Cdc2-mediated Schwann cell migration during peripheral nerve regeneration

Peripheral nerve fibers after injury, unlike those in the central nervous system, regrow towards their original target and can recover functionally. This ability is attributed to intrinsic neuronal properties and surrounding non-neuronal activities in which Schwann cells play a major supportive role. Schwann cells in the injured nerve area proliferate and migrate into the distal end forming the band of Büngner and support axonal regrowth (Fawcett and Keynes, 1990). Schwann cell migration, which also occurs at proximal end of the injury area, provides a guide for regenerating axons by interacting with nerve fibers or basal lamina (Williams et al., 1983; Daniloff, 1991; Guenard et al., 1992; Torigoe et al., 1996). Several studies showed that Schwann cell migration is crucial for successful axonal elongation (Torigoe et al., 1996; Anton et al., 1994a). Of the molecular factors responsible for Schwann cell migration, NGF and its Schwann cell receptor (low-affinity NGF receptor or LNGFR) and 4C5 Schwann cell surface antigen were reported to mediate Schwann cell migration on denervated sciatic nerve (Anton et al., 1994a; Yfanti et al., 2004). Oxidized galectin-1, which is produced from Schwann cells and neurons and functions as a cytokine, and extracellular matrix components, such as laminin (merosin) and products of chondroitin 6-sulphotransferase activity, were reported to promote axonal regeneration by Schwann cell migration (Fukaya et al., 2003; Anton et al., 1994b; Liu et al., 2006). These studies implicate the importance of the molecular interaction of Schwann cells with the environment, but the events within the migrating Schwann cell in response to injury are largely unknown.

Cdc2 (or cdk1) is a prototypical cyclin-dependent kinase that promotes G2-M phase transition in the cell cycle. The range of Cdc2 kinase substrates is broad, implying its importance for diverse cell functions including DNA replication and mitosis, dynamic regulation of spindle assembly and actin polarization (Ubersax et al., 2003). Moreover, Cdc2 in association with cyclin B2 mediates cell migration in prostate cancer cells (Manes et al., 2003). A function of Cdc2 and other cdk family proteins in the nervous system has begun to emerge recently. There have been extensive studies on the role of Cdk5 in neural migration during CNS development, axon guidance, neuronal degeneration and axonal regeneration (Dhavan and Tsai, 2001; Patrick et al., 1999; Namgung et al., 2004). Stimulation and activation of cdk family proteins including Cdc2 and Cdk2 in postmitotic neurons are related to neuronal apoptosis or degeneration (Rideout et al., 2003; Konishi and Bonni, 2003; Vincent et al., 1997). By contrast, induction of Cdk2 expression and activation in non-neuronal cells in the nervous system such as astrocytes, oligodendrocytes and Schwann cells was shown to be associated with proliferation (Tikoo et al., 2000; Tanaka et al., 1998; Belachew et al., 2002).

Given that Cdc2 activation mediates cell proliferation and migration in non-neuronal systems, the functional involvement of Cdc2 in the regenerating peripheral nerves in which Schwann cells proliferate and migrate would be expected. We have recently found that Cdc2 protein expression and its activity were upregulated in sciatic nerves undergoing axonal regeneration promoted by physical training (Seo et al., 2006). Here we explored Cdc2 function in Schwann cells associated with nerve regeneration, and found that Cdc2 expression was induced in Schwann cells and promoted Schwann cell migration, which is functionally related to the nerve regeneration process. We also present evidence of Cdc2-induced caldesmon activation during Schwann cell migration.

To investigate changes in Cdc2 expression, nerve segments proximal and distal to the crush site were prepared. Cdc2 protein was not detected in intact nerves, yet its expression increased strongly at 3 and 7 days post crush (d.p.c.) followed by a decrease at 14 d.p.c. (Fig. 1A). Longitudinal distribution of Cdc2 protein showed higher levels in the distal portion of the nerve at 7 d.p.c. compared with that at 3 d.p.c. (Fig. 1B). Similarly, Cdc2 mRNA was transiently elevated at 3 and 7 d.p.c. showing an increase in the distal segment at day 7 (Fig. 1C). Cdc2 mRNA expression in the nerve was examined further by FISH using a Cdc2 riboprobe. Cdc2 mRNA was clearly detected in the sciatic nerve prepared at 7 d.p.c. and mostly overlapped with S100β protein signals, which label Schwann cells (Fig. 1D). FISH in either intact nerve tissue or using a sense probe did not show any signal. Examination of Cdc2 protein in the injured nerve by immunofluorescence staining also revealed colocalization with S100β (Fig. 1E). These data suggest the transient induction of Cdc2 expression in Schwann cells of the injured sciatic nerve.

