Ca2+/Calmodulin-dependent Protein Kinase IV-mediated LIM Kinase Activation Is Critical for Calcium Signal-induced Neurite Outgrowth*

Actin cytoskeletal remodeling is essential for neurite outgrowth. LIM kinase 1 (LIMK1) regulates actin cytoskeletal remodeling by phosphorylating and inactivating cofilin, an actin filament-disassembling factor. In this study, we investigated the role of LIMK1 in calcium signal-induced neurite outgrowth. The calcium ionophore ionomycin induced LIMK1 activation and cofilin phosphorylation in Neuro-2a neuroblastoma cells. Knockdown of LIMK1 or expression of a kinase-dead mutant of LIMK1 suppressed ionomycin-induced cofilin phosphorylation and neurite outgrowth in Neuro-2a cells. Ionomycin-induced cofilin phosphorylation and neurite outgrowth were also blocked by KN-93, an inhibitor of Ca2+/calmodulin-dependent protein kinases (CaMKs), and STO-609, an inhibitor of CaMK kinase. An active form of CaMKIV but not CaMKI enhanced Thr-508 phosphorylation of LIMK1 and increased the kinase activity of LIMK1. Moreover, the active form of CaMKIV induced cofilin phosphorylation and neurite outgrowth, and a dominant negative form of CaMKIV suppressed ionomycin-induced neurite outgrowth. Taken together, our results suggest that LIMK1-mediated cofilin phosphorylation is critical for ionomycin-induced neurite outgrowth and that CaMKIV mediates ionomycin-induced LIMK1 activation.

neuroblastoma cells. We provide evidence that LIMK1-mediated cofilin phosphorylation plays a crucial role in ionomycininduced neurite outgrowth and that CaMKIV mediates the ionomycin-induced LIMK1 activation.
Knockdown of LIMK1-For testing the effect of LIMK1 knockdown on ionomycin-induced cofilin phosphorylation, Neuro-2a cells plated in 100-mm dishes (1.5 ϫ 10 6 cells/dish) were cultured for 20 h and then transfected with pSUPER. retro.puro-LIMK1 shRNA. Transfected cells were cultured for 24 h and selected by culturing for an additional 48 h with 7 g/ml puromycin. Then cells were replated on 60-mm dishes (2.0 ϫ 10 6 cells/dish), cultured for 16 h, serum-starved for 4 h, and then cultured with or without 1.5 M ionomycin for 30 min. For neurite outgrowth assays, Neuro-2a cells were plated on coverslips (2.5 ϫ 10 4 cells/35-mm dish), cultured for 24 h, and transfected with pSUPER-LIMK1 shRNA. After incubation for 24 h, the cells were treated with 1.5 M ionomycin, cultured for 48 h, and then fixed.
In Vitro Kinase Assay-Endogenous LIMK1 or Myc-LIMK1 expressed in Neuro-2a cells was immunoprecipitated with an anti-LIMK1 or anti-Myc antibody and then subjected to an in vitro kinase reaction using His 6 -cofilin as a substrate, as described previously (12). CaMKI-or CaMKIV-catalyzed LIMK1 activation was analyzed in cell-free assays as follows. GFPtagged CaMKI or CaMKIV mutants were expressed in Neuro-2a cells, immunoprecipitated with an anti-GFP antibody, and incubated in 20 l of lysis buffer containing 50 M ATP and 185 kBq of [␥-32 P]ATP (110 TBq/mmol; PerkinElmer Life Sciences) with 1 g/ml GST-LIMK1 and 2 g/ml His 6 -cofilin at 30°C for 30 min. GST-LIMK1 and His 6 -cofilin were expressed in Sf21 insect cells, using the Bac-to-Bac baculovirus expression system (Invitrogen), and purified by glutathione-Sepharose (GE Healthcare), as described (40). The reaction mixture was separated with SDS-PAGE and analyzed with autoradiography to measure 32 P-labeled cofilin and immunoblotting with an anti-P-LIMK1(T508) antibody.
Neurite Outgrowth Assay-Neuro-2a cells (2.5 ϫ 10 4 cells/ 35-mm dish) were plated on glass coverslips pretreated with 1 g/ml poly-L-lysine, cultured for 24 h, and transfected with expression plasmids or pSUPER shRNA plasmids. After incubation for 24 h, cells were treated or not treated with 1.5 M ionomycin and further cultured for 48 h. Then the cells were fixed with 4% formaldehyde in phosphate-buffered saline for 15 min. The CFP-positive cells with neurites longer than two cell body lengths were scored as neurite-bearing cells.

