Phosphorylation of Collapsin Response Mediator Protein-2 by Rho-kinase

We previously identified Rho-associated protein kinase (Rho-kinase) as a specific effector of Rho. In this study, we identified collapsin response mediator protein-2 (CRMP-2), as a novel Rho-kinase substrate in the brain. CRMP-2 is a neuronal protein whose expression is up-regulated during development. Rho-kinase phosphorylated CRMP-2 at Thr-555 in vitro. We produced an antibody that specifically recognizes CRMP-2 phosphorylated at Thr-555. Using this antibody, we found that Rho-kinase phosphorylated CRMP-2 downstream of Rho in COS7 cells. Phosphorylation of CRMP-2 was observed in chick dorsal root ganglion neurons during lysophosphatidic acid (LPA)-induced growth cone collapse, whereas the phosphorylation was not detected during semaphorin-3A-induced growth cone collapse. Both LPA-induced CRMP-2 phosphorylation and LPA-induced growth cone collapse were inhibited by Rho-kinase inhibitor HA1077 or Y-32885. LPA-induced growth cone collapse was also blocked by a dominant negative form of Rho-kinase. On the other hand, semaphorin-3A-induced growth cone collapse was not inhibited by a dominant negative form of Rho-kinase. Furthermore, overexpression of a mutant CRMP-2 in which Thr-555 was replaced by Ala significantly inhibited LPA-induced growth cone collapse. These results demonstrate the existence of Rho-kinase-dependent and -independent pathways for growth cone collapse and suggest that CRMP-2 phosphorylation by Rho-kinase is involved in the former pathway.

During the development of the nervous system, the nerve growth cones play a central role in axon guidance. They are located at the tip of axons and dynamically change their morphology in response to attractive and repulsive cues to decide the growing direction (1,2). Such morphological changes of growth cones are considered to be achieved by the reorganiza-tion of the cytoskeletons and cell adhesions (3)(4)(5). Although recent studies have identified several guidance cue molecules and their receptors (1,2), the mechanisms of their intracellular signaling are poorly understood.
The Rho family of small GTPases including Rho, Cdc42, and Rac are intracellular signaling molecules that are thought to regulate the cytoskeletons and cell adhesions (6 -8). In fibroblasts, Rho activation is required for the formation of stress fibers and focal adhesions (9). On the other hand, Cdc42 and Rac are required for the formation of filopodia and lamellipodia, respectively (9 -11). These structures regulated by Rho, Cdc42, and Rac in fibroblasts are similar to those of nerve growth cones (2,12). Consistently, accumulating evidence suggests that the Rho family GTPases also regulate the neurite and growth cone morphology (13,14). It is reported that Rho negatively regulates growth cone and neurite formation. Microinjection of RhoA as well as application of LPA, 1 which activates intracellular Rho, induces growth cone collapse and neurite retraction in PC12 and N1E-115 cells (13)(14)(15). On the other hand, trituration of Clostridium botulinum C3 transferase (C3), a specific inhibitor of Rho, stimulates neurite outgrowth in DRG neurons (16). A dominant negative form of RhoA and C3 exoenzyme also promote filopodia and lamellipodia formation in the N1E-115 cell growth cone (14). Interestingly, in dorsal root ganglion (DRG) neurons, C3 treatment inhibits the lamellipodial spreading of growth cones (16), thereby suggesting that Rho activity is also required for the maintenance of growth cones.
Several groups including ours have found that Rho-kinase is involved in LPA-induced neurite retraction of neuronal cells downstream of Rho (27,28) and have identified MLC as one of the major substrates of Rho-kinase-mediated neurite retraction (28). However, the activated MLC could not completely mimic the Rho-kinase-induced neurite retraction, raising the possibility that other Rho-kinase substrates are also involved in the Rho-kinase-induced neurite retraction.
In light of these observations, we here searched for novel substrates of Rho-kinase in brain and identified collapsin response mediator protein-2 (CRMP-2). We also identified Thr-555 as the site of phosphorylation in CRMP-2 by Rho-kinase. Monitoring this phosphorylation, we found the existence of Rho-kinase-dependent and -independent pathways for growth cone collapse.
