Mechanobiology June 26th - June 2nd 2016

Mechanobiology: June 26th  - June 2nd 2016

Summary

Oligodendrocyte precursor cells (OPCs) differentiate into oligodendrocytes (OLs) in order to form myelin, which is required for the rapid propagation of action potentials in the vertebrate nervous system. In spite of the considerable clinical importance of myelination, little is known about the basic molecular mechanisms underlying OL differentiation and myelination. Here, we show that cyclin-dependent kinase (Cdk) 5 is activated following the induction of differentiation, and that the Cdk5 inhibitor roscovitine inhibits OL differentiation. The complexity of the OL processes is also diminished after knocking down endogenous Cdk5 using RNAi. We also show that the focal adhesion protein paxillin is directly phosphorylated at Ser244 by Cdk5. Transfection of a paxillin construct harboring a Ser244 to Ala mutation dramatically inhibits its morphological effects. Importantly, phosphorylation of paxillin at Ser244 reduces its interaction with focal adhesion kinase (FAK). Taken together, these results suggest that phosphorylation of paxillin by Cdk5 is a key mechanism in OL differentiation and may ultimately regulate myelination.

Introduction

Oligodendrocytes (OLs) are the myelinating glial cells of the central nervous system (CNS). The myelin sheath wraps concentrically around axons and maximizes nerve conduction velocity. OLs are generated by the oligodendrocyte precursor cells (OPCs), which are derived from precursors in germinal areas of the CNS (Rowitch, 2004). During CNS development, these precursors proliferate and migrate into subcortical white matter, where they extend complex branched processes, and then initiate myelination (Sherman and Brophy, 2005). In vitro studies demonstrate that OLs develop through several stages (Barres et al., 1992). These stages are characterized by distinct morphological changes, cell surface antigens and different responses to growth factors (Baumann and Pham-Dinh, 2001; Pfeiffer et al., 1993). Since the molecular mechanism(s) that control OL development and function are still largely unknown, it is important to identify the signaling molecule and the pathway leading to each of these developmental stages.

Cyclin-dependent kinase (Cdk) 5 is a proline-directed serine/threonine kinase, which is enriched in neuronal tissues and has diverse biological functions. Since the Cdk5 activators, p35 and p39, are expressed solely in the nervous system, Cdk5 activity is specific as well as restricted to the nervous system (Lew et al., 1994; Tsai et al., 1994). Mice lacking Cdk5 exhibit a severely disrupted laminar structure of the cerebral cortex, which stems from the inability of cortical neurons to migrate (Ohshima et al., 1996). A severe defect of axon guidance is also observed in Cdk5–/– mice (Gilmore et al., 1998). Wheras Cdk5–/– mice display neuronal defects in early developmental stages, little is known about the role of Cdk5 in OL development.

Paxillin is a 68 kDa focal-adhesion-associated protein involved in basic biological functions, such as cell adhesion, cell spreading, migration and neurite outgrowth (Brown and Turner, 2004). Paxillin contains five tandem repeats of a 13-amino acid sequence, called the LD motifs, near the N-terminal and four LIM domains in the C-terminus (Brown and Turner, 2004). Paxillin translates various extracellular stimuli into specific intracellular signals by binding to a number of signaling molecules (Brown and Turner, 2004). Furthermore, the N-terminal half of paxillin contains many putative tyrosine phosphorylation sites, of which two main tyrosine phosphorylation sites (Tyr31 and Tyr118) have recently been identified (Webb et al., 2005). Adhesion-signal-dependent tyrosine phosphorylation leads to docking of Src homology (SH2)-domain-containing proteins, such as CrkII (Brown and Turner, 2004). There are also some serine and threonine phosphorylation sites recognized by proline-directed protein kinases (Webb et al., 2005). For example, Jun N-terminal kinase (JNK) phosphorylates paxillin at Ser178 to regulate migration of the rat bladder tumor epithelial cell line NBT-II (Huang et al., 2003) and mediate neurite extension in the mouse neuroblastoma cell line NIE-115 (Yamauchi et al., 2006). p38 mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase (ERK) phosphorylate paxillin at Ser83 to regulate neurite outgrowth in the rat pheochromocytoma PC12 cell line and induce morphological changes in mouse epithelial mIMCD-3 cells, respectively (Huang et al., 2004; Ishibe et al., 2004). ERK1/2 and glycogen synthase kinase (GSK) 3β-mediated phosphorylation of paxillin at Ser126 and Ser130 is involved in nerve growth factor (NGF)-induced neurite outgrowth of PC12 cells (Cai et al., 2006). Therefore, it is thought that paxillin represents a point of convergence in the signals controlling cell morphology. The roles of paxillin in these basic biological phenomena are also illustrated by the embryonic lethality of paxillin–/– mice (Hagel et al., 2002).

We demonstrate that roscovitine, an inhibitor of Cdk5, significantly reduces the complexity of the morphological changes in primary OLs and in FBD-102b cells, a mouse oligodendrocyte precursor cell line. FBD-102b cells are derived from p53–/– mouse fetal brains and possess the characteristics of OPCs. The cells differentiate in a fashion similar to that of OPCs, by extending multiple branched processes and increasing myelin antigen markers upon differentiation (Horiuchi and Tomooka, 2006). Furthermore, in primary OLs, as well as FBD-102b cells, Cdk5 activity is increased after induction of differentiation. We also describe that Cdk5 regulates OL differentiation through the direct phosphorylation of paxillin at Ser244. These results suggest that Cdk5 is one of the regulators of OL differentiation through the direct phosphorylation of the focal adhesion protein paxillin, and provide new insights into the mechanisms underlying OL differentiation. Furthermore, our results show that FBD-102b cells might be a good in vitro model to analyze signal transduction during OL differentiation.