To investigate Cdc2 function in Schwann cells, we first examined whether Cdc2 expression was induced in cultured Schwann cells prepared from injured sciatic nerves. Cdc2 protein levels in the cultured Schwann cells, which were prepared from the sciatic nerve in rats that had undergone crush injury for 7 or 14 days (`injury-preconditioned' cells), were higher than those in Schwann cells from intact nerve (`non-preconditioned' cells) (Fig. 2A). In injury-preconditioned cells, increases in Cdc2 protein were further upregulated by treatment of serum and forskolin for cell proliferation. A portion of S100β-positive Schwann cell counts in the total cell population was increased to ∼90% by injury preconditioning or by inducing proliferation, as demonstrated previously (Morrissey et al., 1991). Double immunofluorescence staining analysis further showed that the number of Cdc2-positive cells in the total cell population or Schwann cell population was elevated by nerve injury and/or by serum and forskolin treatment (Fig. 2B), suggesting that Schwann cells are the major cell type inducing Cdc2 expression in response to nerve injury.

Schwann cells in the injured peripheral nerves can migrate into the injury site and promote axonal regeneration. Here, we examined Schwann cell migration in both explant culture of sciatic nerve segments and filter culture of cells from sciatic nerve trypsinates. Schwann cell migration was strongly induced in the explant culture of injury-preconditioned sciatic nerve (Fig. 3A) whereas no visible cell motility was observed in the control nerve. Analysis using microfilter chamber culture showed that the number of migrated Schwann cells was much higher in the injury-preconditioned group than non-preconditioned control group (Fig. 3B). Treatment with the cdk inhibitor roscovitine decreased Schwann cell migration in a dose-dependent manner (Fig. 3C), suggesting that cell migration depends on Cdc2 activity.

We further examined the role of Cdc2 activity in Schwann cell migration in vivo. After infection with adenovirus expressing wt- or dn-Cdc2 cDNA together with green fluorescence protein (GFP) reporter, Schwann cells were implanted into the injury site of the sciatic nerve. Infected Schwann cells were identified in the sciatic nerve sections by visualizing GFP. GFP-expressing Schwann cells were found at different distances along the distal area of the injured sciatic nerve, and mostly localized close to the nerve fibers (Fig. 4A). At 3 d.p.c. and 7 d.p.c. cells expressing wt-Cdc2 migrated further towards the distal end of the nerve compared with cells expressing dn-Cdc2 (Fig. 4B,C).

Caldesmon is an actin-binding protein that is regulated by phosphorylation by Cdc2 kinase (Yamashiro et al., 1991). Here we investigated whether caldesmon participated in Schwann cell migration in association with Cdc2 activation. Caldesmon phosphorylation was increased in the injury-preconditioned Schwann cells (Fig. 5A). Western blot analysis of phospho-caldesmon showed a positive band from Schwann cell lysate, which had been immunoprecipitated with anti-actin antibody and phosphorylated in vitro by Cdc2 kinase (Fig. 5B, lane 1), but the band intensity was decreased by addition of roscovitine or by depletion of Cdc2 enzyme in the kinase reaction (Fig. 5B, lanes 2 and 3). Caldesmon immunoprecipitated by actin was further used for in vitro Cdc2 kinase assay. As shown in Fig. 5C, phosphorylated caldesmon was detected, but the addition of cdk inhibitor roscovitine or histone H1 as a competitive substrate to the reaction decreased caldesmon phosphorylation by Cdc2, indicating the reaction specificity of Cdc2 phosphorylation of caldesmon. In injury-preconditioned Schwann cells, cell counts for phospho-caldesmon-positive immunostaining were significantly decreased in the cell population transfected with dn-Cdc2 cDNA compared with empty vector transfection (Fig. 5D). By western blot analysis, we confirmed the downregulation of phospho-caldesmon levels in Schwann cells transfected with dn-Cdc2. We then investigated whether Cdc2 activity affected caldesmon interaction with cytoskeletal proteins present in the periphery of the cell. In injury-preconditioned Schwann cells, caldesmon was concentrated in the central area of the cell. The treatment of roscovitine changed the localization of caldesmon signal to the periphery (Fig. 5E), suggesting that caldesmon interacts with the peripheral cytoskeleton and is released by Cdc2 phosphorylation.

