Identification of Peripherin as a Akt Substrate in Neurons*

Activation of Akt-mediated signaling pathways is crucial for survival and regeneration of injured neurons. In this study, we attempted to identify novel Akt substrates by using an antibody that recognized a consensus motif phosphorylated by Akt. PC12 cells that overexpressed constitutively active Akt were used. Using two-dimensional PAGE, we identified protein spots that exhibited increased immunostaining of the antibody. Mass spectrometry revealed several major spots as the neuronal intermediate filament protein, peripherin. Using several peripherin fragments, the phosphorylation site was determined as Ser66 in its head domain in vitro. Furthermore, a co-immunoprecipitation experiment revealed that Akt interacted with the head domain of peripherin in HEK 293T cells. An antibody against phosphorylated peripherin was raised, and induction of phosphorylated peripherin was observed not only in Akt-activated cultured cells but also in nerve-injured hypoglossal motor neurons. These results suggest that peripherin is a novel substrate for Akt in vivo and that its phosphorylation may play a role in motor nerve regeneration.

Akt (also known as protein kinase B) is a Ser/Thr kinase that plays essential roles in various cellular processes such as cell survival, proliferation, and differentiation (1). In the nervous system, Akt is suggested to be involved in neurogenesis (2,3), neuronal survival (4), axon or dendrite formation (5,6), synaptogenesis (7,8), and synaptic transmission (9). The most evident role of all may be its neuroprotective action. For instance, several previous papers have demonstrated a strong protective effect of Akt on damaged neurons in vivo (10 -13). Of particular interest, Akt was proven to have a crucial role in neuronal survival after peripheral nerve injury (10). In the peripheral nervous system, in which most neurons can survive and regenerate after injury, glial cells secrete various trophic factors to promote survival and regeneration of nerve-injured neurons. Astrocytes and microglia, which are located around the neuronal cell bodies, are thought to secrete various factors toward injured neurons (14,15). Furthermore, in the distal stump of axons far from neuronal cell bodies, Schwann cells also secrete trophic factors (16). Such factors released from those glial cells include a wide range of growth factors such as nerve growth factor, brainderived neurotrophic factor, glial cell line-derived neurotrophic factor, and fibroblast growth factor-2 (17)(18)(19). They are known to activate the phosphatidylinositol 3-kinase-Akt pathway in injured neurons via their respective receptors (20 -22). In fact, our previous study showed that Akt activity was markedly induced in motor neurons after nerve injury (10). We also revealed that activated Akt accelerated axonal elongation, as well as neuronal survival. It is well established that activated Akt exerts its function by phosphorylating its substrates; however, the substrates that specifically exist in neurons are largely unidentified. Thus, identification of novel neuronal substrates is pivotal to gain further insight into the function of Akt in neuronal regeneration.
In this study, we attempted to identify novel Akt substrates in neurons by a proteomic approach, using a unique antibody that recognizes the consensus motif phosphorylated by Akt. Here we demonstrate that peripherin, which is a peripheral nervous system neuron-specific intermediate filament protein, is a novel Akt substrate, and that Ser 66 of peripherin is the phosphorylation site. Peripherin phosphorylation is apparently induced in motor neurons after nerve injury, suggesting that the Akt-mediated peripherin phosphorylation may play a role in motor nerve regeneration.
ies. As secondary antibodies, horseradish peroxidase-conjugated antibodies (Amersham Biosciences) and Alexa Fluorconjugated antibodies (Molecular Probes, Eugene, OR) were used for Western blotting and immunohistochemistry, respectively. All of the inhibitors were obtained from Calbiochem (La Jolla, CA).