Results

Cdk5 mediates morphological differentiation of oligodendrocyte precursor cells

The oligodendrocyte precursor cell line FBD-102b (Horiuchi and Tomooka, 2006) was established from fetal brains of p53-deficient mice and maintains OPC properties in the presence of 10% FBS. When cultured in a medium containing reduced serum levels, FBD-102b cells differentiate into OLs with highly branched processes that contain myelin basic protein (MBP). Thus, the FBD-102b cell line seems to be a suitable model system for analyzing the molecular mechanisms involved in OL differentiation in vitro. Fig. 1A illustrates the morphology of the FBD-102b cells at different times following induction of differentiation. To quantify cell differentiation in detail, we divided the OL cells into three categories according to the complexity of their processes; low (two or less primary processes longer than a cell body, minimal development of secondary and tertiary processes), medium (three or more primary processes, moderate development of secondary and tertiary processes), and high (five or more primary processes, extensive development of secondary and tertiary processes) (Sperber and McMorris, 2001). Fig. 1B illustrates a representative example of each category defined by complexity. Before induction of differentiation, most of cells show only less than two short or unbranched processes. The cells were stained with antibody against NG2, which recognizes unipolar or bipolar, proliferative OPCs (Fig. 1C). At 2 days after induction of differentiation, cells typically possess multiple long and more highly branched processes. However, approximately 20% of the cells still display short and less branched processes. At 4-6 days in culture, cells begin to show morphologies with much longer and more highly branched processes. These cells were stained with an antibody against MBP, whose expression is the hallmark of mature OLs; however, differentiated FBD-102b cells do not form web-like myelin structures (Fig. 1C). As shown in Fig. 1D, the protein levels of MBP and myelin proteolipid protein 1 (PLP1) increased 2-6 days after induction of differentiation. Similar to the expression of myelin proteins, mRNA levels of MBP and PLP1 were also elevated 2-6 days after induction of differentiation (Fig. 1E). Together with the results from a previous report by Horiuchi et al. (Horiuchi et al., 2006), FBD-102b cells exhibit OPC properties and have the potential to differentiate into OLs.

Cdk5 is enriched in developing neuronal tissues and is known to be one of the major regulators of cytoskeletal proteins. Since OL differentiation involves dynamic morphological changes driven by cytoskeletal rearrangements (Asou et al., 1995), we asked whether Cdk5 participates in the process of OL differentiation. We found that roscovitine blocks the expression of MBP and PLP1 proteins and mRNA in FBD-102b cells (Fig. 1D,E). In addition, this inhibitor dramatically reduced the complexity of the processes in FBD-102b cells (Fig. 1F). To further confirm the importance of Cdk5 in OL differentiation, we used small interfering RNA (siRNA) to knockdown endogenous Cdk5 in FBD-102b cells. Cdk5 was markedly downregulated by transfection of siRNA targeting Cdk5, whereas the expression of paxillin and β-actin were unaffected (Fig. 1G). Similar to the effect of roscovitine observed in FBD-102b cells, knockdown of Cdk5 reduced the number of processes in cells transfected with Cdk5 siRNA (2.1±0.6) compared with those transfected with control siRNA (3.6±0.7) (see Fig. 1G), suggesting that Cdk5 is required for differentiation of FBD-102b cells into OLs.

We next examined the role of Cdk5 in the differentiation of primary OLs. OPCs were freshly isolated from embryonic rat brains. After two passages, they were cultured on non-coated dishes for 2 days. The representative morphology of each category is shown in Fig. 2A. OPCs were then allowed to differentiate for 3-5 days in differentiation medium with or without roscovitine. After changing to differentiation medium the OPC processes became highly branched and abundant with web-like myelin structures (Fig. 2A,B). These processes and MBP-positive structures were markedly suppressed in OPCs cultured in roscovitine-containing differentiation medium (Fig. 2B-D). In addition, roscovitine reduced the protein levels of MBP and PLP1 (Fig. 2E) but failed to completely inhibit MBP expression, suggesting that inhibition of Cdk5 activity attenuates differentiation rather than leading to the inhibition of differentiation. Furthermore, we infected primary OPCs with a retrovirus expressing both GFP and a short-hairpin RNA (shRNA) sequence targeting Cdk5 or its activator p35 to knockdown the expression of Cdk5 or p35, respectively. The expression of Cdk5 or p35 was specifically downregulated by infection with these shRNAs, whereas expression of control proteins was unaffected (Fig. 2F). The knockdown of Cdk5 or p35 also reduced the formation of the web-like myelin structures (21.6±5.9% or 20.8±2.6% in cells infected with the control shRNA as compared with 5.9±0.9% or 13.5±2.0% in cells infected with Cdk5 or p35 shRNA, respectively, Fig. 2G,H). Taken together, these results indicate a crucial role of Cdk5 in OL differentiation.

Fig. 1.

Cdk5 is necessary for the differentiation of FBD-102b cells. (A) After induction of differentiation, FBD-102b-cell morphology was observed using phase-contrast microscopy. Representative images of cells are shown. (B) FBD-102b cells were randomly selected and assigned to one of three categories of complexity; low (two or less primary processes longer than a cell body, minimal development of secondary and tertiary processes), medium (three or more primary processes, moderate development of secondary and tertiary processes), and high (five or more primary processes, extensive development of secondary and tertiary. The percentagees of the cells in each category and the representative images of cells are shown. (C) FBD-102b cells were stained with antibodies against NG2 (green) or MBP (red) before or 4 days after induction of differentiation. (D) After induction of differentiation, FBD-102b cells treated with or without 10 μM roscovitine were lysed and immunoblotted with antibodies against MBP, PLP1 and β-actin. (E) RT-PCR analyses for MBP, PLP1, and β-actin were performed on differentiating cells treated with or without 10 μM roscovitine. (F) Representative images of cells treated with or without 10 μM roscovitine 4 days after induction of differentiation. The percentages of the cells in each category are shown in the bar graph. (G) The number of processes of the cells transfected with siRNAs for Cdk5 and luciferase is shown (left). Lysates from the transfected cells were immunoblotted with antibodies against Cdk5, paxillin and β-actin (right). Data were evaluated using Student's t-test (*P<0.01).