To determine the effects of Cdc2-mediated caldesmon phosphorylation on Schwann cell migration, cells were transfected with mutant forms of Cdc2 and caldesmon. A plasmid construct expressing mutant caldesmon in which all seven phosphorylation sites are missing has been developed (Yamashiro et al., 2001) and was used here to inhibit the Cdc2-mediated caldesmon pathway. In injury-preconditioned Schwann cells, which were expected to express Cdc2, the number of migrated cells was significantly lower when transfected with either dn-Cdc2 or caldesmon 7th mutant than vector control (Fig. 6A). In non-preconditioned Schwann cells, migration was significantly facilitated by transfection with wt-Cdc2 DNA compared with control vector (Fig. 6B). However, cell migration was decreased by cotransfection with Cdc2 and caldesmon 7th mutant DNA, suggesting interference of endogenous caldesmon phosphorylation by transfected mutant caldesmon. Expression of wild-type caldesmon alone did not increase cell migration, suggesting the possible requirement for its activation by Cdc2.

Cdc2 upregulation in migrating Schwann cells suggests functional association with axonal regeneration. To delineate the role of Cdc2 in axonal regeneration after injury, neurite outgrowth of DRG sensory neurons in the presence of Schwann cells was investigated. Sciatic nerves with injury preconditioning facilitated neurite outgrowth of DRG sensory neurons as demonstrated previously (Fig. 7A,B) (Smith and Skene, 1997). We found that cultured sensory neurons prepared from DRG without injury preconditioning showed extensive neurite outgrowth when co-cultured with injury-preconditioned Schwann cells. However, incubation with roscovitine significantly inhibited the development of neurite branch points. Roscovitine also caused a 56% decrease in neurite length but this was not statistically significant (P=0.57, n=4; Fig. 7B).

Our data provide new evidence that Cdc2 mediates Schwann cell migration in the injured peripheral nerve and facilitates axonal regeneration. We demonstrate in vitro and in vivo that migration of injury-preconditioned Schwann cells was significantly delayed by Cdc2 inhibition, and that they regained motility by Cdc2 overexpression. Our data further showed that Cdc2 phosphorylated caldesmon in Schwann cells, and blockade of the Cdc2-caldesmon pathway suppressed cell migration. Neurite outgrowth of cultured DRG sensory neurons was largely improved by co-culture with injury-preconditioned Schwann cells.

We have recently reported increased activity of Cdc2 kinase during axonal regeneration of the injured sciatic nerve facilitated by treadmill training in the rat (Seo et al., 2006). Here, we sought the physiological role of Cdc2 in Schwann cells, which had been preconditioned to activate the regeneration process. Cdc2 synthesis was increased in Schwann cells that were prepared from sciatic nerves that had undergone crush injury for 7–14 days. The change in Cdc2 protein levels showed a similar time course to that seen in the injured sciatic nerve. Measurement of Schwann cell migration in the explant culture and microfilter chamber showed strong induction by injury preconditioning of the sciatic nerve. Inhibition of Cdc2 activity by roscovitine or by expression of dn-Cdc2 mutant resulted in decreased cell migration, indicating the requirement of Cdc2 activity for cell migration. A role of Cdc2 function in migration was further demonstrated in Schwann cells that were infected with Cdc2-expressing viral vectors and implanted into the sciatic nerve. Since transplanted Schwann cells expressing wt-Cdc2 migrated faster toward the distal end than those expressing dn-Cdc2, we presume that endogenous Schwann cells in which Cdc2 expression was increased by nerve crush might conduct similar motility along the Wallerian degenerated trunk (Fukaya et al., 2003; Torigoe et al., 1996).