Two-dimensional PAGE-PC12 cells grown on 10-cm cell culture dishes were infected with AxCALNLLacZ (multiplicity of infection (MOI) 100) or AxCALNLHA-CA-Akt (MOI 100) together with AxCANCre (MOI 30). The cells were collected 48 h after infection, washed once with PBS, and lysed in a buffer containing 40 mM Tris base, 8 M urea, and 2% CHAPS. After centrifugation at 10,000 ϫ g for 20 min at 4°C, the supernatants were aliquoted and stored at Ϫ80°C until use. Two-dimensional PAGE was performed according to the previous report with slight modification (28). Immobiline DryStrips (pH 3-10, 7 cm; pH 4.5-5.5, 24 cm; Amersham Biosciences) were rehydrated with rehydration solution containing 60 g (for 7 cm gel) or 240 g (for 24 cm gel) of the supernatants, 8 M urea, 2% CHAPS, 0.5% IPG buffer (Amersham Biosciences), 20 mM dithiothreitol, and bromphenol blue for 12 h at 20°C. Isoelectric focusing was then performed using the IPGphor Isoelectric Focusing System (Amersham Biosciences) (500 V for 1 h, 1000 V for 1 h, and 8000 V for 2-3 h at 20°C). The strips were equilibrated with a buffer containing Tris-HCl, pH 6.8, 6 M urea, 30% glycerol, 2% SDS, and 65 mM dithiothreitol for 20 min at room temperature, fixed vertically on top of the SDS-polyacrylamide gel by 1.5% agarose in running buffer, and subjected to 25 mA/gel in a cold room. The gels were analyzed by Western blotting for immunostaining or SYPRO Ruby for protein staining according to the manufacturer's protocol (Molecular Probes, Eugene, OR).
In-gel Digestion-Protein spots were punched out from the gel, trimmed into small pieces, destained in a solution containing 50% acetonitrile and 25 mM NH 4 HCO 3 , and dehydrated. The gel pieces were then rehydrated in a solution containing 10 mM dithiothreitol and 25 mM NH 4 HCO 3 and subsequently treated with 25 mM NH 4 HCO 3 containing 55 mM iodoacetoamide. Following the dehydration step, gel pieces were rehydrated in trypsin solution containing 10 mg/ml trypsin and 25 mM NH 4 HCO 3 overnight at 37°C, and finally digested peptides were eluted with 50% acetonitrile containing 5% trifluoroacetic acid.
Mass Spectrometry-The eluate containing digested peptides was desalted with ZipTip (Millipore, Bedford, MA). The solution was then mixed with an equal volume of saturated ␣-cyano-4-hydroxycinnamic acid solution dissolved in 30% acetonitrile and 0.1% trifluoroacetic acid and spotted onto a target plate. Mass spectrometry was performed on a matrixassisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometer Reflex III (Bruker Daltonics, Billerica, MA) with reflector mode. Obtained peptide mass fingerprinting data were searched against the NCBI data base using the MASCOT search engine (Matrix Science, Boston, MA).
Western Blotting-Protein extracts were separated by SDS-PAGE, and blots were prepared on polyvinylidene difluoride membranes (Millipore). For two-dimensional gels, whole 7-cm gels or part of 24-cm gels were prepared on the membrane. The blots were probed with primary and subsequent secondary antibodies and visualized by using the chemiluminescence system (Western Lightning; PerkinElmer Life Sciences). If necessary, the membranes were stripped of antibodies by incubating in stripping buffer containing 62.5 mM Tris-HCl, pH 6.8, 2% SDS, and 100 mM 2-mercaptoethanol for 30 min at 50°C and then probed with another antibody.
Preparation of Recombinant Proteins-To generate GST fusion proteins, partial sequences for 1-60, 51-100, 181-251, and 301-350 amino acids of peripherin were amplified from full-length mouse peripherin cDNA (kindly provided by Dr. F. Landon) and subcloned into pGEX 5X-1 (Amersham Biosciences). Site-directed mutagenesis (Ser 66 or Ser 79 to Ala) was introduced by PCR primers carrying these mutations. BL21 bacteria transformed with these vectors were stimulated with 0.2 mM isopropyl-␤-D-thiogalactopyranoside overnight at 20°C, harvested by brief centrifugation, and lysed in PBS containing 1% Triton X-100 for 30 min at 4°C. The supernatants were subsequently incubated with glutathione-Sepharose 4B (Amersham Biosciences) for 1 h at 4°C, and the bound proteins were eluted by adding 10 mM reduced glutathione in 50 mM Tris-HCl, pH 8.0. After removal of glutathione by dialysis against PBS, the proteins were checked by SDS-PAGE followed by Coomassie Brilliant Blue R-250 staining and stored at Ϫ80°C until use.
In Vitro Kinase Assay-2.5 g of GST fusion proteins were incubated with or without 100 ng of recombinant His-tagged CA-Akt (His-CA-Akt) (Upstate Biotechnology) in 20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 20 M ATP, and 30 kBq [␥-32 P]ATP (PerkinElmer Life Sciences) for 30 min at 30°C. The reaction mixtures were subjected to SDS-PAGE, and phosphorylation of the fragments was detected by autoradiography. For Western blot analysis, 1 g of GST fusion proteins was reacted with 100 ng of His-CA-Akt, and one-tenth of the reaction mixtures was analyzed.