Cdk5 is activated following induction of differentiation

To further investigate the role of Cdk5 in OL differentiation, we examined the expression of Cdk5 following induction of differentiation of FBD-102b cells and primary OLs. After induction in FBD-102b cells, Cdk5 protein and mRNA expression were increased (Fig. 3A). In addition to its expression, we measured the bona fide activity of Cdk5. Using an anti-Cdk5 antibody, we immunoprecipitated endogenous Cdk5 from cell lysates and incubated with ATP in a reaction buffer containing histone H1 as the substrate. The immuno-complexes were collected and immunoblotted with an antibody that specifically recognizes a phosphorylated consensus Cdk5 motif. After induction of differentiation, the kinase activity of Cdk5 in FBD-102b cells increased 2.8±0.2 and 4.1±0.6 times at 2 and 4 days, respectively (Fig. 3B). Similarly, in primary OLs, the level of Cdk5 protein began to increase at day 3, and leveled off at day 4 after induction of differentiation (Fig. 3C). Phosphorylated histone H1 also gradually increased following differentiation of primary OLs (Fig. 3D), consistent with Cdk5 expression and indicative of the role of Cdk5 in OL differentiation. Tang et al. have reported that the activity of Cdk5 is increased without an increase in Cdk5 protein levels (Tang et al., 1998). In our experiment, the activity of Cdk5 was stimulated following differentiation, but the protein expression of Cdk5 was also upregulated. This might be due to differences in the culturing conditions and/or induction of differentiation and the cell line used.

Fig. 2.

Cdk5 is necessary for the differentiation of primary OLs. (A) After induction of differentiation, primary OL morphology was observed using phase-contrast microscopy. Representative images of cells in each category are shown. (B) Percentages of the cells treated with or without 10 μM roscovitine in each category after induction of differentiation. (C) Primary OLs treated with or without 10 μM roscovitine immunostained with anti-MBP antibody (red) before or 3 days after induction of differentiation. Cells can only be identified by DAPI staining of the nuclei (blue). (D) Percentages of MBP-positive cells (red) from the total number of cells. Data were evaluated using Student's t-test (*P<0.01). (E) After induction of differentiation, primary OLs treated with or without 10 μM roscovitine were lysed and immunoblotted with antibodies against MBP, PLP1 and β-actin. (F) Primary OPCs were infected with Cdk5, p35 or luciferase shRNA retroviral vectors. At 3 days after induction of differentiation, lysates from the infected cells were immunoblotted with antibodies against Cdk5, p35 or β-actin. (G) Representative fluorescence images of MBP- and GFP-positive cells are shown. Arrows indicate GFP-expressing cells. (H) Percentages of MBP-positive cells (red) against the total number of infected cells. Data were evaluated using Student's t-test (*P<0.01).

Cdk5 phosphorylation of paxillin mediates differentiation

Paxillin has been originally characterized as the binding partner of focal adhesion proteins and actin cytoskeletal proteins. The primary function of paxillin is to integrate extracellular cues and cytoskeletal rearrangement. We previously reported that paxillin is required for morphological changes of neuroblastoma cells (Yamauchi et al., 2006). During these experiments, we investigated the effects that RNA interference (RNAi) of paxillin has on the morphological changes of various neuronal cells and have found that a paxillin siRNA potentially inhibits OL differentiation. To further define the role of paxillin in OL differentiation, we constructed retroviruses encoding both GFP and a shRNA sequence for paxillin. The paxillin shRNAs significantly reduced paxillin expression, whereas the control shRNA did not alter its expression in primary cells (Fig. 4A). We infected primary OPCs with these retroviruses and observed changes in morphology using an anti-MBP antibody. Whereas primary cells infected with control shRNA showed no effect on the formation of the web-like myelin structures, paxillin shRNAs significantly suppressed the cells ability to differentiate into mature OLs (Fig. 4B,C). Therefore, we have evidence that paxillin acts as a positive regulator in OL differentiation. Paxillin has been found to bind to the antiapoptosis protein Bcl2 and is thought to protect cells from apoptosis (Sorenson and Sheibani, 1999). Thus, to check whether the viability of OLs changes after the infection, cells were stained with 0.4% Trypan Blue and viable cells were counted. Trypan-Blue-incorporating cells numbered fewer than 1% in each experiment (0.7±0.010%, 0.4±0.008% or 0.9±0.014% for paxillin103, paxillin1206 or luciferase shRNA), suggesting that the knockdown of paxillin has no effect on the viability of OLs.

Paxillin has many potential phosphorylation sites that are involved in multiple cellular functions, such as cell migration, proliferation and differentiation (Brown and Turner, 2004). Among these phosphorylation sites, amino acid residues 244-247 (SPQR) of paxillin perfectly match the consensus Cdk5 phosphorylation sequence [S/T]PX[K/H/R] (Fig. 5A). To test the potential role of Cdk5 in OL differentiation, we used the non-phosphorylatable paxillin mutant paxillin-S244A tagged to GFP. This GFP-paxillin-S244A showed a dramatically reduced number of processes (Fig. 5B,C) and focal adhesion disassembly (Fig. 5B) in FBD-102b cells. By contrast, transfection of the plasmid encoding GFP-tagged wild-type paxillin had no significant effect on the number of processes (Fig. 5B,C) and clustered in focal adhesions at the tips of the processes (Fig. 5B). There was an approximate twofold reduction in the number of processes between the paxillin-S244A mutant and wild-type paxillin (Fig. 5B,C). Since Huang et al. have reported that paxillin Ser83 is phosophorylated by p38/MAPK and Cdk5 (Huang et al., 2004), we also transfected GFP-paxillin-S83A into FBD-102b cells. Expression of the paxillin-S83A mutant did not result in any significant changes (Fig. 5B,C), and expression levels of the two paxillin mutant proteins were comparable (Fig. 5D). Furthermore, we used a retroviral expression system to manipulate expression of wild-type or mutant paxillin and assessed the ability of primary OPCs to differentiate into mature OLs. OPCs infected with wild-type paxillin or the GFP-paxillin-S83A differentiated normally, as revealed by staining for MBP (Fig. 5E,F). When the paxillin-S244A mutant was expressed in OPCs, differentiation was inhibited (Fig. 5E,F). Formation of the web-like myelin structures was observed in non-infected OPCs (Fig. 5E,F). Collectively, these results indicate that phosphorylation of paxillin at Ser244 by Cdk5 plays an important role in OL differentiation.

