Abstract
The radial migration of newborn neurons is critical for the lamination of the cerebral cortex. Proper neuronal migration requires precise and rapid reorganization of the actin and microtubule cytoskeleton. However, the underlying signaling mechanisms controlling cytoskeletal reorganization are not well understood. Here, we show that Mst3, a serine/threonine kinase highly expressed in the developing mouse brain, is essential for radial neuronal migration and final neuronal positioning in the developing mouse neocortex. Mst3 silencing by in utero electroporation perturbed the multipolar-to-bipolar transition of migrating neurons and significantly retards radial migration. Although the kinase activity of Mst3 is essential for its functions in neuronal morphogenesis and migration, it is regulated via its phosphorylation at Ser79 by a serine/threonine kinase, cyclin-dependent kinase 5 (Cdk5). Our results show that Mst3 regulates neuronal migration through modulating the activity of RhoA, a Rho-GTPase critical for actin cytoskeletal reorganization. Mst3 phosphorylates RhoA at Ser26, thereby negatively regulating the GTPase activity of RhoA. Importantly, RhoA knockdown successfully rescues neuronal migration defect in Mst3-knockdown cortices. Our findings collectively suggest that Cdk5–Mst3 signaling regulates neuronal migration via RhoA-dependent actin dynamics.
Introduction
Cortical functions require precise wiring of distinct laminar-specific neural circuitry (Douglas and Martin, 2004). Radial migration, in which newborn neurons radially migrate to their final locations in the cerebral cortex, is one of the key steps that determines cortical function (Sidman and Rakic, 1973; Guerrini et al., 2008) and disruptions to neuronal migration are frequently associated with cortical malformation disorders and neuropsychiatric diseases such as schizophrenia and autism (Gleeson and Walsh, 2000; Deutsch et al., 2010; Singh et al., 2010; Valiente and Marin, 2010; Wegiel et al., 2010). Multiple cellular events are precisely regulated and coordinated to ensure proper neuronal migration (Bielas et al., 2004; Ayala et al., 2007; Creppe et al., 2009; Guerrier et al., 2009). A transient multipolar migration occurs in the lower intermediate zone (Nadarajah and Parnavelas, 2002; LoTurco and Bai, 2006). The subsequent multi-to-bipolar transition allows the migrating neurons to acquire a single leading process and initiate the glia-guided locomotion (Nadarajah and Parnavelas, 2002; LoTurco and Bai, 2006). Impairment of the transition leads to the inability of migrating neurons entering the cortical plate and eventually causes severe migration defects such as heterotopia (Ayala et al., 2007). However, the molecular mechanisms underlying the regulation of multipolar-to-bipolar transition remain unclear.
Emerging evidence suggests that protein kinases are important regulators for neuronal migration via the modulation of the cytoskeletal dynamics (Kawauchi and Hoshino, 2008; Su and Tsai, 2011). For example, cyclin-dependent kinase 5 (Cdk5), a serine/threonine kinase, plays a key role in neuronal migration by phosphorylating and regulating cytoskeletal regulators such as the actin-binding protein cofilin and the microtubule-binding protein doublecortin (Tanaka et al., 2004; Kawauchi et al., 2006). Cdk5 also triggers the activation of various kinase cascades to coordinate various cellular events during neuronal migration, such as the phosphorylation of focal adhesion kinase and p21-activated kinase (Rashid et al., 2001; Xie et al., 2003). Nonetheless, it remains unclear how these kinases are precisely regulated during neuronal migration.
Mammalian Ste20-like kinase (Mst), a serine/threonine kinase family member originally identified as a regulator of cell death, was recently reported to play important roles in the regulation of neuronal functions (Lehtinen et al., 2006; Ling et al., 2008; Xiao et al., 2011). The Drosophila homolog of Mst, Hpo, regulates dendrite development (Emoto et al., 2006), whereas an isoform of Mst3 has been implicated in axon outgrowth and regeneration (Irwin et al., 2006; Lorber et al., 2009). However, the precise roles of Mst3 in the developing CNS and its molecular regulation are unclear.
