Ezrin Mediates Neuritogenesis via Down-Regulation of RhoA Activity in Cultured Cortical Neurons

Neuronal morphogenesis is implicated in neuronal function and development with rearrangement of cytoskeletal organization. Ezrin, a member of Ezrin/Radixin/Moesin (ERM) proteins links between membrane proteins and actin cytoskeleton, and contributes to maintenance of cellular function and morphology. In cultured hippocampal neurons, suppression of both radixin and moesin showed deficits in growth cone morphology and neurite extensions. Down-regulation of ezrin using siRNA caused impairment of netrin-1-induced axon outgrowth in cultured cortical neurons. However, roles of ezrin in the neuronal morphogenesis of the cultured neurons have been poorly understood. In this report, we performed detailed studies on the roles of ezrin in the cultured cortical neurons prepared from the ezrin knockdown (Vil2kd/kd) mice embryo that showed a very small amount of ezrin expression compared with the wild-type (Vil2+/+) neurons. Ezrin was mainly expressed in cell body in the cultured cortical neurons. We demonstrated that the cultured cortical neurons prepared from the Vil2kd/kd mice embryo exhibited impairment of neuritogenesis. Moreover, we observed increased RhoA activity and phosphorylation of myosin light chain 2 (MLC2), as a downstream effector of RhoA in the Vil2kd/kd neurons. In addition, inhibition of Rho kinase and myosin II rescued the impairment of neuritogenesis in the Vil2kd/kd neurons. These data altogether suggest a novel role of ezrin in the neuritogenesis of the cultured cortical neurons through down-regulation of RhoA activity.


Introduction
Establishment of neural circuits in the central nerve system requires generation and development of multiple dendrites and single axon. Cultured neurons showing a sequence of morphological changes have been well studied for neuronal morphogenesis [1,2]. In particular, neuritogenesis that is the first step in neuronal morphogenesis is driven by exocytic and cytoskeletal machinery [3]. Several neurites extended from a symmetrical cell body become an axon or dendrites, and subsequently, neurons establish synaptic connections and networks. Small GTPases, RhoA, Rac1 and Cdc42 modulate the neuronal morphogenesis through regulating cytoskeletal dynamics in different pathways [4]. Rac1 and Cdc42 promote neurite outgrowth through phosphorylation of p21-activated kinase (PAK) family of serine/threonine kinases [5]. In contrast, RhoA and its downstream effector Rho kinase mediate neurite retractions [6]. Myosin II activity is determined by phosphorylation of myosin light chains (MLCs) and mediated by RhoA/Rho kinase pathway. Activated myosin II generates formation of cortical actin filaments and leads to inhibition of neuritogenesis [7].
Ezrin, radixin and moesin (ERM) proteins are membranecytoskeleton linkers and regulate Rho activity through interaction with Rho guanine nucleotide dissociation inhibitor or Rho GTPase-activating protein [8][9][10]. In the cultured hippocampal neurons, expression of the ERM proteins was first detected by mouse monoclonal 13H9 antibody that recognized all members of the ERM proteins [11]. Among the ERM proteins, radixin and moesin were enriched in growth cone structure and associated with neurite extensions in the cultured hippocampal neurons [12]. Ezrin was associated with axon outgrowth induced by netrin-1 stimulation [13], however, expression of ezrin was mainly detected in cell body [12]. Therefore, the role of ezrin in the neuronal morphogenesis has remained unclear.
In the present study, to examine the functions of ezrin in the neuronal morphogenesis, we newly used cultured cortical neurons prepared from ezrin knockdown (Vil2 kd/kd ) mice in which ezrin expression levels were decreased to less than 5% compared with the wild-type (Vil2 +/+ ) mice [14]. The Vil2 kd/kd mice showed achlorhydria due to impairment of membrane fusion between intracellular gastric vesicles and apical membrane in gastric parietal cells. In this report, the cultured cortical neurons prepared from the Vil2 kd/kd mouse embryo showed reduction in number of neurites compared with the Vil2 +/+ neurons whereas length of neurites and axon was not changed. We studied whether RhoA, Rac1 and Cdc42 activities were modulated in the cultured cortical neurons prepared from the Vil2 kd/kd mouse embryo. Treatment of Rho kinase inhibitor Y-27632 or myosin II inhibitor blebbistatin was reported to promote initiation of axon outgrowth and neuritogenesis, respectively [15][16][17]. We studied effects of these inhibitors on the neuritogenesis in the Vil2 kd/kd neurons. Our data suggest that ezrin is a key player of the neuritogenesis in the cultured cortical neurons through down-regulation of the RhoA activity.