Fig. 3.

Expression and the activity of Cdk5 are increased following differentiation of OLs. (A) After induction of differentiation, FBD-102b cells were lysed and immunoblotted with an anti-Cdk5 antibody (upper panel). RT-PCR analysis for Cdk5 mRNA was performed in FBD-102b cells after induction of differentiation (lower panel). The internal control β-actin is also shown. (B) After induction of differentiation, endogenous Cdk5 proteins were immunoprecipitated from the lysates of FBD-102b cells with anti-Cdk5 antibody, incubated with buffer containing histone H1 as the substrate, and immunoblotted with an antibody specific for a phosphorylated consensus-Cdk5 motif. The band intensity of phosphorylated histone H1 was quantified. (C) After induction of differentiation, primary OLs were lysed and immunoblotted with anti-Cdk5 antibody. (D) Activity of endogenous Cdk5 in primary OLs was measured after induction of differentiation using histone H1 as the substrate. The band intensity of phosphorylated histone H1 was quantified.

Fig. 4.

Paxillin is necessary for OL differentiation. (A) Primary OPCs were infected with paxillin103, paxillin1206 or luciferase shRNA retroviral vectors. At 3 days after induction of differentiation, lysates from infected cells were immunoblotted with antibodies against paxillin, Cdk5 and β-actin. (B) Representative fluorescence images of MBP and GFP-positive cells. Arrows indicate GFP-expressing cells. (C) Percentages of MBP-positive cells (red) against the total number of infected cells. Data were evaluated using Student's t-test (*P<0.01).

Fig. 5.

Cdk5 phosphorylation of paxillin mediates differentiation of OLs. (A) Amino acid sequences of human, mouse and rat paxillin. The consensus Cdk5 phosphorylation sequences are underlined. (B) FBD-102b cells were transfected with plasmid encoding GFP or the GFP-paxillin mutants, and the morphologies of the cells were observed using fluorescence microscopy. Representative fluorescence images of cells are shown. Arrows indicate cell adhesion. (C) Number of processes of cells transfected with the paxillin mutants. Data were evaluated using Student's t-test (*P<0.01). (D) FBD-102b cells were transfected with plasmid encoding GFP or the GFP-paxillin mutants. At 4 days after induction of differentiation, lysates from transfected cells were immunoblotted with antibodies against GFP and β-actin. The asterisk indicates a non-specific band. (E) Primary OPCs were infected with retrovirus expressing GFP, GFP-wt-paxillin, GFP-paxillin-S83A or GFP-paxillin-S244A. Representative fluorescence images of MBP- and GFP-positive cells are shown 3 days after induction of differentiation. Arrows indicate GFP-expressing cells. (F) Percentages of MBP-positive cells (red) against the total number of infected cells. Data were evaluated using Student's t-test (*P<0.01).

Cdk5 forms a complex with paxillin

Since Cdk5 is able to phosphorylate paxillin, we examined whether Cdk5 forms a complex with paxillin. We performed coimmunoprecipitation assays using 293T cells transfected with FLAG-Cdk5 and/or GFP-paxillin. Paxillin coimmunoprecipitated with Cdk5 from the lysates of cells co-transfected with Cdk5 and paxillin, suggesting that paxillin has the ability to form an immunocomplex with Cdk5 (Fig. 6A). Additionally, to identify the domain of paxillin that binds to Cdk5, we co-transfected paxillin fragments (see Materials and Methods) together with FLAG-Cdk5 into 293T cells and found that the LD1 and LD2 domains of paxillin specifically coimmunoprecipitated with Cdk5 (Fig. 6B), suggesting that the specific binding of paxillin to Cdk5.

Next, to investigate the effect of paxillin phosphorylation on a main binding protein, focal adhesion kinase (FAK), we examined the interaction of wild-type paxillin or mutant paxillin-S244A with FAK in 293T cells. Wild-type paxillin associated more efficiently with endogenous FAK than did the paxillin-S244A mutant (Fig. 6C). This finding suggests that phosphorylation at Ser244 affects the interaction of paxillin with its binding partners, probably leading to the formation of the web-like myelin structures.

Cdk5 phosphorylates Ser244 on paxillin

To investigate whether paxillin is a substrate for Cdk5, we performed in vitro phosphorylation assays of wild-type or mutant paxillin using activated recombinant Cdk5. Immobilized paxillin proteins were incubated with Cdk5 and immunoblotted with an antibody that specifically recognizes a consensus phosphorylated Cdk5 target sequence. Consistent with a previous report (Huang et al., 2004), we detected phosphorylation of wild-type paxillin and mutant paxillin-S83A with constant input levels of paxillin (Fig. 7A,B). By contrast, paxillin-S244A showed impaired phosphorylation by Cdk5, leading to the conclusion that Cdk5 phosphorylates paxillin at Ser244 in vitro.

Cdk5 binds and phosphorylates paxillin at Ser244 in primary differentiated OLs

Finally, we investigated the role Cdk5 plays in the phosphorylation of paxillin in differentiated OLs. Increased colocalization of Cdk5 with its substrate paxillin was detected throughout the cell bodies of primary OLs, and partially in the web-like myelin structures in differentiated OLs. (Fig. 8A). Furthermore, we detected an immunocomplex of Cdk5 with paxillin in primary OLs after induction of differentiation (Fig. 8B). This result is supported by the outcome of a reciprocal coimmunoprecipitation assay, suggesting that Cdk5 forms a complex with paxillin and phosphorylates paxillin at Ser244 in differentiated OLs. Therefore, to illustrate the significance of paxillin phosphorylation at Ser244 by Cdk5, endogenous paxillin was immunoprecipitated from primary OLs, and immunoblotted with an antibody recognizing a target sequence of phosphorylated Cdk5. As shown in Fig. 8C,D, endogenous paxillin was phosphorylated at Ser244 dependent on induction of differentiation. By contrast, pretreatment with roscovitine blocked phosphorylation of paxillin, suggesting that the phosphorylation of paxillin at Ser244 is mediated by Cdk5. Furthermore, the interaction of endogenous paxillin with FAK was increased after the induction of differentiation in primary OLs and this interaction was reduced with roscovitine, as suggested by the results of the transfection studies using 293T cells (Fig. 8E). Taken together, these results indicate that Cdk5 binds and phosphorylates paxillin at Ser244 to regulate OL differentiation.