In this study, we report an essential role of Mst3 in neuronal positioning. In utero silencing of Mst3 disrupts the multipolar-to-bipolar transition and significantly retards radial migration. Moreover, we found that the kinase activity of Mst3 is regulated via the Cdk5-dependent phosphorylation at Ser79. Interestingly, Mst3 regulates neuronal migration through modulating the activity of RhoA, a Rho-GTPase critical for actin reorganization (Hall, 1994). RhoA knockdown restored normal neuronal migration in Mst3-knockdown cortices. Our findings collectively suggest that Cdk5–Mst3 signaling regulates neuronal migration via RhoA-dependent actin dynamics.
Materials and Methods
Constructs.
FLAG-tagged full-length mouse Mst3 plasmids were constructed by cloning the cDNAs of Mst3 wild-type (WT) or its kinase-dead mutant into the pCDNA3 vectors. For gene knockdown by RNA interference (RNAi), pSUPER vector-based small hairpin RNAs (shRNAs) of Mst3, Mst3-scramble, and RhoA were constructed. The shRNA target sense sequences for Mst3, Mst3-scramble, and RhoA were 5′-GGACTTGATTATCTACACT-3′, 5′-GTCATAATCGCATGTCTTA-3′, and 5′-GAAAGCAGGTAGAATTGGC-3′, respectively (Conery et al., 2010). RNAi-resistant Mst3 expression constructs were generated by introducing three silent mutations into the cDNA sequence targeted by Mst3 shRNA; the primer sequence was 5′-CGAGAAATTCTGAAAGGACTAGACTACTTACATTCGGAGAAGAAAATTC-3′ and the amino acids were not changed.
Antibodies.
Antibodies against Mst3 (sc-135993), HA (sc-805), Cdk5 (DC17), p35 (C-19), and RhoA (sc-418) were purchased from Santa Cruz Biotechnology. Anti-Mst3 (3723S) and anti-phospho-(Ser) CDKs substrate (2324S) were from Cell Signaling Technology. Antibodies against GAPDH (AM4300), Mst3 (EP1468Y), and phospho-serine (AB1603) were from Ambion, Epitomics, and Millipore, respectively. Anti-FLAG (M2), anti-Tuj1 (T3952), anti-CS-56 (C8035), and anti-β-actin (A5316) were from Sigma. Anti-phospho-Ser79-Mst3 antibody was raised in a rabbit immunized with a synthetic peptide (CVLSQCDS(P)PYVTKYY; Biosynthesis) and purified using the SulfoLink Kit (Thermo). His-RhoA protein (RH01) was from Cytoskeleton and the Mst3 recombinant protein (PV3650) was from Invitrogen.
Experimental animals and in utero electroporation.
Cdk5-knock-out mice were provided by A.B. Kulkarni (National Institutes of Health). In utero electroporation was performed as described previously (Ip et al., 2012). ICR mice of either sex were used for in utero electroporation at indicated ages. For knockdown experiments, the mice were coinjected with pCAG2IG expressing GFP and pSUPER plasmid, scramble shRNA, or Mst3 shRNA in a 1:2 ratio on embryonic day 14 (E14) and brains were collected on E17, E18, or postnatal day 2 (P2) or P5. For rescue experiments, Mst3 shRNA was mixed with pCAG2IG expressing different rescue constructs (in a 2:3 ratio) and coinjected into the mice. For the RhoA double-knockdown experiment, GFP, Mst3 shRNA, and RhoA shRNA were mixed (2:2:0.5) and coinjected into the mouse brains.
Primary neuron cultures, transfection, and treatment.
Primary cortical neuron cultures were prepared from E18 rats and maintained as described previously (Fu et al., 2007). Nucleofector (Amaxa Biosystems) or Lipofectamine 2000 (Invitrogen) was used to deliver plasmids into cultured neurons. For pharmacological treatment, cortical neurons (1 × 107 cells per plate) were treated with 200 nm 6-bromoindirubin-3′-acetoxime, 10 μm roscovitine, 10 μm SP600125, 100 nm UO126 (Calbiochem), or 20 μm H89 (Sigma) for 1 h. Neurons at 4 d in vitro (DIV) were subjected to immunoprecipitation and Western blot analysis.