Purification of a Rho-kinase Substrate, p70 -Bovine brain cytosol fraction (20 mg of protein) prepared as described (31,32) was loaded onto the Mono Q HR5/5 column (Amersham Pharmacia Biotech) preequilibrated with 10 ml of buffer A (20 mM Tris-HCl at pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 5 mM MgCl 2 ). After being washed with buffer A, proteins were eluted with a linear gradient of NaCl (0 -0.5 M) at a flow rate of 0.5 ml/min. Fractions (1 ml each) were collected, and an aliquot of each fraction (20 l) was analyzed by the phosphorylation assay as described below. The peak fractions containing the major phosphorylated protein were loaded onto the Mono S HR5/5 column (Amersham Pharmacia Biotech) using the same flow rate and gradient of NaCl. An aliquot of each fraction (20 l) was used for the phosphorylation assay.
Phosphorylation Assay-The phosphorylation assay of the samples was carried out as described (22). In brief, the kinase reaction for Rho-kinase was carried out in 50 l of a reaction mixture (50 mM Tris-HCl at pH 7.5, 2 mM EDTA, 1 mM dithiothreitol, 7 mM MgCl 2 , 10 M [␥-32 P]ATP (1-20 GBq/mmol), and purified glutathione S-transferase (GST)-dominant active form of Rho-kinase (GST-CAT) (1 g of protein)) for 10 -60 min at 30°C. GST-CAT was produced in Sf9 cells with a baculovirus system (33) and purified on glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) (19). Then the reaction mixtures were boiled in SDS sample buffer and subjected to SDS-PAGE. The radiolabeled bands were visualized by an image analyzer (BAS 2000; Fuji, Tokyo, Japan). To confirm the site of phosphorylation of CRMP-2 by Rho-kinase, HA-CRMP-2 and HA-CRMP-2 T555A were used as substrates. HA-CRMP-2 or HA-CRMP-2 T555A was prepared by immunoprecipitation from COS7 cells transiently transfected with each expression vector as described (32).
Peptide Sequencing-Peptide sequencing of p70 was carried out as described (34). In brief, partially purified p70 was resolved by SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and stained with 0.1% Ponceau S in 1% acetic acid. The band corresponding to p70 was cut out and digested by lysyl endopeptidase, Achromobacter protease I, and endoproteinase, Asp-N. The obtained peptides were fractionated by C18 column chromatography (Shiseido, Japan) and subjected to amino acid sequencing for identification.
Identification of the Phosphorylation Site of CRMP-2 by Rho-kinase-The site of phosphorylation of CRMP-2 by Rho-kinase was identified as described (26,35). In brief, CRMP-2 (77 g of protein) was phosphoryl-ated by GST-CAT (70 g of protein), and the reaction product was digested with Asp-N. The obtained peptides were fractionated by C18 reverse-phase column chromatography and subjected to amino acid sequencing. The fraction obtained from each Edman degradation cycle was measured for 32 P in a Beckman liquid scintillation counter.
Production of Site-and Phosphorylation State-specific Antibody for CRMP-2-A rabbit polyclonal antibody against CRMP-2 phosphorylated at Thr-555 (anti-pT555) was prepared as described (35), using the chemically synthesized phosphopeptide Cys-Ile 550 -Pro-Arg-Arg-Thr-Thr(P)-Gln-Arg-Ile-Val-Ala 560 as an antigen by Peptide Institute Inc. For the confirmation of the specificity of anti-pT555, the fraction 11 from the Mono S column, which contained p70 (Fig. 1B), was subjected to phosphorylation assay and used for immunoblot analysis.