Fig. 6.

Cdk5 forms a complex with paxillin. (A) In 293T cells, the immunoprecipitates of FLAG-Cdk5 were immunoblotted with an anti-GFP antibody to detect GFP-tagged paxillin. The total protein levels of Cdk5 and paxillin are also shown. (B) 293T cells were co-transfected with plasmids encoding LD1, LD2, LD3, LD4/5, LIM1, LIM2, LIM3 or LIM4 domains of paxillin and FLAG-Cdk5. Lysates of the cells were immunoprecipitated with anti-FLAG antibody and then immunoblotted with anti-HA antibody. Expression of paxillin domains and Cdk5 is also shown. Boxed bands indicate specific binding. (C) In 293T cells, the immunoprecipitates of GFP-wt-paxillin and GFP-paxillin-S244A were immunoblotted with anti-FAK antibody. The total protein levels of FAK and paxillin are also shown.

Fig. 7.

Cdk5 phosphorylates paxillin at Ser244 in vitro. (A) Immobilized paxillin constructs were incubated with recombinant Cdk5 and p25, and immunoblotted with antibody recognizing phosphorylated Cdk substrate. Total levels of GFP-paxillin are also shown. (B) Quantification of the intensity of the bands shown in the immunoblots.

Discussion

Cdk5 is a unique member of the Cdk family of serine/threonine kinases. Unlike other members of the Cdk family, Cdk5 is not directly involved in mediating the progression of the cell cycle. During development of the nervous system, Cdk5 activity is required for differentiation, migration, axon outgrowth and synaptogenesis. Cdk5 is expressed in all tissues; however, its kinase activity is likely to be restricted to the nervous system, because the activators of Cdk5, p35 and its isoform p39, are expressed solely in the nervous system (Lew et al., 1994; Tsai et al., 1994). Calpain-mediated proteolysis of p35 and p39 to p25 and p29, respectively, is known to be one of the mechanisms that regulate Cdk5 activity (Kamei et al., 2007; Kusakawa et al., 2000). In addition, the non-receptor tyrosine kinase Abl phosphorylates Cdk5 at Tyr15 and increases its kinase activity in neuronal cells (Zukerberg et al., 2000). Furthermore, the non-receptor Src-family tyrosine kinase Fyn also phosphorylates Cdk5 at Tyr15 and mediates semaphorin3A (Sema3A)-induced growth-cone collapse (Sasaki et al., 2002). Despite increasing evidence for a central role of Cdk5 in neuronal cells, few studies have reported Cdk5 function in glial cells (Tang et al., 1998). Here, we show a crucial role of Cdk5 in OL differentiation. The expression and kinase activities of Cdk5 are dramatically increased after induction of differentiation. We further find that paxillin is a target for Cdk5 in OL differentiation. These results support the idea that Cdk5 phosphorylation of paxillin plays a key role in the morphological changes of OLs during differentiation. Paxillin contains not only several protein-protein interaction regions, such as LD motifs and LIM domains, but also amino-acid motifs that are phosphorylated by tyrosine and serine/threonine kinases. Phosphorylated paxillin provides the binding sites for other proteins or modulates interaction of paxillin with other proteins. Phosphorylation of paxillin at Ser244 likely have an effect on its interaction with FAK. Further studies on the structural analysis will allow us to promote our understanding of how each phosphorylation can regulate paxillin.

As mentioned above, Fyn phosphorylates Cdk5 at Tyr15. Similar to Cdk5, Fyn is the only Src-family tyrosine kinase member that is upregulated during OL differentiation (Osterhout et al., 1999). Evidence indicates a central role for Fyn in OL differentiation and myelination (Osterhout et al., 1999; Umemori et al., 1999). Fyn–/– mice display thinner myelin sheaths than those observed in wild-type mice (Umemori et al., 1999). Interestingly, N-terminal domains of Fyn have been shown to be associated with myelin-associated glycoprotein (MAG) (Umemori et al., 1994). MAG is a transmembrane protein localized in oligodendroglial membranes. It functions not only as a ligand for an axonal receptor that is needed for the maintenance of myelinated axons, but also acts reciprocally as a receptor for an axonal signal that promotes survival and differentiation of OLs, and the maintenance of the myelin sheaths. Crosslinking of MAG proteins with anti-MAG antibodies increases the phosphorylation of Fyn in primary OLs (Marta et al., 2004) and in transfected COS-1 cells (Umemori et al., 1999). It is, thus, possible that MAG and Fyn regulate Cdk5 to modulate OL differentiation.