Phosphorylation-specific Antibody-A rat monoclonal antibody that specifically recognized phosphorylated peripherin (anti-pPer antibody) was raised in accordance with the previous report (29,30). Briefly, a 10-week-old female WKY/NCrj rat was immunized with a synthetic peptide containing phosphorylated Ser 66 (ARLGpS 66 FRAPRC). Three weeks after immunization, lymph nodes obtained from the rat were dispersed, and lymphocytes were fused with mouse myeloma Sp2/0-Ag14 cells. The phosphorylation-specific antibody was screened by enzyme-linked immunosorbent assay using hybridoma supernatants, and clone 2C2 was selected. Finally, 2C2 hybridoma cells were injected into the abdominal cavity of nude mice, and prepared ascites were used for immunological assays.
Detection of Peripherin Phosphorylation in Cultured Cells-HEK 293T cells seeded on 60-mm culture dishes were grown to ϳ80% confluence and transfected with pcDNA3-peripherin and pcDNA3-HA-WT-Akt using Lipofectamin 2000 (Invitrogen). After 8 h, the cells were seeded into 12 well culture dishes and cultured for another 24 h. The cells were then serumstarved for 10 h, treated with insulin (Sigma), and subjected to Western blot analysis using the anti-pPer antibody. If necessary, inhibitors were added to the cultured medium 30 min before insulin treatment. As for the phosphorylation of endogenous peripherin, PC12 cells infected with AxCALNLLacZ (MOI 100), AxCALNLHA-WT-Akt (MOI 100), AxCALNLHA-CA-Akt (MOI 100), or AxCALNLHA-DN-Akt (MOI 100) together with AxCANCre (MOI 30) for 48 h were examined.
Immunoprecipitation-HEK 293T cells seeded on 6-well culture dishes were transfected with pcDNA3-HA-WT-Akt together with pcDNA3 empty vector or FLAG-tagged head domain of peripherin (FLAG-Per 1-103) subcloned into pcDNA3 using Lipofectamin 2000. After 32 h, the cells were serum-starved for 10 h and treated with 100 nM insulin for 20 min. The cells were then washed in Tris-buffered saline briefly and lysed in radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.25% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml aprotinin, 1 mM Na 3 VO 4 , and 10 mM NaF). After centrifugation at 10,000 ϫ g for 10 min at 4°C, the soluble fractions were collected and reacted with anti-FLAG antibody followed by precipitation using protein G-Sepharose 4B (Sigma). Immunoprecipitates were rinsed four times with lysis buffer and eluted by adding 2ϫ SDS sample buffer.
Immunohistochemistry-Adult male Wistar rats weighing ϳ150 g were anesthetized with pentobarbital (40 mg/kg) and positioned supine, and their right hypoglossal nerves were crushed with forceps. The rats were perfused with 4% paraformaldehyde in 0.1 M phosphate buffer 5 days after surgery. The brains were quickly removed, post-fixed overnight at 4°C in the fixative, and immersed in 0.1 M phosphate buffer containing 25% sucrose for an additional day. Sections were cut on a cryostat (18 m in thickness), washed once in PBS, and treated with 10 g/ml proteinase K for 10 min. After two washes in PBS, the sections were blocked with PBS containing 10% normal goat serum for 1 h and subsequently reacted with primary antibodies (anti-peripherin antibody; 1:1000, anti-pPer antibody; 1:1000) in PBS containing 1% normal goat serum overnight at 4°C. After three washes in PBS, the sections were incubated with secondary antibodies for 1 h and finally washed three times in PBS. The sections were visualized by fluorescent microscopy (AX70; Olympus, Tokyo, Japan).

