 interacts with YAP1 in both cytoplasm and nucleus. In the cytoplasm, AMOT-p130 prevented the phosphorylation of YAP1 by blocking the access of the WW domains to the kinase LATS1 [75]. In the nucleus, AMOT-p130 interacted with the transcriptional complex containing YAP1 and TEADs to induce the expression of downstream targets [75]. Based on the facts above, AMOT promoted YAP1 nuclear translocation and acted as a transcriptional co-factor of the YAP1-TEAD complex to facilitate proliferation of biliary epithelial cells and cancer development of the liver [75]. In glioblastoma, phosphorylation of AMOTL2 by the mTORC2 kinase enhances YAP1 signaling, leading to cancer growth and invasiveness [76].