FBD-102b cells upregulate the expression levels of MBP and PLP1, and extend highly branched processes after induction of differentiation. However, FBD-102b cells are not able to form the web-like myelin structures. This cell line can be used as a model system for analyzing the process of early morphological differentiation but is not suitable for observing the entire differentiation process. In both FBD-102b cells and primary OPCs, Cdk5 inhibition by roscovitine appears to have a more inhibitory effect on the formation of multiple and long, branched processes than that on primary processes. In addition, studies using primary OPCs suggest that roscovitine significantly inhibits the formation of MBP-positive web-like myelin structures. Therefore, Cdk5 activity may participate in later differentiation stages rather than in the early process. This possibility is also supported by the result that paxillin-S244A mutant – which cannot be phosphorylated by Cdk5 – blocks formation of the web-like myelin structures. It is clear that Cdk5 phosphorylates paxillin, but substrates other than paxillin may also be responsible for OL differentiation. Given the important role for the dramatic cytoskeletal changes in OL differentiation, cytoskeletal proteins might act as potential substrates of Cdk5. Several recent studies identified the target proteins of Cdk5 that are involved in cytoskeletal changes of neuronal cells. In chick DRG neurons, the sequential phosphorylation of collapsing-response-mediating protein 2 (CRMP2) by Cdk5 and GSK3β is involved in Sema3A-induced growth-cone collapse (Uchida et al., 2005). In hippocampal neurons, Cdk5 phosphorylates the Wiskott-Aldrich syndrome protein (WASP) family verprolin homologous protein 1 (WAVE1) at multiple sites, resulting in an increase in the formation of dendritic spines (Kim, Y. et al., 2006). Similar to neuronal differentiation, OLs require dynamic changes in their cytoskeletons to generate the myelin sheath. Importantly, CRMP2 is detected in cultured OLs and participates in Sema3A-induced inhibition of process extension (Richard et al., 2000). WAVE1 is also expressed in OLs and specifically localizes to their leading edges (Kim, H. et al., 2006). WAVE1 is likely to be one of the main regulators of actin polymerization. Actually, fewer processes are observed in OLs isolated from mice that lack WAVE1 (Kim, H. et al., 2006). Conversely, the expression of WAVE1 in the CNS significantly increases following the progression of myelination. These observations suggest that WAVE1 mediates OL process outgrowth and is necessary for myelination (Kim, H. et al., 2006). It is conceivable that CRMP2 and/or WAVE1 are targets for Cdk5 during OL differentiation.

Fig. 8.

Cdk5 phosphorylates paxillin and forms a complex with paxillin after OL differentiation. (A) Primary OLs were immunostained with antibodies against Cdk5 (red) and paxillin (green) before or 3 days after induction of differentiation. Panels a-f are magnifications of the boxed areas as indicated. (B) Three days after induction of differentiation, immunoprecipitates of endogenous Cdk5 and paxillin in the lysates of OLs were immunoblotted with their respective antibodies. The total proteins levels of Cdk5 and paxillin are also shown. (C) Immunoprecipitates of endogenous paxillin from the lysates of OLs treated with or without 10 μM roscovitine immunoblotted with antibody recognizing phosphorylated Cdk substrate before or 3 days after induction of differentiation. Total protein levels of paxillin and β-actin are also shown. (D) Quantification of the intensity of the bands shown in the immunoblots. Data were evaluated using Student's t-test (*P<0.01). (E) Before or 3 days after induction of differentiation, immunoprecipitates of endogenous FAK in lysates of OLs treated with or without 10 μM roscovitine were immunoblotted with antibody against paxillin. Total protein levels of paxillin and FAK are also shown.

It has been suggested that Cdk5 participates in the pathogenesis of neurodegenerative disorders (Shelton and Johnson, 2004). Peptide p25, the proteolytic product of p35, accumulates in the brains of Alzheimer disease and amyotrophic lateral sclerosis patients (Patrick et al., 1999; Nguyen et al., 2001). At present, Cdk5 inhibitors are considered as one of several possible therapeutic agents for these diseases. However, side effects owing to the inhibition of Cdk5 must be also considered. At first, inhibition of Cdk5 results in the suppression of neurite outgrowth in neuronal cells (Shelton and Johnson, 2004). Here, we find a pivotal role of Cdk5 in OL differentiation and identify paxillin as the Cdk5 substrate involved in the morphologic differentiation of OLs. In contrast to these potential side effects, neurodegenerative disorders are occasionally accompanied by demyelination (Sherman and Brophy, 2005). Defects in axon guidance and neuronal migration are observed in Cdk5–/– mice (Gilmore et al., 1998). These defects might also be associated with an incomplete OL differentiation and/or defective myelination processes. Chemical compounds that have an effect on Cdk5 activity could be applied to treat demyelination diseases. Further studies will shed light on using Cdk5 inhibitors in the treatment of these diseases. Overall, further detailed studies investigating the molecular mechanisms in OLs that are regulated by Cdk5 will be valuable – not only in the understanding of how the development of the CNS is regulated but also in the development of new therapeutic strategies for the treatment of neurodegenerative disorders.

Materials and Methods

Materials

The following antibodies were purchased: anti-MBP (clone SKB3), anti-FAK (clone 4.47), and anti-paxillin (Upstate, Charlottesville, VA); anti-NG2, anti-MBP (119-131), anti-PLP1, and anti-paxillin (Chemicon, Temecula, CA); anti-β-actin (BD Biosciences Pharmingen, Franklin Lake, NJ); anti-Cdk5 (C-8) and anti-p35 (C-19) (Santa Cruz Biotechnology, Santa Cruz, CA); anti-phosphorylated (Ser) Cdk and anti-phosphorylated MAPK/Cdk (Cell Signaling Technology, Beverly, MA); anti-GFP (MBL, Nagoya, Japan); anti-FLAG (M2) (Sigma Biosciences, St Louis, MO); anti-RFP (Evrogen Moscow, Russia); and horseradish peroxidase (HRP)-conjugated anti-mouse or anti-rabbit secondary IgGs (GE Healthcare Bio Sciences, Piscataway, NJ). Alexa-Fluor-488 anti-mouse IgG and Alexa-Fluor-546 anti-rabbit IgG were obtained from Molecular Probes (Eugene, OR). Roscovitine was purchased from Biomol (San Diego, CA). Poly-L-lysine (PLL) was purchased from Sigma Biosciences.