Cell cultures, transfection, protein extraction, and immunoprecipitation.
HEK293T cells were cultured in DMEM with 10% heat-inactivated fetal bovine serum, 50 units/ml penicillin, and 100 μg/ml streptomycin at 37°C in a 5% CO2 humidified atmosphere. Lipofectamine and Plus transfection reagents were from Invitrogen. Protein extraction and immunoprecipitation were performed as described previously (Fang et al., 2011). In brief, 1–2 mg of protein lysate was incubated with 2 μg of antibody at 4°C for 3 h with end-to-end rotation; the antibody was subsequently pulled down by protein G-Sepharose beads (GE Healthcare) for 1 h.
In vitro phosphorylation assay.
The in vitro phosphorylation assay was performed as described previously (Fu et al., 2007). In brief, recombinant kinases or immunoprecipitated kinases were incubated in kinase buffer [20 mm MOPS; 3-(N-morpholino)-propanesulfonic acid, pH 7.4, 15 mm MgCl2, 20 mm MnCl2, and 100 μm ATP] containing 1–2 μCi [γ-32P] ATP with 8 μg of histone H1 protein or immunoprecipitated proteins at 30°C for 30 min with shaking at 900 rpm. The phosphorylated protein was subsequently separated using SDS-PAGE and visualized by autoradiography. In particular, to examine the activity of Mst3 in the mouse brain, equal amounts of brain lysates at the indicated stages were immunoprecipitated with anti-Mst3 antibody, followed by incubation with recombinant histone H1 protein in kinase reaction buffer for in vitro phosphorylation assay. For identification of the Cdk5 phosphorylation site on Mst3, HEK293T cells expressing WT or one of three phospho-deficient mutants of Mst3 at putative Cdk5 phosphorylation sites were pulled down by anti-FLAG antibody, incubated with or without Cdk5/p35 complex, and subjected to an in vitro kinase assay.
Filamentous/globular actin ratio and growth cone collapse assay.
For determining the ratio of filamentous actin (F-actin) and globular actin (G-actin) (Meller et al., 2008), cortical neurons were transfected with pSUPER or Mst3 shRNA by nucleofection at 0 DIV. Four days after transfection, the neurons were homogenized in F-actin stabilization buffer and ultracentrifuged to separate F-actin (pellet fraction) and G-actin (supernatant fraction) pools. The two fractions were separated by SDS-PAGE and actin was quantified by Western blotting using anti-actin antibody. Growth cone collapse was performed as described previously (Shi et al., 2007). In brief, cultured cortical neurons from E14 ICR mice were transfected with GFP and pSUPER or Mst3 shRNA by nucleofection. At 4 h after transfection, the neurons were treated with 10 μm ROCK inhibitor (RI: Y27632) for 20 h and subsequently fixed. The fixed neurons were stained with anti-GFP, anti-Tuj1 antibodies, and fluorescently labeled phalloidin.
Electrophysiology.
In utero electroporation was performed at E14. Electrophysiological recordings were performed on cortical slices prepared at P15–P20. Briefly, mice of either sex were anesthetized and brains were rapidly dissected in ice-chilled artificial CSF (ACSF) containing the following (in mm): 126 NaCl, 3.5 KCl, 2 CaCl2, 1.3 MgCl2, 25 NaHCO3, 1.2 NaHPO4, and 10 glucose at 95% O2, 5% CO2, pH 7.4. Coronal brain slices (250 μm) prepared using a vibratome were kept in oxygenated ACSF at room temperature for 1 h before recordings. The slices were transferred to a chamber mounted onto a fixed-stage upright microscope (BX51WI; Olympus) equipped with an epifluorescent attachment and superfused with ACSF at physiological temperature. Conventional whole-cell recordings were made from target cortical neurons, including GFP+ neurons, aided by differential interference contrast optics. Microelectrodes had a resistance of 6–8 MΩ. All electrophysiological data were digitized by the Digidata pClamp system and analyzed with MiniAnalysis (Synaptosoft).
Quantification and statistical analysis.