Mice
Vil2 kd/kd mice were prepared as described previously [14]. All works with animals were performed with approval from the Animal Ethics Committee of Ritsumeikan University. Neurons at the stages 1, 2 and 3 were stained with an anti-ezrin antibody, rhodamine phalloidin, and an anti-a-tubulin antibody, respectively. In the bottom lane, neurons were triple stained with an anti-ezrin antibody (green), rhodamine phalloidin (red) and an anti-a-tubulin antibody (blue). Scale bars, 50 mm. doi:10.1371/journal.pone.0105435.g001
Immunofluorescence 1610 4 cells were fixed with 4% paraformaldehyde and 4% sucrose in PBS for 10 min at 4uC. Fixed cells were permeabilized with 0.1% Triton X-100 in PBS for 10 min at room temperature. Cells were treated with 1% BSA in PBS for 30 min at room temperature, and incubated with primary antibodies overnight at 4uC, followed by the treatment with secondary antibodies for 45 min at room temperature. The primary antibodies used in this study were a mouse anti-a-tubulin antibody (1:1000, DM1A, Abcam), a rabbit anti-ezrin antibody (#3145, 1:100, Cell Signaling Technology), a rabbit anti-neuronal class III b-tubulin antibody (1:1000, TUJ1, Covance). The secondary antibodies used in this study were a fluorescein isothiocyanate (FITC)-conjugated anti-rabbit IgG (Jackson ImmunoResearch) and an Alexa Fluor 633-conjugated anti-mouse IgG (Invitrogen). For filament actin staining, rhodamine phalloidin (Invitrogen) was added to secondary antibody. Fluorescence images were acquired with a confocal laser scanning microscope (FV-1000D, FV-10i, Olympus).

Morphological analysis
To categorize each stage of cells, cells were defined by the length of the longest neurite: stage 1 (nonpolar), ,10 mm; stage 2 (multipolar), ,40 mm; stage 3 (axon-forming), .40 mm as reported previously [19]. Axon was defined by the following criteria: a process more than twice longer than other processes [20]. The length and number of neurites and length of axon in each neuron were measured by manual tracing using ImageJ software with NeuronJ plugin. Neurons were identified by immunofluorescence with an anti-neuronal class III b-tubulin antibody.

Statistical analysis
Results are expressed as mean 6 SE. The significance of differences was determined by Student's t test.

Ezrin knockdown showed impairment of neuritogenesis
We first confirmed the distribution of ezrin in the wild-type cultured neurons by immunofluorescence ( Fig. 1). Ezrin was partly overlapped with microtubules and mainly expressed in the cell body. In actin-rich region such as the lamellipodia or growth cone, only a little colocalization of ezrin with actin filaments was detected as reported in the sensory and hippocampal neurons [12,21].
To examine the role of ezrin in the cultured cortical neurons, we performed loss-of-function studies. Previously, Antoine-Bertrand et al. [13] performed ezrin-specific siRNA treatment of cultured cortical neurons, which resulted in partial reduction of ezrin expression to 55%, and reported that the treatment impaired axon-outgrowth induced by netrin-1. Here, we newly studied the roles of ezrin in the cultured cortical neurons prepared from the Vil2 kd/kd mouse embryo. A very small signal of ezrin was detected by immunofluorescence in the Vil2 kd/kd neurons (Fig. S1A). The same antibody recognized a single band of ezrin with a molecular mass of 80 kDa by immunoblotting in the Vil2 +/+ neurons. Conversely, the band was not detected in the Vil2 kd/kd neurons (Fig. S1B). The band was detected in the Vil2 +/+ neuron extracts (10 mg, 1 mg and 0.5 mg) whereas it was not detected even in 10 mg of the Vil2 kd/kd neuron extracts, indicating that amount of ezrin expressed in the Vil2 kd/kd neurons is less than 5% compared with the Vil2 +/+ neurons, which is consistent with the previous report [14]. In addition, the expression of radixin or moesin was not upregulated in the Vil2 kd/kd neurons in a compensatory manner (data not shown).
To perform the morphological analysis, we stained the cultured cortical neurons with cytoskeletal markers, anti-neuronal class III b-tubulin antibody, and rhodamine phalloidin ( Fig. 2A,B). First, we classified and counted several stages of neurons at 48 h after plating (2 DIV). The population of stage 1 neurons was significantly increased in the Vil2 kd/kd neurons compared with the Vil2 +/+ neurons (Vil2 +/+ : 8.861.4%, Vil2 kd/kd : 17.362.3%, Fig. 2C). Conversely, the population of stage 3 neurons was significantly decreased in the Vil2 kd/kd neurons (Vil2 +/+ : 55.864.2%, Vil2 kd/kd : 43.164.1%). Moreover, the stage 3 Vil2 kd/kd neurons exhibited decreases in the number of neurites (Vil2 +/+ : 2.760.2, Vil2 kd/kd : 1.560.2, Fig. 2D). On the other hand, no significant differences were detected in the length of both neurites and axon (Fig. 2E,F). We confirmed this morphological deficit in stage 3 neurons cultured for longer period (Fig. S2A,B). Although length of both neurites and axon was not changed (Fig.  S2D,E), the number of neurites was decreased in Vil2 kd/kd neurons (Vil2 +/+ : 4.260.6, Vil2 kd/kd : 1.460.2, Fig. S2C). There were no significant differences in the number and length of neurites between the neurons prepared from radixin knockout mice and wild-type neurons (unpublished data). These results suggest that ezrin is a key player of neuritogenesis.