RESULTS
Ionomycin Induces LIMK1 Activation and Cofilin Phosphorylation-Previous studies showed that ionomycin promotes neurite outgrowth in Neuro-2a cells (23,24). To examine whether LIMK1 is involved in ionomycin-induced neurite outgrowth, we first analyzed changes in the kinase activity of LIMK1 and the level of P-cofilin in ionomycin-stimulated Neuro-2a cells. In an in vitro kinase assay using cofilin as a substrate, the kinase activity of LIMK1 increased 1.6-fold by 30 min after ionomycin stimulation and then gradually decreased LIMK1 Is Involved in Ionomycin-induced Cofilin Phosphorylation-To assess whether LIMK1 is involved in ionomycininduced cofilin phosphorylation, we knocked down LIMK1 by transfecting Neuro-2a cells with shRNA targeting LIMK1. Neuro-2a cells were transfected with pSUPER.retro.puro plasmids coding for control or LIMK1 shRNA target sequences and selected with puromycin. An immunoblot analysis showed that the LIMK1 shRNA suppressed expression of endogenous LIMK1 in Neuro-2a cells (Fig. 2A). The LIMK1 shRNA but not the control shRNA blocked ionomycin-induced cofilin phosphorylation (Fig. 2B). Furthermore, overexpression of a kinasedead form of LIMK1, LIMK1(D460A), in which the catalytic Asp-460 was replaced by Ala, significantly suppressed the ionomycin-induced elevation of the P-cofilin level in Neuro-2a cells (Fig. 2C). These results suggest that LIMK1 plays a critical role in ionomycin-induced cofilin phosphorylation in Neuro-2a cells.
LIMK1 Is Required for Ionomycininduced Neurite Outgrowth-To examine the role of LIMK1 in ionomycin-induced neurite outgrowth, we analyzed the effect of LIMK1 knockdown on neurite outgrowth in Neuro-2a cells. Neuro-2a cells were cotransfected with CFP and control or LIMK1 shRNA plasmids, cultured for 24 h, and then stimulated with ionomycin for 48 h. The cells were fixed, and the CFP fluorescence of the shRNA-transfected cells was visualized (Fig. 3A, top). We then scored the percentage of neurite-bearing cells with neurites longer than two cell body lengths in CFP-positive cells. Knockdown of LIMK1 almost completely blocked ionomycin-induced neurite outgrowth (Fig. 3A, bottom). Overexpression of kinase-dead LIMK1(D460A) also inhibited ionomycin-induced neurite outgrowth (Fig. 3B). These results suggest that LIMK1 plays a crucial role in ionomycin-induced neurite outgrowth in Neuro-2a cells and that LIMK1(D460A) functions as a dominant negative form.

Inhibitors of CaMKs and CaMKK Suppress Ionomycin-induced Cofilin
Phosphorylation and Neurite Outgrowth-To determine whether CaMKs and CaMKK are involved in ionomycin-induced cofilin phosphorylation and neurite outgrowth, we examined the effects of KN-93 and STO-609. KN-93 inhibits the kinase activities of CaMKI, CaMKII, and CaMKIV, and STO-609 inhibits CaMKK, which activates CaMKI and CaMKIV but not CaMKII. When Neuro-2a cells were pretreated with KN-93 or STO-609, ionomycin-induced cofilin phosphorylation was almost completely blocked (Fig.  4A), and ionomycin-induced neurite outgrowth was significantly suppressed (Fig. 4, B and C). Thus, CaMKK and one or both of its downstream kinases, CaMKI and CaMKIV, are involved in ionomycin-induced cofilin phosphorylation and neurite outgrowth.

JOURNAL OF BIOLOGICAL CHEMISTRY 28557
affect Myc-LIMK1 activity. The kinase-inactive Myc-LIMK1-(D460A), when coexpressed with active CaMKs, did not exhibit any cofilin-phosphorylating activity (data not shown). These results indicate that CaMKIV but not CaMKI has the potential to activate LIMK1.
CaMKIV Is Activated by Ionomycin and Induces Cofilin Phosphorylation-CaMKIV is activated by CaMKK-catalyzed phosphorylation at Thr-196 (34). To examine whether CaMKIV is activated by ionomycin stimulation, we analyzed changes in the level of Thr-196 phosphorylation of CaMKIV in ionomycin-stimulated Neuro-2a cells by immunoblotting with an anti-P-CaMKIV(T196) antibody. The level of P-CaMKIV(T196) increased about 2.0-fold by 10 min after ionomycin stimulation (Fig.  6A), which suggests that ionomycin induces CaMKIV activation in Neuro-2a cells. We also analyzed whether active CaMKIV induces cofilin phosphorylation in Neuro-2a cells. Expression of CaMKIV-(1-313) but not CaMKIV-(K71E) increased the level of P-cofilin (Fig.  6B). These results further support the notion that CaMKIV mediates ionomycin-induced cofilin phosphorylation. In addition, we also examined whether CaMKI is activated by ionomycin by measuring the level of Thr-177 phosphorylation of CaMKI (35). The level of P-CaMKI(T177) increased 5-10 min after ionomycin treatment (Fig. 6C). Thus, CaMKI is also activated by ionomycin treatment in Neuro-2a cells.