Caldesmon is present as different isoforms in actin- and calmodulin-binding proteins of smooth muscle and non-muscle cells. Its binding to actin can be regulated by phosphorylation by many kinases such as Erk, p38, Cdc2, and calmodulin-dependent protein kinase II (Childs et al., 1992; Ikebe and Reardon, 1990; Yamashiro et al., 1991; Mirazapoiazova et al., 2005). Non-muscle 83 kDa caldesmon is released from actin filaments when phosphorylated by Cdc2 kinase (Yamashiro et al., 1991). Increased levels of αvβ3 integrin in prostate cancer cells upregulate Cdc2 phosphorylation of caldesmon in events that lead to increased cell migration (Manes et al., 2003). Since cell proliferation and migration are the key events for carcinogenesis, the study by Manes et al. raises an interesting issue that a single molecule controlling the cell cycle can also regulate cell migration (Juliano, 2003). Since Schwann cells undergo the stages of proliferation and migration in the injured peripheral nerve before differentiation stage for myelination, the cell-cycle protein Cdc2 could be functionally linked to migration. In Schwann cells, we found caldesmon phosphorylation by Cdc2 and changes in its subcellular localization by roscovitine treatment. Previous studies show that caldesmon bound to actin filaments at the cell periphery is released by its phosphorylation, and thus modulates cell migration (Ishikawa et al., 1998; Manes et al., 2003; Mirzapoiazova et al., 2005). In our study, addition of roscovitine to injury-preconditioned Schwann cells changed the subcellular localization of caldesmon protein to the periphery, suggesting a dynamic regulation of its interaction with actin via Cdc2. We further found that the blockade of Cdc2-mediated caldesmon pathway by expression of caldesmon 7th mutant cDNA significantly reduced cell migration. Caldesmon 7th mutant was constructed by substituting all Cdc2 phosphorylation site and its effects have been well demonstrated in the regulation of actin and calmodulin, and mitosis (Yamashiro et al., 1991; Yamashiro et al., 1995; Yamashiro et al., 2001). Thus, Cdc2 phosphorylation of caldesmon in Schwann cells appears to regulate cell motility via modulation of the interaction with actin.

Our data suggest that Cdc2-mediated Schwann cell migration may be involved in axonal elongation. DRG sensory neurons when preconditioned by sciatic nerve injury showed enhanced neurite outgrowth compared with non-preconditioned control (Smith and Skene, 1997). In the present study, non-preconditioned DRG sensory neurons displayed significantly increased neurite extension by co-culturing with injury-preconditioned Schwann cells, which was then attenuated by roscovitine treatment. Facilitated Schwann cell migration via the activation of Cdc2 signaling pathway might contribute to guide neurite extension in culture as demonstrated in regenerating axons in vivo (Torigoe et al., 1996). Although the present data implicate a potential role of Cdc2 in axonal elongation via Schwann cell migration, the importance of Cdc2-induced Schwann cell proliferation deserves to be mentioned. Cdk2 activity, which is necessary for the transition from G1 to S phase, is important for Schwann cell proliferation (Tikoo et al., 2000), and increased Cdk2 protein levels were observed in the injured sciatic nerves (T.B.S. and U.N., unpublished observation). While these data suggest the possible involvement of Cdk2 in peripheral nerve regeneration, Cdc2 activity could be required for the complete cell-cycle progression in Schwann cells. Indeed, our data indicate that Cdc2 activity is positively associated with Schwann cell proliferation (see Fig. 2). One possible mechanism explaining the dual roles of Cdc2 for cell migration and cell-cycle progression would be Cdc2 displacement and differential activation by cyclin B1 and B2 between cytoplasm for caldesmon phosphorylation and nucleus for M-phase transition. In human cultured cells, cyclin B2 is found in the Golgi complex and cell membrane, and unlike cyclin B1, is not located in the nucleus at prophase (Jackman et al., 1995). Whether cyclin B1 and cyclin B2 are similarly distributed in Schwann cells and activate Cdc2 kinase to regulate cell proliferation and migration at different stages of the cell cycle remains to be determined.