Plasmids

The region encoding mouse paxillin protein was amplified by reverse transcriptase-polymerase chain reaction (RT-PCR) from mouse brain total RNA and ligated into the GFP-expressing mammalian vector pEGFP-C1 (Takara, Kyoto, Japan), as previously described (Yamauchi et al., 2006). GFP-tagged paxillin-S83A and paxillin-S244A constructs were created by Pfu-based PCR using the QuickChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). cDNA of paxillin was divided into eight fragments and each was inserted into pTurboRFP-C (Evrogen) containing a hemagglutinin (HA)-tag at the C-terminus. Fragments were LD1 (amino acids 1-140), LD2 (amino acids 141-207), LD3 (amino acids 208-260), LD4/5 (amino acids 261-320), LIM1 (amino acids 321-380), LIM2 (amino acids 381-440), LIM3 (amino acids 441-500), and LIM4 (amino acids 501-557). The pCMV-FLAG-Cdk5 plasmid was constructed as previously described (Kamei et al., 2007; Kusakawa et al., 2000). All sequences were confirmed by DNA sequencing.

RNAi

siRNA

The 21-nucleotide siRNA duplexes for RNA interference (RNAi) experiments were synthesized by Nippon EGT (Toyama, Japan). The target nucleotide sequence corresponding to nucleotides (5′-AAGCTGTACTCCACGTCCATC-3′) of mouse Cdk5 cDNA was designed according to the online software: RNAi Designer (http://bioinfo.clontech.com/rnaidesigner). The target sequence of a control (luciferase) siRNA was 5′-AAGCCATTCTATCCTCTAGAG-3′, which does not have significant homology to any mammalian gene sequences.

short-hairpin RNAs (shRNAs)

The oligonucleotides used were: Cdk5-sense (starting from nucleotide 531 of rat Cdk5), 5′-GATCCGCTGTACTCCACGTCCATCTTCAAGAGAGATGGACGTGGAGTACAGCTTTTTTACGCGTG-3′; Cdk5 antisense, 5′-AATTCACGCGTAAAAAAGCTGTACTCCACGTCCATCTCTCTTGAAGATGGACGTGGAGTACAGCG-3′; p35-sense (starting from nucleotide 811 of rat p35), 5′-GATCCGATGCTGCAGATCAATGCTTTCAAGAGAAGCATTGATCTGCAGCATCTTTTTTACGCGTG-3′; p35 antisense, 5′-AATTCACGCGTAAAAAAGATGCTGCAGATCAATGCTTCTCTTGAAAGCATTGATCTGCAGCATCG-3′; paxillin-sense 103 (starting from nucleotide 103 of rat paxillin), 5′-GATCCGCTGGAAACCACACTTACCATTCAAGAGATGGTAAGTGTGGTTTCCAGCTTTTTTACGCGTG-3′; paxillin-103 antisense, 5′-AATTCACGCGTAAAAAAGCTGGAAACCACACTTACCATCTCTTGAATGGTAAGTGTGGTTTCCAGCG-3′; paxillin-1206-sense (starting from nucleotide 1206 of mouse/rat paxillin), 5′-GATCCGGACTACCACAGCCTCTTCTTCAAGAGAGAAGAGGCTGTGGTAGTCCTTTTTTACGCGTG-3′; paxillin-1206 antisense, 5′-AATTCACGCGTAAAAAAGGACTACCACAGCCTCTTCTCTCTTGAAGAAGAGGCTGTGGTAGTCCG; luciferase-sense, 5′-GATCCGGCCATTCTATCCTCTAGAGTTCAAGAGACTCTAGAGGATAGAATGGCCTTTTTTAGATCTC-3′; and luciferase-antisense, 5′-AATTCAGATCTAAAAAAGGCCATTCTATCCTCTAGAGTCTCTTGAACTCTAGAGGATAGAATGGCCG-3′. The sense oligonucleotides were annealed with their respective antisense oligonucleotides, and annealed duplexes were ligated into the pSIREN-RetroQ-ZsGreen plasmid used to infect cells (Takara, Kyoto, Japan).

Cell line culture and transfection

Human epithelial 293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% heat-inactivated fetal bovine serum (FBS), 50 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO2. Mouse oligodendrocyte precursor FBD-102b cells were maintained in a 1:1 mixture of DMEM and F-12 medium containing heat-inactivated FBS at 10%, 50 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO2. To differentiate FBD-102b cells, cells were plated on PLL-coated dishes in a Sato medium (Itoh, 2002). Plasmid DNAs were transfected into 293T cells using the calcium phosphate transfection Kit (Takara) and FBD-102b cells using the Lipofectamine Plus (Invitrogen, Carlsbad, CA). Transfection efficiency typically exceeded 98% and 40%, respectively, using pEGFP-C1 as the control in 293T cells and FBD-102b cells. The siRNAs were transfected into FBD-102b cells using Lipofectamine 2000 (Invitrogen), according to the manufacturer's protocol. The phase-contrast and GFP-fluorescence images were captured with a Nikon (Kawasaki, Japan) Eclipse TE300 fluorescence microscope system.

Primary cell culture

Primary rat oligodendrocyte precursor cells (OPCs) were isolated from day-16 embryos of Sprague-Dawley rats according to the modified protocol from Itoh (Itoh, 2002). Briefly, cerebral cortices were dissected, dissociated with 0.25% trypsin (Invitrogen), triturated and passed through mesh (pore size 70 μm). Cells were collected at 184 g for 5 minutes, resuspended in modified Eagle's medium (MEM) containing 10% FBS, 50 U/ml penicillin, and 50 μg/ml streptomycin, and seeded on PLL-coated dishes at a density of 1×107 cells/dish. After 7 days in culture, the cells were passaged using 0.05% trypsin, resuspended in MEM containing 10% FBS, and cultured on PLL-coated dishes at a density of 8×106 cells per dish (first passage). After further 7 days in culture, the cells were passaged using 0.05% trypsin, resuspended in MEM containing 5% FBS, and cultured at a density of 3×106 cells/dish on non-coated Petri dishes (Sterilin, Staffordshire, UK) (2nd passage). On the second day of culture, the medium was changed to DMEM-based serum-free medium supplement with N-2 supplement (Invitrogen), 100 μg/ml BSA (Sigma), 60 μg/ml N-acetyl-L-cysteine (Sigma), 1 μM forskolin (Sigma), 5 nM neurotrophin-3 (Peprotech, London, UK), 10 nM basic fibroblast growth factor (Peprotech), and 10 nM platelet-derived growth factor (PDGF) (Peprotech). The cells were cultured for an additional 3 days. To differentiate OPCs, the cells were continuously cultured with a culture medium with 20 ng/ml T3 (triodothyronine) and 20 ng/ml T4 (thyroxine) without PDGF for 3-5 days.