For the morphological analysis of dendrite length at P5, z-series confocal images of cortical sections were collected at 1 μm steps with 10 optical sections in Fluoview (Olympus). In the stacked images, an apical dendrite was traced manually and the dendrite length was measured using ImageJ software. Quantification results are showed as mean ± SEM. Statistical significance was determined using unpaired Student's t test.
Results
Mst3 regulates radial migration and neuronal morphology in the mouse cerebral cortex
To investigate the role of Mst3 in the brain, we first examined its expression profile in the mouse brain during development. Although Mst3 protein was prominently expressed in the cerebral cortex from early embryonic stages through adulthood (Fig. 1A), its kinase activity was particularly robust from E14–E17, when extensive active neuronal migration occurs (Fig. 1B). Three Mst3 shRNAs were generated to study the function of Mst3; one of them in particular efficiently depleted the expression of endogenous Mst3 in cultured cortical neurons, whereas scrambled shRNA did not reduce Mst3 expression levels (Fig. 1C). In utero electroporation of this Mst3 shRNA at E14 resulted in migratory defects that were rescued by the reexpression of a shRNA-resistant Mst3-expressing construct. Although most (∼50%) of the neurons expressing pSUPER vector or scrambled shRNA migrated into the cortical plate at E17, the majority (∼60%) of Mst3-knockdown neurons remained stacked in the intermediate zone (Fig. 1D,E). To determine whether this migratory defect is associated with abnormal neuronal morphology, we examined leading process formation in knockdown neurons at the intermediate zone. Notably, whereas control neurons formed a single leading process pointing toward the pial surface, most Mst3-knockdown neurons in the intermediate zone exhibited disrupted neuronal polarity with no extended process (Fig. 1F,G).
Mst3 knockdown leads to altered neuronal positioning and cortical activity
The accumulation of Mst3-knockdown neurons in the intermediate zone may be due to either the inhibition or delay of neuronal migration. Intriguingly, whereas only some of the Mst3-knockdown neurons migrated up into the cortical plate at E18 (Fig. 1H,I), most of them ultimately migrated to the upper layers of the cortical plate at P2 and P5 (Fig. 2A,B). These results suggest that Mst3 knockdown delayed rather than arrested neuronal migration. However, although the neurons were able to migrate up into the cortical plate, the final positioning of Mst3-knockdown neurons was abnormal. Compared with the control neurons, the cell bodies of which were deposited in the upper part of layers II–IV (Cutl1+), the knockdown neurons were located in a relatively deeper part of layers II–IV. In addition to the positional defect, Mst3 knockdown also led to abnormal neuronal morphology and impaired dendrite development. The Mst3-knockdown neurons exhibited significantly shorter apical dendrites, altered orientation, and distorted cell bodies compared with the control neurons (Fig. 2C–E). Collectively, these findings suggest that Mst3 is important for the precise positioning and morphogenesis of neurons in the cerebral cortex.
To determine whether the migration defect and morphological abnormalities caused by Mst3 knockdown leads to altered cortical excitability, we performed whole-cell recordings of the pyramidal neurons in upper layers II–III. Intriguingly, the frequency of spontaneous EPSCs was significantly lower in the Mst3-knockdown pyramidal neurons compared with the control neurons (Fig. 2F–H). Therefore, Mst3 knockdown impaired the excitatory synaptic transmission of the pyramidal neurons in the upper cortical layers.
Mst3 kinase activity is important for neuronal migration
The prominent kinase activity of Mst3 from E14 to E17 prompted us to investigate whether Mst3-regulated radial migration is dependent on its kinase activity. To generate a kinase-dead Mst3 mutant (Mst3-KD), the ATP-binding site (Lys53) was substituted with arginine (Fig. 3A). Loss of kinase activity of Mst3-KD was confirmed by its inability to phosphorylate histone H1 in an in vitro phosphorylation assay (Fig. 3A). We subsequently generated shRNA-resistant forms of Mst3 WT and Mst3-KD (Fig. 3B). Notably, reexpression of Mst3-WT but not KD fully restored normal migration and neuronal morphology in Mst3-knockdown cortices (Fig. 3C–F). Nonetheless, overexpression of Mst3-KD alone does not affect the neuronal migration (Fig. 3G,H). Given the relatively high Mst3 activity in mouse brains at embryonic stages (E14–E17), the expression of Mst3-KD may be ineffective to compete with endogenous Mst3 and inhibit its activity and functions in migrating neurons. This suggests that Mst3 kinase activity is important for radial migration. Therefore, understanding the regulation of Mst3 kinase activity may provide insights into the molecular mechanisms of neuronal migration.