Culture Preparation-COS7 cells were seeded on 100-mm culture dishes at 5.7 ϫ 10 3 cells/mm 2 and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, streptomycin, and penicillin in an atmosphere of 5% CO 2 . Chick DRG neurons were dissociated from 7-day-old chick embryos using papain as described (36). They were then seeded on 13-mm round glass coverslips coated by laminin at 1.5 ϫ 10 4 cells/mm 2 in 24-well plates for immunostaining or on 35-mm culture dishes coated by laminin at 1.5 ϫ 10 4 cells/mm 2 for immunoblot analysis. The cells were cultured in Dulbecco's modified Eagle's medium containing 100 ng/ml 2.5 S nerve growth factor (NGF) (Upstate Biotechnology, Inc., Lake Placid, NY), 10% fetal bovine serum, streptomycin, and penicillin in an atmosphere of 5% CO 2 .
Detection of CRMP-2 Phosphorylated at Thr-555 in COS7 Cells and DRG Neurons-Transfection of plasmids into COS7 cells was carried out using a LipofectAMINE-mediated DNA transfection procedure (Life Technologies, Inc.). After a 24-h incubation, the transfected cells were incubated in serum-free medium for 24 h. The cells were treated with 10% (w/v) trichloroacetic acid. The resulting precipitates were subjected to immunoblot analysis using anti-pT555, anti-CRMP-2, and anti-Myc antibodies. DRG neurons were cultured in serum-plus medium for 20 h and then cultured in serum-free medium without NGF for 4 h. The pretreatment with Rho-kinase inhibitor, HA1077 or Y-32885, was performed for 1 h before collapse assay. For immunoblot analysis, the cells were stimulated by 5 M LPA or 5 units/ml Sema3A for the indicated time at 37°C in serum-free medium without NGF. The cells were treated with 10% (w/v) trichloroacetic acid. The resulting precipitates were subjected to immunoblot analysis using anti-pT555 and anti-CRMP-2 antibody.
Collapse Assay-Three hours after seeding, DRG neurons were transfected with plasmids using the calcium phosphate procedure as described (37). Both the transfected neurons and nontransfected neurons were cultured as described above. The collapse assay was performed as described (38). In brief, the collapsed and noncollapsed growth cones were counted by two independent persons in a blind manner. The criterion for the collapsed growth cones was a total loss of filopodia and lamellipodia. For immunofluorescence analysis, the cells were stimulated by 1 M LPA or 5 units/ml Sema3A for 30 min at 37°C in serum-free medium without NGF. We observed similar responses of Sema3A-stimulated DRG neurons in the presence or absence of serum.
Immunofluorescence Analysis-The cells were fixed with 3.7% formaldehyde in phosphate-buffered saline for 10 min at room temperature and then treated with methanol for 10 min at Ϫ30°C or 0.05% Triton X-100 for 10 min at 4°C. They were then treated with 10% normal goat serum in phosphate-buffered saline for 1 h at room temperature and incubated with anti-␤-galactosidase antibody (Promega, Madison, WI), anti-Myc antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), or tetramethylrhodamine B isothiocyanate-phalloidin. The cells were also counterstained with anti-neurofilament antibody (Sigma). The immunoreactivities were visualized by incubation with fluorescein isothiocyanate-conjugated anti-rabbit or mouse Ig antibody and Texas Redconjugated anti-rabbit or mouse Ig antibody. The samples were examined using a Zeiss axiophoto microscope LSM510 (Carl Zeiss, Obeokochem, Germany) or PXL cooled CCD camera (Photometric, Tucson, AZ) with DeltaVision processing software (Applied Precision Inc.).