Identification of Peripherin as an Akt Substrate in Neurons-
To identify novel neuronal substrates for Akt, we utilized the anti-phospho-Akt substrate antibody. Akt preferentially phosphorylates Ser or Thr in the RXRXX(S/T) motif, and the antibody specifically recognizes this motif only when Ser or Thr is phosphorylated. PC12 cells infected with adenovirus expressing LacZ or CA-Akt were subjected to Western blot analysis using this antibody, and proteins exhibiting more intense signal in the CA-Akt-expressing preparation were searched. Our preliminary experiment using ordinary SDS-PAGE demonstrated stacked positive bands where isolation of the individual positive band was impossible (data not shown). We therefore performed two-dimensional PAGE to also separate proteins by their isoelectric points, and the two-dimensional gels were analyzed by Western blotting using the antibody. We initially used a wide pH range gel for the first dimension and found numerous spots were intensely stained in the CA-Akt-expressing preparation; in particular in the region in which the isoelectric point was 5.0 -5.5 and molecular mass was ϳ60 kDa (Fig. 1A). We therefore focused on this region and separated proteins more precisely by using narrow pH range gels for the first dimension (Fig.  1B). Six spots that exhibited the intense positive immunostaining were identical to the protein spots in the protein-stained gels (spots 1-4, 8, and 9 in Fig. 1C). Judging from their sequential spot patterns, we assumed that spots 1-4 were the same proteins, each of which might have different post-translational modifications. Similarly, the spots 8 and 9 were assumed to be the same protein. As representative samples, spots 1 and 9 were punched out from the gel and analyzed by MALDI-TOF mass spectrometry to identify the corresponding proteins. The subsequent data base search revealed that both spots were identical to peripherin. All spots (spots 1-4, 8, and 9) were confirmed as peripherin by Western blot analysis using the anti-peripherin antibody (Fig. 1D).
Akt Phosphorylates Ser 66 of Peripherin in Vitro-Peripherin, whose expression is mostly restricted to neurons in the peripheral nervous system, is a member of type III intermediate filament proteins (31). Because peripherin has not been identified as an Akt substrate, we performed further analysis. First, we aimed to determine the phosphorylation site by Akt in vitro using recombinant proteins. Although no typical consensus sequence for the Akt substrate, RXRXX(S/T), was found in peripherin, five potent sequences existed ( Fig. 2A). Because several previous papers indicated that Akt could possibly recognize some similar sequences as its target (details are described under "Discussion"), we examined the possibility that Akt was able to recognize and phosphorylate some similar sequences. Four types of GST fusion proteins that contained one or two potent sequences were generated and reacted with recombinant CA-Akt protein in the presence of [␥-32 P]ATP (Fig. 2B). Autoradiography showed that one fragment containing 51-100 amino acids of peripherin (GST-Per 51-100) was exclusively phosphorylated by CA-Akt among four fragments. Because GST-Per 51-100 contained two potent sequences, SARLGS 66 and ALRLPS 79 , we then introduced site-directed mutagenesis to GST-Per 51-100 to produce unphosphorylated mutants. These proteins were analyzed by an in vitro kinase assay (Fig.   2C). The S66A mutation entirely prevented Akt phosphorylation, whereas the S79A mutation did not cause any alterations. These results demonstrate that Akt phosphorylates Ser 66 of peripherin in vitro. The sequence containing Ser 66 is highly conserved among mammalian species (Fig. 2D).
Ser 66 of Peripherin Is Phosphorylated in Akt-activated Cultured Cells-To evaluate peripherin phosphorylation in vivo, a monoclonal antibody (anti-pPer antibody) was raised against the synthetic peptide ARLGpS 66 FRAPRC. Specificity of this antibody was tested by Western blot analysis using GST-Per 51-100 in vitro (Fig. 3A). The anti-pPer antibody could detect GST-Per 51-100 only when the fragment was reacted with CA-Akt, and the intense immunoreactivity entirely disappeared when the S66A mutant was used. Using this antibody, peripherin phosphorylation was examined in HEK 293T cells. HEK 293T cells were transfected with WT-Akt and peripherin, because they have no endogenous peripherin, subsequently stimulated with insulin to activate Akt, and peripherin phosphorylation was detected by Western blot analysis. First, HEK 293T cells were treated with increasing doses of insulin, and peripherin phosphorylation was examined (Fig. 3B). Both Akt activation, which was evaluated by the phosphorylation state of Akt (32), and peripherin phosphorylation occurred in a dosedependent manner. Next, we observed changes in peripherin phosphorylation over time after insulin treatment (Fig. 3C). Peripherin was phosphorylated in a time-dependent manner, which paralleled Akt activation. To further demonstrate that Akt kinase activity regulated peripherin phosphorylation, we used several inhibitors to modulate Akt activity. Both Akt activation and peripherin phosphorylation were almost prevented by pretreating cells with LY294002 (phosphatidylinositol 3-kinase inhibitor, which inhibited upstream signaling of Akt). In contrast, peripherin phosphorylation was not prevented by Me 2 SO (the vehicle for control), U0126 (MEK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase) inhibitor, which prevented mitogen-activated protein kinase signaling), or rapamycin (mTOR inhibitor, which prevented one of the downstream signaling of Akt). We also examined the phosphorylation of endogenous peripherin in PC12 cells by Western blot analysis (Fig. 3E). Peripherin phosphorylation was hardly detected in PC12 cells infected with adenovirus expressing LacZ, WT-Akt, or DN-Akt. In contrast, peripherin phosphorylation was clearly observed in cells expressing CA-Akt. Although some minor additional bands were observed at different molecular masses in this blotting using PC12 cells, we assumed those bands would be nonspecific because their intensity was not affected by WT-, CA-, or DN-Akt expression. Together with the results obtained by HEK 293T cells, these results demonstrate that Ser 66 of peripherin is phosphorylated in an Akt-mediated pathway in cultured cells.