Prediction of potential phosphorylation sites by a computational method (Obenauer et al., 2003) identified various potential phosphorylation sites in Mst3 recognized by various kinases, including GSK3, Cdk5, JNK, MEK, and PKA. After treating cortical neurons with different kinase inhibitors, only the Cdk5 inhibitor dramatically reduced autophosphorylation (p-Mst3; Fig. 4A) and thus Mst3 activity. Furthermore, Mst3 kinase activity was reduced in cdk5-knock-out mouse brains (Fig. 4B). These results suggest that Cdk5 regulates Mst3 kinase activity. Indeed, we found that active Cdk5 phosphorylated Mst3 (Fig. 4C) and endogenous Mst3 phosphorylation were significantly attenuated in cdk5-knock-out brains (Fig. 4D). These findings suggest that Mst3 is phosphorylated by Cdk5 in vivo. Three sites on Mst3, Ser4, Ser79, and Ser369, were identified as potential Cdk5 phosphorylation sites (Fig. 4E). We subsequently generated phosphorylation-deficient mutants by mutating the Ser4, Ser79, or Ser369 site to alanine. Mutation of Ser79, but not the other two sites, abolished Cdk5-dependent Mst3 phosphorylation (Fig. 4F), indicating that Ser79 is the major Cdk5 phosphorylation site. To further confirm this result, we raised a phosphospecific Mst3 antibody against phosphorylated Mst3 at Ser79 (p-S79 Mst3). The p-S79 Mst3 antibody specifically recognized phosphorylated Mst3 protein after Cdk5 phosphorylation (Fig. 4G,H). Consistently, the Ser79 phosphorylation of Mst3 was dramatically reduced in cdk5-knock-out mouse brains (Fig. 4I), corroborating Mst3 as an in vivo substrate of Cdk5.
To investigate whether Cdk5-dependent Mst3 phosphorylation is important for radial migration, we reexpressed Mst3 WT and S79A mutant in Mst3-knockdown cortices. Mst3 WT fully restored normal migration, whereas reexpression of the Mst3 S79A mutant failed to rescue the migration defects and altered the neuronal morphology of Mst3-knockdown neurons (Fig. 4J–L). Similar to Mst3 KD, the S79A mutant alone has no effect on neuronal migration (Figs. 3G,H, 4M,N). Therefore, Cdk5-dependent phosphorylation of Mst3 at Ser79 is important for Mst3 kinase activity and thus proper neuronal migration.
Mst3 regulates actin dynamics and RhoA activity
Next, we investigated how Mst3 and its kinase activity regulate neuronal migration. Mst3 was enriched in growing neurite tips of neurons, where it colocalized with F-actin staining (Fig. 5A). Notably, Mst3 knockdown in neurons substantially increased the ratio of F-actin to G-actin (Fig. 5B,C), suggesting that Mst3 regulates the polymerization status of the actin cytoskeleton. To determine how Mst3 regulates the morphology of migrating neurons, Mst3 shRNA was introduced into cortical neurons cultured from E14 embryos and growth cone morphology was examined 1 d later. Mst3 knockdown resulted in remarkably increased growth cone collapse. Interestingly, such defects were largely rescued by Y27632 (RI), a pharmacological inhibitor of RhoA signaling (Fig. 5D,E). Together with the interaction between Mst3 and RhoA, but not the other two important Rho-GTPases (i.e., Rac1 and Cdc42, data not shown; Fig. 5F), these results suggest that RhoA is a downstream signaling molecule of Mst3. Importantly, Mst3 knockdown led to RhoA activation, whereas Mst3 overexpression inhibited RhoA activity in cultured neurons (Fig. 5G,H). These findings suggest that Mst3 plays important roles in the regulation of actin dynamics and RhoA inhibition.