Identification of CRMP-2 as a Novel Rho-kinase Substrate in
Bovine Brain-To search for Rho-kinase substrates in brain, we separated bovine brain cytosol proteins by Mono Q column chromatography and subjected the fractions to phosphorylation assay using a dominant active form of Rho-kinase (GST-CAT). Among the several proteins detected in this assay, one with a mass of about 70 kDa (p70) was recognized as a major band phosphorylated in a GST-CAT-dependent manner (Fig. 1A, fractions 8 -12). p70 was eluted in peaks between 200 and 300 mM NaCl and further purified by Mono S column chromatography (Fig. 1B). To clarify the molecular identity of p70, the purified protein was subjected to amino acid sequencing as described under "Experimental Procedures." Six peptide sequences derived from p70 were determined: KQIGENLIVP, KSSAEVIAQARK, KMDENQFVAV, KVFNLYPR, KIVLED, and KAIEALAELRXVP. These sequences were almost identical to that of bovine CRMP-2 (Fig. 2). Furthermore, the antibody raised against CRMP-2 (amino acids 1-572) cross-reacted with p70 (Fig. 5). Recently, CRMP-2 homologues have been identified from various species (29, 39 -43). CRMP-62, the chick CRMP-2 (98% identity), is reported to be required for the growth cone collapse of DRG neurons induced by Sema3A (also known as collapsin-1) (29). UNC-33, the C. elegans homologue (30% homology), is identified by a mutation resulting in severely uncoordinated movement, abnormalities in axon guidance, and superabundance of microtubules in neurons (44,45).
Identification of the in Vitro CRMP-2 Phosphorylation Site by Rho-kinase as Thr-555-Next, we determined phosphorylation sites of CRMP-2 by Rho-kinase. The purified CRMP-2 was phosphorylated by GST-CAT in the presence of [␥-32 P]ATP in vitro and digested with endoproteinase, Asp-N. The digested peptides were separated by C18 column chromatography. One radioactive peak was obtained (Fig. 3A). The phosphorylated peptide in the fraction was subjected to amino acid sequencing. The sequence obtained from this fraction was DNIPRRT-TQRIVAPPGGR, corresponding to amino acids 548 -565 of CRMP-2. The fraction obtained from each Edman degradation cycle was measured for 32 P in a liquid scintillation counter. As a result, Thr-555 was found to be phosphorylated by GST-CAT (Fig. 3B). To rule out the possibility that GST-CAT phosphorylates not only Thr-555 but also Thr-554, we produced a CRMP-2 mutant (HA-CRMP-2 T555A) in which Thr-555 was replaced by Ala. HA-CRMP-2 or HA-CRMP-2 T555A was expressed in COS7 cells, and the cell lysates were immunoprecipitated with anti-HA antibody. The resulting immunoprecipitate was subjected to the phosphorylation assay (Fig. 4). HA-CRMP-2 was phosphorylated by GST-CAT, whereas HA-CRMP-2 T555A was not. Taken together, these results indicate that Thr-555 is the major site of CRMP-2 phosphorylation by Rho-kinase in vitro.
Production of a Site-and Phosphorylation State-specific Antibody for CRMP-2 at Thr-555-To examine in vivo CRMP-2 phosphorylation by Rho-kinase, we prepared a rabbit polyclonal antibody that specifically recognizes CRMP-2 phosphorylated at Thr-555 (anti-pT555). We used a phosphopeptide corresponding to amino acids 550 -560 of CRMP-2 in which Thr-555 is phosphorylated as an antigen. The specificity of this antibody was examined by immunoblot analysis (Fig. 5). Two pmol of purified CRMP-2 (p70) in the fraction 11 from the Mono S column (Fig. 1B) containing increasing amounts of the phosphorylated CRMP-2 was loaded on the gel. Anti-pT555 bound to the phosphorylated CRMP-2 in a dose-dependent manner but did not react with the unphosphorylated form. The binding of anti-pT555 to CRMP-2 phosphorylated by Rho-kinase was inhibited by preincubation of the antibody with the antigen phosphopeptide (data not shown). These results indicate that FIG. 1. Screening of Rho-kinase substrates from bovine brain. A, bovine brain cytosol fraction (20 mg of protein) was loaded onto the Mono Q column. Proteins were fractionated by elution with the indicated linear gradient of NaCl. The protein content of eluted fractions was monitored by UV absorbency at 280 nm. Each fraction was subjected to the phosphorylation assay. The reaction mixture was subjected to SDS-PAGE, and the radioactive proteins were detected by autoradiography. Fractions 8 -12 contained a major Rho-kinase substrate, p70. B, fractions 10 and 11 obtained above were further purified by Mono S column chromatography. Twenty l of each fraction were subjected to the phosphorylation assay. The reaction mixture was subjected to SDS-PAGE and silver staining. The radioactive proteins were detected by autoradiography. anti-pT555 specifically recognizes the phosphorylation of CRMP-2 at Thr-555.