Akt Interacts with the Head Domain of Peripherin in Vivo-It is likely that Akt may directly phosphorylate Ser 66 of peripherin in vivo. To provide further support for this possibility, we examined whether these two proteins could interact in vivo using a co-immunoprecipitation experiment. Full-length peripherin, most of which may form intermediate filament in cells, is almost detergent-insoluble (33), and we assumed that peripherin might not be solubilized in a typical lysis buffer for immu- . PC12 cells infected with adenovirus expressing LacZ or HA-CA-Akt for 48 h were subjected to two-dimensional PAGE using wide pH range gels (pH 3-10). The gels were then analyzed by Western blotting using antiphospho-Akt substrate antibody. B and C, protein staining (B) and immunoblotting (C) of the gels (pH 5.10 -5.45; molecular mass (MW), 44 -68 kDa) of the CA-Akt-expressing preparation. Extracted proteins from PC12 cells expressing HA-CA-Akt were separated by two-dimensional PAGE using a narrow pH range gel. The gel was stained with SYPRO Ruby, a protein detection reagent (B) or transferred to a nitrocellulose membrane followed by immunoblotting using anti-phospho-Akt substrate antibody (C). Six spots (spots 1-4, 8, and 9) in B were recognized by anti-phospho-Akt substrate antibody in C. D, peripherin (Per) spots in the HA-CA-Akt-expressing preparation. The same membrane used in C was reprobed with anti-peripherin antibody. All of the peripherin spots are indicated by arrows (spots 1-10).
noprecipitation. Therefore, we used a deletion form of peripherin for the immunoprecipitation experiment. Because our preliminary experiment showed that the head domain of peripherin (1-103 amino acids), which contained Ser 66 , could be solubilized entirely in radioimmunoprecipitation assay buffer (data not shown), we used the head domain instead of full-length peripherin in this assay. HEK 293T cells transfected with FLAG-Per 1-103 and WT-Akt were treated with or without insulin and subjected to immunoprecipitation using the anti-FLAG antibody (Fig. 4). Akt was co-precipitated with FLAG-Per 1-103, indicating that they could interact in vivo. It was of note that this interaction was not dependent on Akt activity because insulin treatment did not enhance the interaction. A similar activity-independent binding has also been reported on several other Akt substrates (34,35).

Ser 66 of Peripherin Is Phosphorylated in Regenerating Hypoglossal Motor Neurons-Previous reports have revealed that
Akt was activated in response to neuronal injury (10,13,36). In particular, Akt activation is crucial for nerve-injured motor neurons to regenerate (10). We therefore examined the phosphorylation of endogenous peripherin in nerve-injured hypoglossal motor neurons. We crushed the hypoglossal nerve, and then peripherin expression and phosphorylation were examined by immunohistochemistry 5 days after injury (Fig. 5). Peripherin expression was induced in the cell bodies and also axons of injured neurons (Fig. 5A). The higher magnification photographs showed a significant increase of peripherin immunoreactivity in the cell bodies of injured neurons (Fig. 5, D and  G) and a simultaneous induction of peripherin phosphorylation in the cell bodies (Fig. 5, E and H). However, in injured nerves, immunoreactivity for phosphorylated peripherin was hardly observed, although the peripherin protein was induced and abundantly expressed (Fig. 5, A-C and J-L). This observation may suggest that enhanced immunoreactivity for phosphorylated peripherin in the cell body of injured neurons is not simply caused by increased peripherin expression. These results indicate that the Ser 66 of peripherin is phosphorylated in regenerating neurons and that phosphorylated peripherin was predominantly localized to their cell bodies but not to axons.