Both CaMKI and CaMKIV Are Involved in Ionomycin-induced Neurite Outgrowth-Previous studies showed that
CaMKI is involved in neurite outgrowth (24 -26). To further examine the roles of CaMKI and CaMKIV in ionomycin-induced neurite outgrowth in Neuro-2a cells, we analyzed the effects of expressing the constitutively active and dominant negative forms of these CaMKs. Expression of a dominant negative form of either CaMKI or CaMKIV (CaMKI-(K49E) or CaMKIV-(K71E)) significantly suppressed ionomycin-induced neurite outgrowth (Fig. 7). In addition, expression of a consti- tutively active mutant of CaMKI or CaMKIV (CaMKI-  or CaMKIV-(1-313)) promoted neurite outgrowth in Neuro-2a cells (Fig. 7). These results suggest that both CaMKI and CaMKIV are involved in ionomycin-induced neurite outgrowth in Neuro-2a cells.

DISCUSSION
In this study, we identified a novel signaling pathway that is crucial for calcium signal-induced neurite outgrowth. We showed that ionomycin induces LIMK1 activation and cofilin phosphorylation and that depletion of LIMK1 or expression of kinase-dead LIMK1 blocks ionomycin-induced cofilin phosphorylation and neurite outgrowth. This indicates that LIMK1 mediates calcium signal-induced cofilin phosphorylation and neurite outgrowth. The findings that ionomycin-induced cofi-lin phosphorylation and neurite outgrowth were blocked by STO-609 and KN-93 indicate that CaMKK-mediated activation of CaMKI and/or CaMKIV is involved in calcium signal-induced cofilin phosphorylation and neurite outgrowth. Interestingly, an active form of CaMKIV but not CaMKI increased the kinase activity of LIMK1 in both cell-free and coexpression assays. Furthermore, expression of active CaMKIV enhanced cofilin phosphorylation and neurite outgrowth, and a dominant negative form of CaMKIV inhibited ionomycin-induced neurite outgrowth. Based on these observations, we propose a novel signaling pathway composed of Ca-MKK-CaMKIV-LIMK1-cofilin that plays a crucial role in calcium signal-induced neurite outgrowth in Neuro-2a cells. The level of LIMK1 activation in response to ionomycin is small relative to the effect on neurite extension. This may reflect that ionomycin induces LIMK1 activation in the limited region of the cell, such as the site where neurites start to grow.
It has been proposed that the CaMKK-CaMKIV cascade is involved in calcium-induced neurite growth through phosphorylation and activation of the transcription factor CREB (27)(28)(29)(30)(31). On the other hand, we showed here that the CaMKK-CaMKIV pathway contributes to ionomycin-induced neurite outgrowth through LIMK1 activation and cofilin phosphorylation. These results suggest that CaMKIV plays dual roles in calcium-induced neurite outgrowth: the short term role of promoting the initiation of neurite outgrowth via LIMK1 activation and actin reorganization and the long term role of contributing to neurite extension and maintenance via activation of CREB-dependent gene transcription.
Although CaMKIV is predominantly localized in the nucleus, it is also distributed in the cytoplasm and has the potential to shuttle between the nucleus and cytoplasm (41)(42)(43). Thus, it is possible that LIMK1 is activated by CaMKIV in the cytoplasm. On the other hand, given that LIMK1 also shuttles between the nucleus and cytoplasm using its intrinsic nuclear export and nuclear localization signals (44,45), LIMK1 may be activated in the nucleus and then transported to the cytoplasm to regulate actin cytoskeleton in this region.
Similar to the effects of CaMKIV mutants, expression of an active form of CaMKI induced neurite outgrowth, and a dominant negative form of CaMKI suppressed ionomycin-induced neurite outgrowth in Neuro-2a cells, thus indicating that CaMKI, like CaMKIV, plays a crucial role in calcium signalinduced neurite outgrowth. Although CaMKI has been shown to activate Rac (an upstream regulator of LIMK1) through activation of Rac-guanine nucleotide exchange factors, such as STEF and ␤PIX (46,47), CaMKI failed to activate LIMK1 in our coexpression experiments in Neuro-2a cells (Fig. 5A). Thus, CaMKI acts in neurite outgrowth by a mechanism distinct from Rac-mediated LIMK1 activation, at least in Neuro-2a cells. In this respect, previous studies showed that CaMKI induced neurite outgrowth by activating extracellular signal-regulated kinase, which in turn stimulated CREB-dependent transcription (22,26), or by activating MARK2 (24). We showed that expression of kinase-dead MARK2 suppressed CaMKI-induced neurite outgrowth, indicating that MARK2 mediates CaMKI-induced neuritogenesis. Because