Rho/Rho-kinase-dependent Phosphorylation of CRMP-2 at Thr-555 in COS7 Cells-Next, we examined whether Rho-kinase can phosphorylate CRMP-2 at Thr-555 in the cells, using anti-pT555 and COS7 cells. Because endogenous CRMP-2 was undetectable in COS7 cells by immunoblot analysis using anti-CRMP-2 antibody, HA-CRMP-2 was expressed exogenously with a dominant active or negative form of Rho family members or wild type or a dominant active form of Rho-kinase. An equivalent amount of HA-CRMP-2 was expressed in each case (Fig. 6, middle). The phosphorylation of HA-CRMP-2 was not detectable in serum-starved COS7 cells expressing HA-CRMP-2 alone. The expression of a dominant active form of RhoA (RhoA V14 ) induced the CRMP-2 phosphorylation, whereas that of a dominant negative RhoA (RhoA N19 ) did not (Fig. 6, top). The expression of wild-type Rho-kinase (Rhokinase) or a dominant active Rho-kinase (Rho-kinase CAT) also induced CRMP-2 phosphorylation. Furthermore, the coexpression of a dominant active RhoA (RhoA V14 ) further enhanced the Rho-kinase-induced CRMP-2 phosphorylation. In addition, HA-CRMP-2 T555A coexpressed with Rho-kinase CAT was not recognized by anti-pT555. Taken together, these results indicate that CRMP-2 was phosphorylated at Thr-555 by Rhokinase downstream of Rho in COS7 cells. Interestingly, the expression of a dominant active Rac1 (Rac1 V12 ) or a dominant active Cdc42 (Cdc42 V12 ) also resulted in a small increase in the level of HA-CRMP-2 phosphorylation (Fig. 6, upper panel). This result raises the possibility that CRMP-2 is also phosphorylated downstream of Rac1 and Cdc42 in COS7 cells.
LPA Stimulation of DRG Neurons Induced CRMP-2 Phos- phorylation by Rho-kinase-CRMP-2 is reported to be highly expressed in the developing nerve system (39,41,42) and implicated in Sema3A-induced growth cone collapse (29). In addition, the C. elegans homologue of CRMP-2, UNC-33, is thought to control the guidance and outgrowth of neuronal axons (44,45). Recently, we found that CRMP-2 plays a critical role in the axon formation of primary cultured hippocampal neurons. 2 On the other hand, recent studies reported that Rho is involved in the regulation of growth cone morphology (16,46). These observations raise the possibility that the CRMP-2 phosphorylation by Rho-kinase is involved in the regulation of the growth cone morphology, especially growth cone collapse. To address this issue, we monitored the phosphorylation of CRMP-2 in chick DRG neurons during growth cone collapse induced by LPA or Sema3A using anti-pT555 (Fig. 7). DRG neurons cultured for 24 h were stimulated by 5 M LPA or by 5 units/ml Sema3A for 1, 3, 10, or 30 min, and these stimuli induced growth cone collapse as reported previously (16) (Fig.   8A). The addition of 5 M LPA induced rapid phosphorylation of endogenous CRMP-2 at Thr-555 (Fig. 7A). The phosphorylation of CRMP-2 was increased during the first 3 min up to about 5-fold the basal level. On the other hand, we could not observe the increase of the level of CRMP-2 phosphorylation induced by the addition of Sema3A (Fig. 7A). These results suggest that CRMP-2 is phosphorylated during LPA-induced growth cone collapse of DRG neurons but not during Sema3A-induced collapse. Next, to examine whether LPA-induced phosphorylation of CRMP-2 at Thr-555 was mediated by Rho-kinase, DRG neurons were stimulated by LPA in the presence of Rho-kinase inhibitor, HA1077 or Y-32885, for 1 h. HA1077 and Y-32885 inhibited basal level and LPA-induced phosphorylation of CRMP-2 (Fig. 7B). These results indicate that CRMP-2 is phosphorylated at Thr-555 by Rho-kinase during LPA-induced growth cone collapse of DRG neurons.