DISCUSSION
We performed a proteomic approach to identify novel Akt substrates in neurons using the anti-phospho-Akt substrate antibody. The present study revealed that Akt phosphorylates Ser 66 of peripherin both in vitro and in vivo. The antibody we used recognizes the RXRXX(pS/pT) motif, which is preferentially recognized and phosphorylated by Akt. However, the sequence containing Ser 66 (SARLGS 66 ) is not typical for Akt substrates where only one Arg residue exists at the Ϫ3 position, although the typical one has Arg residues at both Ϫ3 and Ϫ5 positions. Because a similar variation has been reported for several Akt substrates, the Arg at the Ϫ5 position may not always be necessary. For instance, the sequences of PSRTAS in ATPcitrate lyase, LSRRPS in cAMP response element-binding protein, GARRSS in 14-3-3, PMRNTS in p21-activated protein kinase 1, and HVRAHS in Yes-associated protein can be phosphorylated by Akt both in vitro and in vivo (37)(38)(39)(40)(41). As for the ϩ1 position, peripherin has a Phe residue that would be suitable as an Akt substrate because previous studies have shown that a large hydrophobic residue in the ϩ1 position is preferable (42). . B and C, in vitro kinase assay using recombinant peripherin fragments. GST fusion proteins containing potent sequence(s) were generated. 2 g of each fragment was subjected to SDS-PAGE followed by Coomassie Brilliant Blue (CBB) staining for checking (left panels). These fragments were reacted with recombinant His-CA-Akt protein in the presence of Mg 2ϩ and [␥-32 P]ATP for 30 min, and phosphorylation was detected by autoradiography (right panels). Phosphorylation of GST-peripherin 51-100 by CA-Akt was clearly observed (B). GST-Per 51-100 carrying the S66A mutation entirely prevented Akt phosphorylation, whereas the S79A mutation did not show any alterations (C). D, conservation of the motif among different mammalian species. Conserved amino acids between two species and those among three species are shown as gray and black boxes, respectively. Phosphorylated Ser residue in the motif is indicated by an asterisk.
These studies support the finding that Ser 66 in peripherin is likely to be phosphorylated by Akt.
The RXRXX(S/T) motif is fairly specific as an Akt substrate; however, this motif can also be the phosphorylation target by another evolutionally related kinase, p70 S6 kinase (p70S6K) (43). Because p70S6K is activated downstream of the phosphatidylinositol 3-kinase-Akt pathway (44), the result obtained from our screening raised an alternative possibility in which peripherin might be phosphorylated by p70S6K rather than Akt. This possibility might be disregarded by the results of the in vitro kinase assay, where only recombinant proteins were used in these experiments (Figs. 2, B and C, and 3A). However, that result does not entirely rule out a possibility that p70S6K may phosphorylate peripherin in vivo, because an in vitro kinase is not always an in vivo kinase (45,46). To eliminate this possibility, we examined the effect of p70S6K inhibition on peripherin phosphorylation in vivo using cultured cells (Fig.  3D). We treated HEK 293T cells with rapamycin, which is known to inhibit p70S6K activation via inhibition of its upstream mTOR kinase (44). The phosphorylation state of peripherin was not changed by rapamycin treatment, whereas it was markedly decreased by phosphatidylinositol 3-kinase inhibitor LY294002 treatment when the cells were stimulated by insulin. Furthermore, we performed a co-immunoprecipitation experiment using the head domain of peripherin and revealed that peripherin was capable of interacting with Akt in vivo (Fig. 4). Taken together, these data suggest that Akt directly phosphorylates peripherin in vivo.