Increase of RhoA activity in Vil2 kd/kd neurons
RhoA is a member of Rho family proteins that promote cell signaling pathway and cytoskeletal organization in cultured neurons [4,22]. It was shown that RhoA activated the ERM proteins by phosphorylating their C-terminal threonine residues in 3T3 cells [23]. Conversely, ezrin was shown to be a negative regulator for RhoA because a dominant-negative form of ezrin (a mutant with its actin-binding domain being deleted) or ezrin knockout increased RhoA activity in non-neuronal cells [24,25]. Therefore, we examined whether ezrin knockdown increases RhoA activity in the cultured cortical neurons (Fig. 3). Total amount of RhoA was not changed between the Vil2 +/+ and Vil2 kd/ kd neurons whereas GTP-bound RhoA in the Vil2 kd/kd neurons was increased more than three-folds compared with the Vil2 +/+ neurons (Fig. 3A,D). In contrast to RhoA, other Rho family members, Rac1 and Cdc42 were not affected by ezrin knockdown (Fig. 3B,E and C,F). To confirm the involvement of RhoA activation in the neuritogenesis, we observed the phosphorylation of downstream effector in the Vil2 kd/kd neurons. The phosphorylation of MLC2, which negatively regulates the actin organization, was enhanced in the Vil2 kd/kd neurons (Fig. 4A,B). However, the immunoblotting with a polyclonal antibody that recognized all members of ERM proteins confirmed that there were no differences between the Vil2 +/+ and Vil2 kd/kd neurons in the phosphorylation of radixin and moesin (Fig. 4C,D). Next, we tested whether phosphorylation of MLC2 is affected by Y-27632. Phosphorylated MLC2 was decreased in Y-27632-treated Vil2 +/+ neurons compared with DMSO-treated Vil2 +/+ neurons (Fig. 5A,B). Similar decrease in the phosphorylated MLC2 was also observed in Y-27632-treated Vil2 kd/kd neurons, indicating that MLC2 is a downstream effector of RhoA/Rho kinase pathway. In contrast, phosphorylated ezrin, radixin and moesin were not affected by Y-27632 (Fig. 5A,C-E).

Inhibition of Rho kinase and myosin II rescued neuritogenesis
We then attempted to determine whether inhibition of RhoA activation rescues neuritogenesis in the Vil2 kd/kd neurons (Fig. 6A-D). The Rho kinase was inhibited by a specific inhibitor Y-27632, resulting in the alternation of neurite outgrowth [15,26,27]. Our results showed that Y-27632 was able to rescue the effect of ezrin knockdown on the neuritogenesis. In both the Vil2 +/+ and Vil2 kd/kd neurons, the number of neurites was increased by the treatment of Y-27632 in comparison to vehicle-treated cells (Fig. 6E). The number of neurites was similar between Y-27632-treated Vil2 +/+ neurons and Y-27632-treated Vil2 kd/kd neurons (Y-27632-treated Vil2 +/+ : 4.060.3, Y-27632-treated Vil2 kd/kd : 3.46.3). Similar increase in the length of neurites and axon was also observed in the Vil2 +/+ and Vil2 kd/kd neurons (Fig. 6F,G). We next examined inhibition of myosin II by addition of 50 mM blebbistatin (Fig. 7A,B). Similar to Y-27632, treatment of blebbistatin enhanced number of neurites and rescued impairment of neuritogenesis (blebbistatin-treated Vil2 +/+ : 7.760.8, blebbistatintreated Vil2 kd/kd : 8.861.0, Fig. 7C). There were no significant differences in the length of neurites and axon between DMSOtreated neurons and blebbistatin-treated neurons (Fig. 7D,E). These results suggest that ezrin is essential in the neuritogenesis through down-regulation of RhoA activities (Fig. 8).