Inhibition of Rho-kinase Blocks LPA-induced, but Not Sema3A-induced, Growth Cone Collapse-We next examined whether Rho-kinase mediates LPA-or Sema3A-induced growth cone collapse. In the control cultures, exposure to LPA or Sema3A for 3, 10, or 30 min increased the percentage of collapsed growth cones (Fig. 8). However, in the HA1077-or Y-32885-treated cultures, LPA-induced growth cone collapse was completely inhibited from 10 to 30 min. On the other hand, Sema3A-induced growth cone collapse was a little inhibited for 30 min (Fig. 8). Next, we examined the effects of a dominant negative Rho-kinase on LPA-induced or Sema3A-induced growth cone collapse. In neurons transfected with the control vector, exposure to LPA or Sema3A increased the percentage of the collapsed growth cones from about 20% to about 70%. The population of the collapsed growth cones was higher than that of untransfected neurons (Fig. 9B). The reason for the increase in the collapsing rate in transfected neurons is unclear, and the possibility of the effects of gene overexpression cannot be ruled out. The expression of a dominant negative Rho-kinase, RB/ PH(TT), completely blocked LPA-induced growth cone collapse (Fig. 9). However the Sema3A-induced growth cone collapse was insensitive to RB/PH(TT) (Fig. 9). In COS7 cells, CRMP-2 was phosphorylated by the expression of a dominant active Rac1 or Cdc42 (Fig. 6). We have also found that CRMP-2 was phosphorylated at Thr-555 by myotonic dystrophy kinase-related Cdc42-binding kinase ␤, an effector of Cdc42 (47), at a lower level than Rho-kinase in COS7 cells (data not shown). However, RB/PH(TT), which is a specific inhibitor of Rho-kinase and has no effect on the activity of myotonic dystrophy kinase-related Cdc42-binding kinase ␤ (48), inhibited LPAinduced growth cone collapse. Thus, we conclude that the LPAstimulated phosphorylation of CRMP-2 was induced by Rhokinase. We consider that slight inhibition of Sema3A-induced growth cone collapse by HA1077 and Y-32885 (Fig. 8) may be due to the inactivation of other kinases than Rho-kinase by the drugs, because Sema3A-induced growth cone collapse was not inhibited by a more Rho-kinase specific inhibitory molecule RB/PH(TT) (30,48). Taken together, these results suggest that Rho-kinase mediates LPA-induced growth cone collapse but does not play a central role in Sema3A-induced growth cone collapse.
CRMP-2 Mutant T555A Partially Inhibits LPA-induced Growth Cone Collapse-To examine the roles of CRMP-2 phosphorylation by Rho-kinase in growth cone morphology, we expressed the CRMP-2 mutant T555A, in which the Rho-kinase phosphorylation site was mutated, in DRG neurons. The expression of wild-type HA-CRMP-2 had no effect on LPA-induced growth cone collapse (Fig. 10). On the other hand, the expression of the mutant HA-CRMP-2 T555A partially but significantly inhibited LPA-induced growth cone collapse (Fig.  10). These results suggest that the phosphorylation of CRMP-2 by Rho-kinase is at least partly involved in LPA-induced growth cone collapse. DISCUSSION In the present study, we identified CRMP-2 as a novel Rhokinase substrate in brain. CRMP-2 was recognized as a major band phosphorylated in the bovine brain cytosol fractions. Rhokinase phosphorylated CRMP-2 at Thr-555 in vitro and in COS7 cells. Furthermore, we demonstrated that CRMP-2 is phosphorylated by Rho-kinase in DRG neurons during LPAinduced growth cone collapse.