Negative regulation of RhoA by Mst3 phosphorylation is important for radial migration
Mst3 phosphorylated RhoA at serine residue(s) (Fig. 6A). Mass spectrometry and 3D structural analysis identified two potential Mst3 phosphorylation sites on RhoA: Ser26 and Ser160 (Ihara et al., 1998). An in vitro phosphorylation assay confirmed that Mst3 phosphorylates RhoA at Ser26 (Fig. 6B). Importantly, blocking the Mst3-dependent phosphorylation of RhoA at Ser26 enhanced its GTPase activity (Fig. 6C), suggesting that Ser26 phosphorylation at RhoA negatively regulates its activity. To determine whether Mst3-dependent RhoA phosphorylation is important for neuronal migration, we overexpressed RhoA WT or S26A in E14 mouse brains. Overexpression of RhoA WT significantly inhibited neuronal migration and disrupt normal neuronal polarization, whereas overexpression of RhoA S26A, which has higher RhoA activity, led to a more severe defect in neuronal migration and polarization (Fig. 6D,E). Similar to the effect observed in Mst3-knockdown cortices, RhoA WT or S26A-expressing migrating neurons exhibited an abnormal morphology with no neuronal polarity or processes (Fig. 6F). Moreover, inhibition of RhoA activity by suppressing the RhoA protein expression restored migration and neuronal morphology defects in Mst3-knockdown neurons (Fig. 6G–I). Together, these findings reveal a critical role of Mst3 in the regulation of radial migration via the inhibition of RhoA activity.
Discussion
In this study, we report that Mst3 is a critical regulator of corticogenesis through controlling the final positioning and morphogenesis of migrating neurons. Our results show that the activity of Mst3 is regulated by Cdk5 phosphorylation and activated Mst3 in turn inhibits RhoA activity, thus modulating the actin cytoskeleton of migrating neurons. Together with previous reports on the regulation of RhoA-dependent actin dynamics by Cdk5 (Ye et al., 2012), these findings suggest that Cdk5 regulates neuronal migration through integration of various kinase signaling pathways, including that mediated by Mst3.
It has become increasingly clear that subtle alterations in neuronal morphogenesis may underlie the pathophysiology of various neurological disorders. For example, aberrant dendritic patterning and altered spine morphology have been implicated in psychiatric disorders with delayed onset, such as autism spectrum disorders (de Anda et al., 2012). Therefore, it is important to understand how pyramidal neurons acquire their dendritic morphology after terminal translocation. Here, we show that Mst3 knockdown retards the migration of cortical neurons, resulting in mispositioning of Mst3-deficient neurons. Importantly, the Mst3-depleted neurons exhibit aberrant dendrite morphology. These findings collectively suggest that Mst3 plays multiple roles in different cellular events during corticogenesis, including neuronal migration and dendrite development, in a cell-autonomous manner. In support of this notion, we found that dendrite extension and arborization was inhibited in cultured neurons after Mst3 knockdown (data not shown). However, given that Mst3 kinase activity is robust in embryonic brains and becomes much lower in postnatal brains, it is likely that the defect in synaptic transmission in Mst3-depleted cortex is due to the mispositioning of neurons rather than being the direct result of Mst3 knockdown. Further studies on how Mst3 knockdown may affect circuit connection among layer II–IV pyramidal neurons may provide insights on the pathogenesis of psychiatric disorders.
Inhibition of RhoA activity is required for F-actin depolymerization at the tip of leading process during multipolar-to-bipolar transition (Pacary et al., 2011). Nonetheless, how RhoA activity is regulated in migrating neurons remains largely unknown. We found that silencing Mst3 in cultured neurons results in enhanced RhoA activity and increased F-actin content, whereas depletion of the RhoA protein expression rescues the radial migration defect of the Mst3-knockdown neurons. These results support the hypothesis that the inactivation of RhoA activity by Mst3 causes F-actin depolymerization, ultimately allowing the polarization of neurons and their exit from the intermediate zone, which is an initial critical step in radial migration. In addition to Mst3, RhoA activity is regulated through phosphorylation by other kinases, such as cAMP- and cGMP-dependent kinases (i.e., PKA and PKG; Ellerbroek et al., 2003). In addition, RhoA-dependent actin reorganization is modulated by various signaling regulators, including an atypical Rho-GTPase Rnd3 and p27 (Nguyen et al., 2006; Pacary et al., 2011). Interestingly, the activity and functions of these RhoA regulators can be modulated by phosphorylation. In particular, p27 and Mst3 are substrates of Cdk5 (Kawauchi et al., 2006, and present results). Cdk5-mediated phosphorylation of p27 or Mst3 leads to the inactivation of RhoA, suggesting that Cdk5 acts as a functional switch through phosphorylation of multiple substrates during the initiation of radial migration.