Discussion
ERM proteins that link plasma membrane proteins and the actin cytoskeleton are expressed in various cultured neurons; hippocampal, cortical and dorsal root ganglion neurons [11,13,28]. Radixin and moesin, not ezrin are enriched in the actin-rich structure of growth cone such as radial striations, lamellipodial veils and filopodial extensions [29]. Radixin was abundant in the leading edge of growth cone, and the microscale chromophore-assisted laser inactivation (micro-CALI) of radixin in growth cones caused a 30% reduction of lamellipodial area in chick sympathetic neurons [30]. Moesin was phosphorylated by the treatment with glutamate in H19-7/IGF-IR cells [31]. The siRNA inhibition of moesin and inactivation of the phosphorylated ERM proteins caused reduction of active synaptic boutons induced by the glutamate treatment in the cultured hippocampal neurons [32]. In addition, in the hippocampal neurons, suppressions of both radixin and moesin, neither of ezrin-moesin nor ezrin-radixin, by the antisense oligonucleotide treatment displayed the impairment of growth cone morphology and neurite extensions [12], suggesting that radixin and moesin promote the actin organization in neuronal morphogenesis. However, among the ERM proteins, the roles of ezrin in the neuronal morphogenesis have been poorly understood.
Previously, Antoine-Bertrand et al. [13] demonstrated that the phosphorylated ERM proteins formed a complex with the receptor Deleted in Colorectal Cancer (DCC), and were associated with netrin-1-induced axon outgrowth. The netrin-1 phosphorylated the ERM proteins especially ezrin, and enhanced accumulation of the phosphorylated ERM proteins in the growth cone. The DCC-mediated neurite outgrowth in N1E-115 cells was inhibited by the expression of dominant-negative form of ezrin, in which the actin-binding domain was deleted. Moreover, ezrin siRNA treatment, which induced 55% down-regulation of its expression, impaired axon outgrowth on the netrin-1 stimulation, suggesting that ezrin is required for the axon outgrowth in the cultured cortical neurons. In our present report, we newly demonstrated that ezrin was involved in the neuritogenesis using the cultured cortical neurons prepared from the Vil2 kd/kd mice embryo in which ezrin expression was down-regulated less than 5% compared with the Vil2 +/+ neurons. The Vil2 kd/kd neurons showed decrease in the number of neurites (Fig. 2D) with their lengths of neurites and axon being unaltered (Fig. 2E,F). In the Vil2 kd/kd neurons, the retardation of stage progression was also observed at the stages 1 and 3 (Fig. 2C). We also showed impairment of neuritogenesis in Vil2 kd/kd neurons beyond 2 DIV ( Fig. S2C), suggesting that ezrin knockdown caused impairment, not delay, of neuritogenesis. These observations suggest that ezrin is necessary for sprouting, not extension in neurites and axon.
Previous reports suggested the possible functional redundancy of ezrin and among the ERM proteins at the cellular level [33]. However, no compensatory up-regulation of radixin and moesin was observed in the Vil2 kd/kd neurons. Although we could not use ezrin knockout (Vil2 2/2 ) mice in the present study, the expression level of ezrin in the Vil2 kd/kd neurons was less than 5% compared with the Vil2 +/+ neurons. Therefore, it can be expected that functional role of ezrin was observed more clearly in our present loss-of-function study overcoming redundancy.
Rho family small GTPases, RhoA, Rac1 and Cdc42 coordinate actin filaments and microtubule dynamics in the neuronal morphology [4]. The inactivation of RhoA/Rho kinase pathway facilitated stability of actin filaments resulting in the initiation of neurite sprouting in the cultured hippocampal neurons and 1C11 neuronal cell line [26]. Myosin II is a downstream effector of RhoA/Rho kinase pathway and important for the neuritogenesis. It was required for the maintenance of neuronal sphere with cortical actin filaments in the early stage of neurons [7]. In fact, both length and number of minor processes were reduced by the increase of myosin II activity in the cultured cortical neurons transfected with constitutively active RhoA. In addition, the phosphorylation of MLC2 following RhoA/Rho kinase activation regulated the activity of myosin II [34,35]. Our present findings demonstrate that the impairment of neuritogenesis was concomitant with the abnormal cytoskeletal organizations caused by activation of RhoA/Rho kinase/MLC2 pathway in the Vil2 kd/kd neurons.