Recent studies have identified various axon guidance mole- cules and their receptors (1,2). However, little information is available about the intracellular mechanisms responsible for axon guidance and growth cone regulation. In this study, we examined whether Rho-kinase is involved in growth cone collapse induced by LPA and Sema3A. Interestingly, while both LPA and Sema3A induced growth cone collapse of DRG neurons, only the former stimulated CRMP-2 phosphorylation by Rho-kinase. Furthermore, the inhibition of Rho-kinase activity by Rho-kinase inhibitors prevented LPA-induced growth cone collapse completely but not Sema3A-induced collapse. These results clearly indicate that there are Rho-kinase-dependent and -independent pathways for growth cone collapse. LPAinduced growth cone collapse occurs via the former, while Sema3A-induced collapse involves the latter.
Current reports have shown that Rho-family GTPases are involved in growth cone collapse induced by repulsive guidance cues. A constitutively active form of RhoA inhibited myelininduced growth cone collapse of motor neurons (49). In addition, retinal neurons treated with C3 extended neurites on myelin and myelin-associated glycoprotein substrates (50). Thus, myelin and myelin-associated glycoprotein are good candidates for extracellular guidance cues to activate the Rhokinase signaling pathway. Rac1 is implicated in Sema3A-induced growth cone collapse. A dominant negative form of Rac1 inhibited Sema3A-induced DRG and motor neuron growth cone collapse (16,49). However, we did not detect the phosphorylation of CRMP-2 by Rho-kinase downstream of Sema3A. We consider that Sema3A signaling is activated without affecting Rho-kinase activity. The morphology of growth cones is thought to be regulated by multiple factors such as assembly, disassembly, and retrograde flow of actin filaments, microtubule organizations, and cell adhesions (6 -8, 51). Therefore, it is not unexpected that multiple signaling pathways regulate the growth cone morphology. In line with this, Poo and colleagues (52) recently reported differential signaling pathways for two groups of guidance cues in terms of dependence upon Ca 2ϩ , phospholipase C-␥, and phosphoinositide 3-kinase. Further analysis of cell signaling by a series of attractive and repulsive guidance cues is absolutely needed in the future. Anti-pT555 produced in the present study will provide a powerful tool to monitor the Rho-kinase activation by various guidance cues.
What are the functions of the CRMP-2 phosphorylation by Rho-kinase? CRMP-2 is implicated in Sema3A-induced growth cone collapse (29). The present study proposed that the phosphorylation of CRMP-2 by Rho-kinase is not involved in the Sema3A signaling pathway. Therefore, CRMP-2 may mediate the signaling of Sema3A irrespective of phosphorylation by Rho-kinase. On the other hand, we demonstrated that CRMP-2 was phosphorylated by Rho-kinase during LPA-induced growth cone collapse and that the mutant that mimics the dephosphorylated form of CRMP-2 partially but significantly inhibited LPA-induced growth cone collapse. These results suggest that CRMP-2 phosphorylation by Rho-kinase plays a role in LPAinduced growth cone collapse. Previously, we also reported that the mutant that mimics the Rho-kinase-phosphorylated MLC could not completely induce Rho-kinase-induced neurite retraction in N1E-115 cells (28), thereby suggesting that MLC phosphorylation by Rho-kinase also plays only a partial role in Rho-kinase-induced neurite retraction. Given that the growth cone morphology is regulated by multiple factors, Rho-kinase is likely to phosphorylate multiple substrates including CRMP-2 and MLC, for the achievement of LPA-induced growth cone collapse.
Recently, we found that the long term overexpression of CRMP-2 disrupted the neuronal polarity and induced the formation of multiple axons in primary cultured hippocampal neurons. 2 These results suggest that CRMP-2 is involved in the axonogenesis. It is known that the Rho-Rho-kinase signaling pathway is activated by the extracellular signals and prevents neurite elongation. This raises the possibility that the Rho-Rho-kinase signaling pathway may link extracellular signals to CRMP-2-mediated axonogenesis.
In conclusion, the present study identified CRMP-2 as a novel substrate of Rho-kinase in brain. We also showed that the CRMP-2 phosphorylation by Rho-kinase plays a partial role in the regulation of growth cone morphology. The molecular mechanism of the growth cone regulation by Rho-kinase and CRMP-2 remains an important issue for future investigations.