In summary, we found that upregulation of Cdc2 protein in Schwann cells is important for cell migration and has a growth-promoting activity for regenerating axons. Together with the present finding, recent studies on Cdc2-mediated neuronal apoptosis demonstrate a novel function of Cdc2 in the nervous system (Konishi and Bonni, 2003; Konishi et al., 2002). Moreover, we found upregulation of Cdc2 activity in a non-neuronal cell population of the injured spinal cord, which positively correlated with the extent of axonal sprouting of injured neurons (T.B.S. and U.N., unpublished observation), which implicates the pathophysiological significance of Cdc2 in CNS function. Growing bodies of evidence show that non-neuronal cells such as astrocytes in the CNS could actively participate in the recovery of injured spinal cord via enhanced migration (Okada et al., 2006) and information processing by directly communicating between them via chemical signals such as ATP, Ca²⁺ and leukemia inhibitory factors (LIF) for myelination (Ishibashi et al., 2006; Fields and Stevens-Graham, 2002). Since non-neuronal cells in the nervous system function in diverse ways, further investigation of Cdc2 in this system could provide insight into the regulatory mechanism of nerve regeneration.

The sciatic nerve of Sprague-Dawley rats (7-8 weeks old, male) was exposed on the middle thigh and crush injury on the nerve was given by holding twice with forceps for 30 seconds, as described previously (Seo et al., 2006). All procedures were in strict accordance with the NIH guide for the care and use of laboratory animals and approved by the Committee on Use of Live Animals for Teaching and Research at Daejeon University. Nerves were prepared 0-14 d.p.c. for western blot analysis, immunohistochemistry, or primary cell culture.

For Schwann cell and DRG sensory neuron culture, sciatic nerve and DRG at lumbar 4-5 were dissociated by treatment with 125 U/ml type XI collagenase (Sigma, St Louis, MO) in DMEM for 80 minutes at 37°C, and washed twice with DMEM. Cells were treated with 0.5 mg/ml type SII trypsin for 15 minutes and followed by inhibition reaction for 5 minutes in 1 mM EDTA, 100 μg/ml of soybean trypsin inhibitor and 40 μg/ml of DNase I. Cells (1×10⁶ cells per dish) were then plated onto 12-mm coverslips (Bellco, Glass Inc. Vineland, NJ) precoated with 0.01% poly-L-ornithine (Sigma, St Louis, MO) and laminin (0.02 mg/ml, Collaborative Research, Bedford, MA). Cells were cultured for 12 hours, and changed to DMEM containing 10% serum [5% fetal bovine serum (Gibco, Australia) plus 5% horse serum] and 2 mM glutamine and 1% penicillin-streptomycin. Cells were incubated for 48 hours before the harvest for immunofluorescence staining or western blot analysis. In experiments for roscovitine inhibition of Cdc2 activity, injury-preconditioned Schwann cells were cultured in the presence of 5 or 10 μM roscovitine or the equivalent volume of DMSO vehicle for 6 hours before harvest.

For DRG sensory neuron and Schwann cell co-culture, DRG sensory neurons (1.5×10² cells/well) were cultured for 24 hours and Schwann cells (1×10⁴/well) were added. The co-cultures were maintained in 500 μl DMEM supplemented with 10% serum in the presence of 10 μM roscovitine or equivalent volume of 100% of DMSO for 2 days before cell harvest. After immunofluorescence staining with anti-βIII-tubulin antibody (1:200, TUJ1, rabbit polyclonal, Covance, Berkeley, CA), digital images of neuronal process were captured and transferred to Adobe Photoshop (version 5.5). The number and length of neurite processes exhibiting clear outgrowth (longer than cell body size) from the cell body were analyzed by using i-Solution software (Image and Microscope Technology, Goleta, CA). Mean branch points and neurite length were determined by analyzing at least 50 sensory neurons that were randomly selected in each experiment.

After preliminary culture for 3 days, Schwann cells (1×10⁴) were seeded on Costar filter chambers (bottom diameter: 6.5 mm, 8 μm pore size, Costar, Cambridge, MA) which had been coated with poly-L-ornithine and laminin mixture. Cells were allowed to grow on the surface of the chamber and to migrate to a precoated coverslip (12 mm in diameter) beneath the chamber for 2 days. Migrated cells were fixed with 4% paraformaldehyde and 4% sucrose in PBS for immunofluorescence staining with anti-S100β antibody (1:200, Dako, Denmark) or Cresyl Violet staining (0.2% cresyl violet in water for 25 minutes). Digital images of stained cells were measured and transferred to Photoshop. The number of cells in each image was counted by using i-Solution software program. A similar procedure was also applied for digital image collection and quantitative analysis from the explant culture (see below).