Retrovirus-mediated transfection

The retroviral expression vectors and the envelope expression vector pVSV-G were cotransfected into GP2-293 cells by the Calcium Phosphate Transfection Kit (Takara) and the supernatants were harvested at 2 days after transfection. To concentrate the recombinant viruses, 30 ml of culture supernatants were centrifuged at 9000 g for 8 hours. The virus pellets were suspended in 3 ml of culture medium, and 0.3 ml of the virus solution was used to infect purified OPCs. Cells infected with the recombinant retroviruses were grown in culture medium overnight, and then cultured without PDGF for 3 days to induce differentiation.

Immunofluorescence

FBD-102b cells and primary cells were fixed in 4% paraformaldehyde in PBS for 8 minutes at room temperature. The fixed cells were permeabilized and blocked in PBS containing 20% normal goat serum and 0.1% Tween-20 for 1 hour at room temperature. Blocked cells were incubated with primary antibodies overnight at 4°C. Unbound primary antibodies were removed by washing three times with PBS containing 0.05% Tween-20. Washed cells were then incubated with secondary antibodies at room temperature for 1 hour. After three rinses with PBS containing 0.05% Tween-20, the coverslips were mounted onto slides using Vectashield containing DAPI (Vector Laboratories, Burlingame, CA) for observation by confocal microscope. The images were captured with an FV300 confocal microscope (Olympus, Tokyo, Japan) and processed by the Fluoview software (Olympus).

RNA preparation and RT-PCR analysis

Total RNA was prepared using Trizol reagent (Invitrogen). The cDNAs were prepared from 1 μg of total RNA with the SuperscriptIII enzyme (Invitrogen), according to manufacturer's instructions. PCR amplification was carried out for the indicated number of cycles, each cycle consisting of denaturation at 94°C for 1 minute, annealing at 59-62°C (depending on Tm values of primer pairs) for 1 minutes, and extension at 72°C for 1 minutes. The primers used were: Cdk5 (30 cycles), 5′-ATGCAGAAATACGAGAAACTGGAGAAG-3′ (sense) and 5′-GTACCACAGCGTGACCAC-3′ (antisense); MBP (40 cycles), 5′-ATGGCATCACAGAAGAGACCCTCAC-3′(sense) and 5′-TCAGCGTCTCGCCATGGG-3′ (antisense); PLP1 (30 cycles), 5′-ATGGGCTTGTTAGAGTGTTGTGCTAG-3′(sense) and 5′-CACCAGGAGCCATACAACAGTCAG-3′ (antisense); and β-actin (30 cycles), 5′-ATGGATGACGATATCGCTGCGCTC-3′ (sense) and 5′-CTAGAAGCATTTGCGGTGCACGATG-3′ (antisense).

Immunoprecipitation and immunoblotting

Cells were lysed in lysis buffer [50 mM HEPES-NaOH pH 7.5, 20 mM MgCl2, 150 mM NaCl, 1 mM dithiothreitol (DTT), 1 mM phenylmethane sulfonylfluoride (PMSF), 1 μg/ml leupeptin, 1 mM EDTA, 1 mM Na3VO4, 10 mM NaF, and 0.5% NP-40] and centrifuged as previously described (Miyamoto et al., 2006). Aliquots of the supernatants were mixed with a protein G resin preabsorbed with an anti-FLAG antibody. The immunoprecipitates or the proteins in the cell lysates were denatured and then separated on SDS-polyacrylamide gels. The electrophoretically separated proteins were transferred to PVDF membranes, blocked and immunoblotted. Three to five separate experiments were carried out, and a representative experiment is shown in Fig. 1D,G; Fig. 2E,F; Fig. 3A,C; Fig. 4A; Fig. 5D; Fig. 6A-C; Fig. 8B,C,E.

Cdk5 kinase assay

Immunoprecipitated Cdk5 was incubated with 100 ng/μl histone H1 as the substrate in 30 μl of reaction buffer (20 mM HEPES-NaOH pH 7.5, 1 mM DTT, 10 mM NaF, 0.1 mM PMSF, 0.1 μg/ml leupeptin, and 0.1 mM EDTA) containing 20 μM ATP at 30°C for 30 minutes, and then chilled on ice. The supernatants were collected and then immunoblotted with an antibody against phosphorylated MAPK/Cdk. Three to five separate experiments were performed, and a representative experiment is shown in Fig. 3B,D. The band intensity was also quantified.

In vitro paxillin phosphorylation

Immobilized paxillin protein was incubated with 100 ng/μl recombinant Cdk5/p25 protein (Upstate) in 30 μl of reaction buffer containing 20 μM ATP at 30°C for 15 minutes, and then chilled on ice. The immobilized phosphorylated paxillin was washed with reaction buffer and then immunoblotted with antibody recognizing phosphorylated Cdk. Three to five separate experiments were performed, and a representative experiment is shown in Fig. 7A. The band intensity was also quantified.

Statistical analysis

Values shown represent the mean ± s.d. from separate experiments. Student's t-test was carried out for intergroup comparisons (*P<0.01).

Acknowledgements

We thank E. M. Shooter and Y. Kaziro for helpful discussions. This work was supported by grants from the Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, Technology of Japan, the Ministry of Human Health and Welfare of Japan, research fellowships of the Japan Society for the Promotion of Science for Young Scientists, research grants from the Astellas Metabolic Disease Foundation, the Kampou Medical Foundation, the Kowa Life Science Foundation, the Nakajima Foundation, the Samuro Kakiuchi Foundation, the Takeda Science Foundation, and the Uehara Memorial Foundation.

  • Accepted October 6, 2007.

References

View Abstract