How does Mst3 regulate RhoA? The activity of RhoA is tightly regulated by three classes of regulatory proteins: guanine dissociation inhibitors, guanine nucleotide exchange factors (GEFs), and GTPase-activating proteins (Takai et al., 2001). In addition, the phosphorylation of Ser188 in the C-terminal region regulates RhoA activity by protecting the protein from ubiquitin/proteasome-mediated degradation (Rolli-Derkinderen et al., 2005). Our results show that the phosphorylation of RhoA at Ser26 by Mst3 negatively regulates its RhoA GTPase activity. Interestingly, Ser26 is adjacent to Lys27, which is a residue in the Switch I region of RhoA that is critical for the specific interactions between RhoA and its GEFs (Ihara et al., 1998). Therefore, it would be interesting to determine whether this specific phosphorylation of RhoA alters its association with GEFs during neuronal migration. Furthermore, given that Ser26 is conserved among different Rho members, phosphorylation at this specific site may serve as a general negative regulatory mechanism for members of this small GTPase family (Madaule and Axel, 1985; Chardin et al., 1988; Fagan et al., 1994).
Although phosphorylation of Mst3 at Ser79 results in the increase of its kinase activity, its underlying molecular mechanism remains to be elucidated. Our preliminary results show that suppression of Ser79 phosphorylation apparently does not affect the substrate-binding or ATP-binding abilities of Mst3 (data not shown). Structural analysis reveals that Ser79 is located near the C-helix within the kinase domain, which is an element important for the conformational change of the kinase domain during catalysis (Ko et al., 2010). Therefore, Ser79 phosphorylation of Mst3 may be important for the catalytic activity of the kinase. Although Ser79 is conserved among different members of the Mst family (Record et al., 2010), it would be intriguing to determine whether this serine phosphorylation regulates the kinase activities of other Mst members and thus their functions during neurodevelopment. It will also be important to determine whether the status of Ser79 phosphorylation of Mst3 is altered in animal models of neuropsychiatric disorders. Further characterization of this serine phosphorylation will provide important insights into the pathogenesis of neurodevelopmental disorders such as autism spectrum disorders.
In conclusion, our results demonstrate that the Cdk5–Mst3 pathway controls the precise organization of the actin cytoskeleton during initiation of neuronal migration. The convergence of various kinase signaling pathways toward cytoskeletal organization and remodeling during neuronal migration indicates that the patterning and lamination of the cerebral cortex is under precise regulation by a complex and coordinated system of biochemical signaling mechanisms.
Footnotes
The authors declare no competing financial interests.
This work was supported by the Research Grants Council of Hong Kong SAR (Grants HKUST660810, HKUST660711, HKUST661111, HKUST661212, and HKUST660213), the National Key Basic Research Program of China (Grant 2013CB530900), the Theme-based Research Scheme of the University Grants Committee (Grant T13-607/12-R), the Innovation and Technology Fund for State Key Laboratory (Grant ITCPT/17-9), the Shenzhen Peacock Plan, and the S.H. Ho Foundation. We thank Cara Kwong, Busma Butt, Wanting Zhong, and William Chau for excellent technical assistance; other members of the Ip laboratory, in particular Dr. Lei Shi, for helpful discussions; and Xuemin Zhang and Weihua Li (National Center of Biomedical Analysis) for providing help in mass spectrometry analysis.
- Correspondence should be addressed to either Amy K.Y. Fu or Nancy Y. Ip, Division of Life Science, The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong, China, boip{at}ust.hk or boamy{at}ust.hk