Sciatic nerve with or without crush injury was placed in ice-cold DMEM, and placed on 12-mm coverslips precoated with a mixture of poly-L-ornithine and laminin. Tissue was incubated in 200 μl culture medium (DMEM containing 5% fetal bovine serum, 5% horse serum and 2 mM glutamine and 1% penicillin-streptomycin) for 1 hour and supplemented with 300 μl medium. The explant culture was replaced with 500 μl of fresh medium 16 hours later and incubated for 48 hours. Then the culture was fixed with 4% paraformaldehyde/4% sucrose for immunofluorescence staining with anti-S100β antibody (1:200, Dako) and anti-Cdc2 antibody (1:100, p34, mouse monoclonal, Santa Cruz Biotechnology, Santa Cruz, CA). Average migration distance in each explant (30-100 cells per explant) was determined by measuring the distance between explant border and stained Schwann cells.

Recombinant pRcCMV5 plasmids encoding wild-type (wt) caldesmon or 7th mutant caldesmon in which all seven Cdc2 phosphorylation sites were replaced with Ala were kindly provided by S. Matsumura (Rutgers University, Piscataway, NJ). pRcCMV5-Cdc2wt-HA and pRcCMV5-Cdc2dn-HA were from Sander van den Heuvel (Harvard University, Boston, MA). Transient transfection into Schwann cells (1×10⁴ cells per 24-well plate) was performed using 3 μg pCMV5-Cdc2wt-HA, pCMV5-Cdc2dn-HA, pRcCMV-caldesmon wt, pRcCMV-7th caldesmon or pCMV5 together with 1 μg pmaxGFP plasmid, by the calcium phosphate protocol as described (Namgung and Xia, 2000). With this procedure, cotransfection efficiency was routinely higher than 70%. Cells were harvested 48 hours later. For migration analysis, 1×10⁴ cells were plated on the filter chambers (Costar) and harvested 2 days later for counting GFP-positive cells on the coverslip. Adenoviral vector (Ad) expressing HA-tagged Cdc2 wt or Cdc2 dn cDNA was prepared using AdEasy system (Stratagene, La Jolla, CA). Ad-Cdc2 DNA was subcloned into pShuttle-CMV vector and cotransformed into E. coli BJ5183 cells with Adeasy-1 vector. Selected recombinant plasmids were transfected into HEK293 cells and replication-deficient recombinant adenoviruses were generated. Adenoviruses containing Cdc2 were isolated using a commercially available adenovirus purification kit (Cell Biolabs, San Diego, CA) following the manufacturer's instructions. Purified viruses were aliquoted and stored at –80°C. Viral titer (PFU per milliliter: 1×10¹¹) was determined using the agarose overlay method (Becker et al., 1994).

Schwann cells (1×10⁴) were mixed with 40 μl adenoviral stock solution and harvested 2 days later. Infected Schwann cell suspension (1×10⁴ cells in 5 μl) was injected at the injury site of the sciatic nerve using the microsyringe (Microliter Spritze, Innovative Labor System, Germany). An average of 75% of the Schwann cells used for transplantation expressed GFP, and 59% and 63% of the GFP-positive cells expressed wt-Cdc2 and dn-Cdc2 in viral constructs respectively when HA-tagged Cdc2 expression was analyzed by immunofluorescence staining with anti-HA antibody (1:500, Santa Cruz Biotechnology). The implanted cells were allowed to migrate along the nerve for 3 or 7 days, and nerve sections (20 μm thickness) were prepared for immunofluorescence staining with anti-Neurofilament 200 (NF-200) antibody (1:200, rabbit, Sigma). For quantitative analysis on migrated cells, images of GFP-positive cells in the nerve sections were captured by a digital camera, and transferred to Photoshop. The numbers of infected Schwann cells in a defined area (470 μm × 590 μm) 0 mm, 1.5 mm, 3 mm, 5 mm and 7 mm from the injury site were counted.