MARK2 regulates microtubule reorganization by phosphorylating microtubule-associated proteins (48), it is likely that CaMKI contributes to neurite outgrowth by regulating microtubule dynamics via MARK2, in addition to activating CREB-mediated gene expression. Thus, calcium signals induce neurite outgrowth by stimulating several pathways, including the CaMKIV-mediated LIMK1 activation that induces actin remodeling and the CaMKI-mediated MARK2 activation that regulates microtubule dynamics as well as the CaMKI/CaMKIV-mediated transcriptional activation of CREB. It is thus likely that the coordinated regulation of actin filaments and microtubules is necessary for the induction of neurite outgrowth.
Because CaMKIV increased the level of Thr-508 phosphorylation of LIMK1 in a cell-free assay, it is likely that CaMKIV directly activates LIMK1 by Thr-508 phosphorylation. We and other investigators previously showed that ROCK and PAK activate LIMK1 by phosphorylation at Thr-508, which is in the activation loop within the kinase domain of LIMK1 (10 -12). Thus, various signaling pathways, such as Rho-ROCK, Rac-PAK1, and CaMKK-CaMKIV, activate LIMK1 by the same mechanism, Thr-508 phosphorylation. This suggests that LIMK1 activation is a point of convergence that links various signaling pathways to the regulation of actin cytoskeletal reorganization in cells.
Previous studies showed that knockdown of LIMK1 or inhibition of LIMK1 activity suppresses neurite outgrowth in PC12 rat pheochromocytoma cells, chick dorsal root ganglion neurons, and hippocampal pyramidal neurons (18 -20), which suggests that LIMK1 has a critical role in neurite outgrowth. On the other hand, overexpression of LIMK1 or depletion of SSH1/ SSH2 also suppressed neurite outgrowth in PC12 cells and chick dorsal root ganglion neurons (16,20). These results indicate that a precise level of LIMK1 activation and a proper balance of cofilin phosphorylation and dephosphorylation by LIMK and SSH activities are necessary for neurite outgrowth. We previously showed that calcium signals induce SSH1 activation and cofilin dephosphorylation via calcineurin in other types of cells (49). Thus, it is possible that calcium signals stimulate both LIMK1 and SSH1 to regulate the actin cytoskeleton  and that the spatial and temporal regulation of LIMK1 and SSH1 activities and the balance between them are important for the control of neurite outgrowth. Because LIMK1 depletion inhibits stimulus-induced actin filament assembly (20,50), LIMK1 is probably involved in the actin filament assembly and stabilization required for the initial phase of neurite outgrowth. In contrast, cofilin is required for neurite extension because it stimulates actin filament disassembly and thereby supplies actin monomers to the leading edge of the extending growth cones (5,51); therefore, SSH is likely involved in the neurite extension by promoting actin filament turnover via cofilin dephosphorylation and reactivation.
In conclusion, we have identified a novel signaling pathway that transduces calcium elevations into LIMK1 activation and actin cytoskeletal reorganization via CaMKK and CaMKIV. We have the data that brain-derived neurotrophic factor induces LIMK1 activation in primary rat cortical neurons and that expression of dominant negative CaMKIV or depletion of LIMK1 blocked brain-derived neurotrophic factor-induced dendritogenesis, which suggests that the CaMKIV-LIMK1 pathway is probably involved in brain-derived neurotrophic factor-induced dendritogenesis in primary neurons. In further studies, we will elucidate the roles of this signaling pathway in a variety of calcium-induced actin cytoskeletal responses in neuronal and non-neuronal cells. Calcium signals regulate various important neural functions and development, including axon guidance, dendrite arborization, spine morphology, synaptic plasticity, and learning and memory formation. Actin cytoskeletal remodeling mediated by the LIMK1-cofilin pathway plays an important role in the dynamics of dendritic spine structures, persistence of long term potentiation, and synaptic plasticity (52)(53)(54). Therefore, CaMKIV-mediated LIMK1 activation and cofilin phosphorylation may be among the important mechanisms by which calcium signals regulate synaptic plasticity and higher order brain functions, in addition to those needed for neurite outgrowth. Our identification of the CaMKIV-LIMK1 pathway will help to elucidate the mechanisms by which actin cytoskeletal remodeling is regulated in response to various stimuli mediated by calcium signaling.