ERM proteins were thought to be regulators of Rho activity through interaction with Rho guanine nucleotide dissociation inhibitors or Rho GTPase-activating proteins [8,10]. In nonneuronal cells, it was reported that down-regulation of ezrin induced activation of RhoA. Speck et al. [24] reported that activation of RhoA was observed in LLC-PK1 epithelial cells expressing a dominant-negative form of ezrin, in which the actinbinding domain was deleted. Casaletto et al. [25] demonstrated that loss of ezrin increased RhoA activity and phosphorylation of MLC2 in both colonic and small intestinal epithelia of Vil2 2/2 mice. In this report, we first demonstrated that ezrin downregulated RhoA activity in neuronal cells. However, the mechanism by which ezrin inhibits RhoA activity has yet to be determined.
In Fig. 8 we propose our model in which ezrin is involved in the neuritogenesis through the regulation of RhoA/Rho kinase/ MLC2 activity. In the Vil2 +/+ neurons, ezrin down-regulates RhoA/Rho kinase pathway, which leads to the inhibition of phosphorylation of MLC2 and activation of myosin II, and promotes the neuritogenesis (Fig. 8A). On the other hand, in the Vil2 kd/kd neurons, activated RhoA/Rho kinase phosphorylates MLC2, which leads to the impairment of neuritogenesis (Fig. 8B). In fact, the inhibition of Rho kinase and myosin II rescued the morphological deficits found in the Vil2 kd/kd neurons (Fig. 6E, 7C). Although treatment of blebbistatin resulted in a significant increase in number of neurites similar to treatment of Y-27632, the length of neurites and axon was not affected in both Vil2 +/+ and Vil2 kd/kd neurons (Fig. 7D,E). These results suggest that myosin II activity which is controlled by phosphorylation of MLC2 and located downstream in the RhoA/Rho kinase pathway is mainly involved in neuritogenesis. These observations indicate that regulation of myosin II activity by ezrin is crucial for neuritogenesis.
In contrast to RhoA, activation of Rac1, Cdc42 and their downstream effectors such as PAK family of serine/threonine kinases promoted neuronal morphology through actin remodeling [5,6]. It should be noted that other ERM protein, merlin, which is the neurofibromatosis type 2 gene product, inhibited neurite extensions through the inactivation of Rac1 activity in cultured cerebellar Purkinje cells [36]. Conversely, in the present study, ezrin knockdown in the cultured cortical neurons specifically activated RhoA without any effects on Rac1 and Cdc42 (Fig. 3). Future studies are necessary to understand the mechanisms how each ERM protein regulates specific Rho family proteins.
In this study, we demonstrated that ezrin facilitated neuritogenesis by regulating RhoA/Rho kinase pathway. Conversely, ERM proteins are reported to be phosphorylated by several kinases. Here, we demonstrated that there were no differences in the phosphorylation of ezrin, radixin and moesin between the DMSOtreated and Y-27632-treated neurons (Fig. 5A,C-E). In addition, the phosphorylation level of radixin and moesin remained unaltered in the Vil2 kd/kd neurons where RhoA/Rho kinase pathway was activated (Fig. 4C,D). These findings suggest that endogenous phosphorylation of ERM proteins is not directly regulated by Rho kinase in the cultured cortical neurons. In fact, phosphoinositide 3-kinase (PI3 kinase), leucine-rich repeat kinase 2 (LRRK2) and protein kinase C (PKC) have been reported to be involved in phosphorylation of ERM proteins and neuronal development [32,37,38].
In the central nervous system, ezrin was detected in the developmental brain, rostral migratory stream and subventricular zone [39][40][41]. The expression of ezrin was also detected in radial glial cells characterized as stem cells in the intermediate zone in the prenatal human cerebrum [42]. In addition, survival and migration were affected by the inhibition of Rho kinase in neuronal stem cells in vivo [43]. Our results suggest a possible role for ezrin and its downstream effector, Rho kinase on the neuronal development in vivo.
In conclusion, our study revealed a new function of ezrin in the neuritogenesis using the Vil2 kd/kd neurons. Ezrin is involved in the neuritogenesis via down-regulation of the RhoA activity and inhibition of MLC2 phosphorylation in the cultured cortical neurons. Figure S1 Detection of ezrin in the Vil2 +/+ and Vil2 kd/kd neurons. A, Immunofluorescence of the Vil2 +/+ and Vil2 kd/kd neurons at stage 1 using an anti-ezrin antibody. Scale bar, 50 mm. B, Immunoblotting of cell extracts (10 mg, 1 mg, 0.5 mg or 0.2 mg) from the Vil2 +/+ and Vil2 kd/kd neurons (2 DIV) with an anti-ezrin antibody. (EPS)