Total RNA was extracted from sciatic nerve by Easy Blue (Intron Biotechnology, Korea), and was used for reverse transcriptase reaction (MMLV RT, Promega, Madison, WI) for 1 hour at 37°C, followed by PCR amplification of Cdc2 or actin DNA for 30 cycles using Taq DNA polymerase (Takara, Japan). RT-PCR was performed as suggested by the manufacturer. The primer sequences for Cdc2 were 5′-ATCGGAGAAGGGACTTATGG-3′ as a forward primer and 5′-TGCAGGGATCTACTTCTGG-3′ as a reverse primer, and the sequences for actin were 5′-CACACTGTGCCCATCTATGA-3′ as a forward primer and (5′-TACGGATGTCAACGTCACAC-3′) as a reverse primer. The amplified PCR products were 677 bp and 409 bp for Cdc2 and actin respectively.

Cross sections (20-μm thick) of sciatic nerves cut on a cryostat and mounted on positively charged Superfrost/Plus slides (Fisher Scientific, Pittsburgh, PA) were fixed with 4% paraformaldehyde for 10 minutes, rinsed twice in 0.1% DEPC water, washed in 0.1 M triethanolamine (TEA, pH 8.0), and treated with 0.25% acetic anhydride in 0.1 M TEA for 10 minutes. After washing twice in 2× SSC, sections were dehydrated in an ethanol series (50%, 70%, 90%), and hybridized overnight at 58°C using 3 ng/μl fluorescein-UTP (Roche, Germany)-labeled Cdc2 riboprobes in hybridization buffer (0.3 M NaCl, 50 mM Tris-HCl pH 8.0, 5 mM EDTA, 50% formamide, 1× Denhardt's, 10% dextran sulphate, 10 mM DTT). pcDNA3-Cdc2 was linearized with HindIII and EcoRI digestion for anti-sense and sense probe preparation respectively. Sense and anti-sense RNA probes were prepared by fluorescence labeling reaction with RNA polymerase SP6 and T7 respectively as suggested by the manufacturer (Promega, Madison, WI). After hybridization, slides were rinsed for 10 minutes in 2× SSC containing 10 mM β-mercaptoethanol and 1 mM EDTA, washed for 30 minutes in RNase solution (0.5 M NaCl, 10 mM Tris-HCl pH 8.0, 20 μg/ml RNase), then washed for 10 minutes with 2× SSC. After incubating for 2 hours at 60°C in 0.1× SSC containing 10 mM β-mercaptoethanol and 1 mM EDTA, sections were rinsed in 0.5× SSC and dehydrated in an ethanol series (50%, 70%, 90% containing 0.3 M ammonium acetate). Sections after hybridization reaction were permeabilized in 0.5% nonidet P-40 solution and subsequently treated for immunofluorescence staining with anti-S100β primary antibody (1:200, Dako) and Rhodamine-goat anti-rabbit secondary antibody (1:400; Molecular Probes, Eugene, OR) (see below). Fluorescence images were analyzed by confocal laser-scanning microscopy (LSM 510, Carl Zeiss, Germany).

Nerve segments or cultured Schwann cells in 50-200 μl of Triton lysis buffer (20 mM Tris-HCl, pH 7.4, 137 mM NaCl, 25 mM β-glycerophosphate, pH 7.14, 2 mM sodium pyrophosphate, 2 mM EDTA, 1 mM Na₃VO₄, 1% Triton X-100, 10% glycerol, 5 μg/ml leupeptin, 5 μg/ml aprotinin, 3 μM benzamidine, 0.5 mM DTT, 1 mM PMSF) were sonicated, and the supernatant was removed after centrifugation at 14,000 g for 10 minutes at 4°C. Antibodies used are: anti-Cdc2 antibody (1:1000, p34, mouse monoclonal, Santa Cruz Biotechnology), anti-phospho-caldesmon antibody (1:1000, rabbit polyclonal, Santa Cruz Biotechnology), anti-actin antibody (1:5000, clone no.C4, mouse monoclonal, MP Biomedicals, Aurora, OH), and horseradish peroxidase (HRP)-conjugated secondary antibodies (1:1000, goat anti-rabbit; Santa Cruz Biotechnology, or sheep anti-mouse; Amersham Biosciences, France). Protein (10 μg) was used for western analysis as described previously (Seo et al., 2006).