Although various types of Akt substrates have been identified in a variety of cell types to date, to our knowledge this is the first report that identified an intermediate filament protein as an Akt substrate. Peripherin is a member of type III intermediate filament proteins, which include vimentin, desmin, and glial  fibrillary acidic protein (31). Intermediate filament proteins generally consist of a central coiled-coil ␣-helical rod domain flanked by a head and a tail domain (47), and Ser 66 resides in the center of the head domain of peripherin. Generally, head and tail domains of intermediate filament proteins contain several phosphorylation sites, whereas rod domains do not (48). It is well known that a similar neuron-specific intermediate filament, the neurofilament (NF), which belongs to type IV intermediate filament proteins, has multiple phosphorylation sites within its head or tail domains (49). With regard to peripherin, it had been hypothesized that its head domain might contain multiple phosphorylation sites like other intermediate filament proteins (50), and in fact, it has been reported that the N terminus of peripherin, which contains a head domain and a half of rod domain, was phosphorylated in cultured neurons, although the exact phosphorylation sites have not been determined yet (33,51). Therefore, Ser 66 could be one of phosphorylation sites previously suggested in those papers.
Of the intermediate filament proteins, peripherin is probably one of the most unknown members in terms of physiological function. However, there are several studies demonstrating its characteristic expression pattern. In contrast to NF proteins, which are widely expressed in various kinds of neurons, periph-erin expression is rather restricted to peripheral nervous system neurons (50,(52)(53)(54). It is of interest that peripherin expression is transiently induced in injured neurons (55)(56)(57)(58). Helfand et al. (60) recently revealed that peripherin might be required for both formation and elongation of neurites in PC12 cells, although there remains controversy (59). Therefore, it is likely that peripherin may be involved in neuronal regeneration, probably by contributing to rearrangement of intermediate filaments in neurons during nerve regeneration processes. Although the functional significance of the Akt-mediated peripherin phosphorylation remains unclear, Akt, which is necessary for nerve regeneration, would transfer divergent signals to the substrates, including peripherin, for proper nerve regeneration.
The present immunohistochemical study demonstrated that peripherin phosphorylation occurred endogenously in hypoglossal neurons after nerve injury (Fig. 5). These data suggest that not only peripherin expression but also peripherin phosphorylation may be associated with neuronal regeneration. Although it remains unclear how peripherin phosphorylation plays a role in motor nerve regeneration, immunohistochemical localization of phosphorylated peripherin may provide clues to address this issue. In adult rats, peripherin expression was induced in both cell bodies and axons of injured neurons (Fig. 5, A, D, G, and J). However, phosphorylated peripherin was only observed in cell bodies of injured neurons (Fig. 5, E and H) and could hardly be detected in injured axons (Fig. 5K). This cell body-specific localization is reminiscent of phosphorylated NF-L. It has been shown that Ser 55 within the head domain of NF-L is phosphorylated by protein kinase A (61,62). This occurred immediately after its synthesis in neuronal cell bodies, and thereafter the phosphorylation of NF-L disappeared, along with its translocation into axons (61). In addition, transgenic mice in which Ser 55 was replaced with Asp to mimic permanent phosphorylation showed aberrant NF-L inclusions in neuronal perikarya (63). Because NF-L phosphorylation by protein kinase A can lead to disassembly of the filament (64), it is assumed that phosphorylation may block premature assembly of NF-L in cell bodies before transport into axons (49,61,63). As mentioned above, Ser 66 of peripherin is located at the head domain, and some phosphorylation of head domains often causes disassembly of intermediate filaments (48). If Ser 66 phosphorylation is able to trigger disassembly of peripherin fil- ament, Akt may control the dynamics of intermediate filament in regenerating axons by phosphorylating newly synthesized peripherin and prevent them from assembly in neuronal cell bodies until the appropriate timing.
It is also possible that peripherin phosphorylation may modulate interactions with other proteins. Previous reports suggest that peripherin has interactive proteins. Peripherin was shown to bind to dystonin (also known as BPAG1-n), which was assumed to be a cross-linker between intermediate filaments and microfilaments (65). Peripherin was also reported to interact with the small heat shock protein ␣B-crystallin (66). Although it has not been addressed which part of peripherin may be responsible for these interactions, these interactive properties may be regulated by Akt-mediated phosphorylation.
In conclusion, we have identified peripherin as a novel neuronal substrate for Akt both in vitro and in vivo, and Akt-mediated phosphorylation was induced in regenerating motor neurons. Because Akt is known to play a pivotal role in neuronal regeneration, peripherin would be one of the significant substrates for Akt during nerve regeneration processes. To gain a better understanding of Akt function in regenerating neurons, in particular the functional significance of Akt for the cytoskeletal rearrangement in neurons after nerve injury, further studies, such as how Ser 66 phosphorylation is capable of changing physiological property of peripherin, are required.