Cdc2 kinase assay was performed as described previously with some modifications (Namgung and Xia, 2000; Qiao et al., 2006). Anti-actin antibody (3 μl; clone no. C4, mouse monoclonal, MP Biomedicals, Aurora, OH) was incubated in 250 μl PBS containing 8% protein A Sepharose for 16 hours at 4°C and beads were washed once with cold Triton lysis buffer. Cell lysate (200 μg of protein in 100 μl reaction) was immunoprecipitated with anti-actin antibody conjugated to protein A Sepharose for 3 hours at 4°C. The beads were washed twice each with Triton lysis buffer and by cold kinase buffer (25 mM HEPES, pH 7.4, 25 mM β-glycerophospohate, pH 7.14, 25 mM MgCl₂, 0.1 mM Na₃VO₄, 0.5 mM DTT). The immunocomplexes were precipitated by brief centrifugation (1500 g, 1 minute) and incubated with 30 μl of kinase reaction buffer containing 1 nmole ATP, 10 μCi of [γ-³²P]ATP (3000 μCi/ml, Amersham), and 1 unit of Cdc2 recombinant kinase (New England Biolabs, Beverly, MA) for 30 minutes at 30°C. In the control reaction, 2 μg histone H1 or 10 μM roscovitine was included in the reaction mixture. Samples were boiled, and the supernatants were resolved by 10% SDS-PAGE. The gels were dried and analyzed by autoradiography. For the purpose of western analysis of kinase reaction products, the immunocomplexes were incubated with kinase reaction mixture including 1 nmole ATP and 1 unit of Cdc2 recombinant kinase, and were used for western blot analysis using anti-phospho-caldesmon antibody (1:1000, rabbit polyclonal, Santa Cruz Biotechnology, Inc.).

Nerve sections or cultured cells were fixed with 4% paraformaldehyde and 4% sucrose in PBS at room temperature for 40 minutes, permeabilized with 0.5% nonidet P-40 in PBS, and blocked with 2.5% horse serum and 2.5% bovine serum albumin for 4 hours at room temperature. The staining reaction was performed by incubating with primary antibodies including anti-Cdc2 antibody (1:100, p34, mouse monoclonal, Santa Cruz Biotechnology), anti-S100β antibody (1:200, Dako), anti-caldesmon antibody (1:200, mouse monoclonal Sigma), anti-phospho-caldesmon antibody (1:200, rabbit polyclonal, Santa Cruz Biotechnology), anti-neurofilament 200 antibody (1:200, rabbit, Sigma), and anti-βIII-tubulin antibody (1:200, TUJ1, rabbit polyclonal, Covance, Berkeley, CA), followed by Fluorescein-labeled goat anti-mouse (1:400; Molecular Probes) or Rhodamine-labelled goat anti-rabbit secondary antibodies (1:400; Molecular Probes) in 2.5% horse serum and 2.5% bovine serum albumin for 1 hour at room temperature. Cellular nuclei were stained with 2.5 μg/ml of Hoechst 33258 (bis-benzimide, Sigma) for 10 minutes before the final washing with 0.1% Triton X-100 in PBS, and the sections or cells were coverslipped with gelatin mount medium. We always included control sections treated with secondary antibody alone, which usually did not have any visible images. In cases where the nonspecific signals were high, data from the same experiments were not analyzed further. Samples were viewed with a Nikon fluorescence microscope, and the images were captured using Nikon camera. The merged images were produced using layer-blending mode options of Adobe Photoshop (Version 5.5).

Data were presented as mean ± s.e.m. The mean numbers of data in individual groups were compared by one-way ANOVA followed by Tukey test (SPSS computer software version 12.0), and statistically significant differences were reported as ^*P<0.05, ^**P<0.01, ^***P<0.001.

We are grateful to S. Matsumura (Rutgers University) and Sander van den Heuvel (Harvard Medical School) for providing the plasmids and David Elzi (University of Washington) for critical reading of the manuscript. This work was supported by the Korea Research Foundation (C00180 and C00161).