Mitotic reorganization of the intermediate filament protein nestin involves phosphorylation by cdc2 kinase.

The intermediate filament protein nestin is expressed during early stages of development in the central nervous system and in muscle tissues. Nestin expression is associated with morphologically dynamic cells, such as dividing and migrating cells. However, little is known about regulation of nestin during these cellular processes. We have characterized the phosphorylation-based regulation of nestin during different stages of the cell cycle in a neuronal progenitor cell line, ST15A. Confocal microscopy of nestin organization and (32)P in vivo labeling studies show that the mitotic reorganization of nestin is accompanied by elevated phosphorylation of nestin. The phosphorylation-induced alterations in nestin organization during mitosis in ST15A cells are associated with partial disassembly of nestin filaments. Comparative in vitro and in vivo phosphorylation studies identified cdc2 as the primary mitotic kinase and Thr(316) as a cdc2-specific phosphorylation site on nestin. We generated a phosphospecific nestin antibody recognizing the phosphorylated form of this site. By using this antibody we observed that nestin shows constitutive phosphorylation at Thr(316), which is increased during mitosis. This study shows that nestin is reorganized during mitosis and that cdc2-mediated phosphorylation is an important regulator of nestin organization and dynamics during mitosis.

The intermediate filament (IF) 1 protein nestin is expressed during early stages of development in progenitor cells of the central nervous system. Nestin is also found in developing muscle and myocardial cells (1)(2)(3)(4) as well as in developing sertoli cells (5). Upon differentiation, nestin expression is down-regulated and replaced by other tissue-specific IF pro-teins, such as glial fibrillary acidic protein (GFAP) in astrocytes, neurofilaments in neurons, and desmin in muscle tissue, respectively. Interestingly, nestin expression is reinduced during various regenerative and degenerative conditions in the fully differentiated organism (6,7).
Coexpression of nestin and the class III IF proteins, vimentin, GFAP, and desmin, has been documented in developing neuronal and embryonic cells. In these cells, nestin shows a similar intracellular organization as the class III IF proteins (1,8). Based on these observations, a possible copolymerization of nestin and the class III IF proteins, has been postulated. Recent data showed that nestin fails to assemble into filaments in the absence of a coexpressed IF network in astrocytes (9). In addition, it was recently shown that nestin cannot polymerize on its own but is integrated into type III IFs, most likely as a heterodimer (10).
Nestin has been shown to be associated with dividing and migrating cells and with cells rapidly changing their morphology. In light of the fact that nestin is expressed particularly in dynamic cells, it is of interest to study posttranslational modifications of nestin and the involvement of such modifications in regulating nestin organization. Intermediate filament proteins form dynamic structures that change their intracellular organization during various conditions, such as mitosis, differentiation, and different pathological situations (reviewed in Refs. [11][12][13][14][15]. The regulatory mechanisms behind these changes are still not fully understood, but phosphorylation has been implicated as an important regulator of IFs (reviewed in Refs. 13 and 16 -18). The role of phosphorylation in regulating IF organization is supported by studies showing that several kinases, including protein kinase C, cAMP-dependent kinase, Ca 2ϩ -calmodulin-dependent protein kinase, and cdc2 kinase, induce disassembly of IFs in vitro (18 -21). Changes in IF morphology can also be induced by elevating kinase activities in vivo, for example by expression of constitutively active forms of protein kinase C or Ca 2ϩ /calmodulin-dependent protein kinase II in astrocytes (22) or by microinjection of the catalytic subunit of protein kinase A into fibroblasts (23). Several IF proteins have been shown to undergo increased phosphorylation with subsequent reorganization of IF structure during mitosis. These proteins include vimentin (24,25), GFAP (26,27), keratins 8 and 18 (28,29), and the nuclear lamins (30,31). cdc2 kinase has been identified as a key regulator of IF organization during mitosis (reviewed in Refs. 13 and 18). Phosphorylation may also be involved in regulating the specific cellular distribution of IFs, as demonstrated by the preferential localization of phosphorylated neurofilaments in axons (32,33), and the localization of K18, phosphorylated on Ser 52 , to different cellular domains depending on the tissue (34). In addition to changes in IF protein phosphorylation during mitosis, there appears to be a continuous phosphate turnover on IFs in interphase, which could be involved in regulating the dynamic properties of IFs in interphase cells (13,16,35,36). Phosphorylation has also been implicated in the regulation of the interactions of IFs with both cytoskeleton-associated proteins, e.g. plectin, desmoplakin (37,38), or signaling proteins, e.g. 14-3-3 (39 -41).
The intracellular organization of nestin has not been characterized in detail, and little is known about the structural regulation of nestin during different cellular processes. In view of the previous observations on mitotic vimentin phosphorylation by cdc2, we have characterized the role of cdc2 kinase in regulating nestin organization during mitosis in an immortalized central nervous system precursor rat cell line, ST15A. In addition, we have identified Thr 316 as an in vitro cdc2-specific nestin phosphorylation site. We have generated a phosphospecific nestin antibody against the phosphorylated Thr 316 peptide to characterize phosphorylation of nestin at this site in vivo. The phosphorylation of this site is elevated during mitosis but also shows a certain degree of constitutive phosphorylation in interphase cells.

MATERIALS AND METHODS
Cell Culture and Synchronization-Immortalized rat central nervous system precursor cells, ST15A (42), were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin in a humidified incubator in a 5% CO 2 atmosphere at 37°C. For synchronization, cells were arrested in mitosis with 0.3 g/ml nocodazole (Sigma) for 12 h. The mitotic cells were harvested by mechanical shake off, whereas interphase cells were scraped off. Equal amounts of control and nocodazole-treated cells were harvested. The protein concentrations were determined by Bradford assay or by densiometric scanning of SDS-PAGE-separated protein bands, and the samples were normalized before immunoprecipitation or immunoblotting.
Antibodies, Double Immunofluorescence, and Confocal Microscopy-For double label immunofluorescence experiments, cells were grown on coverslips coated with 0.02 mg/ml poly-L-lysine (Sigma). Cells were washed in phosphate-buffered saline (PBS) and fixed in dehydrated methanol for 10 min at Ϫ20°C. Cells were rinsed in PBS, blocked with 1% bovine serum albumin, and incubated with nestin antiserum 130 diluted 1:500 or nestin antiserum 6 diluted 1:200 for 30 min at room temperature. The cells were then rinsed three times in PBS and incubated with fluorescein isothiocyanate-conjugated goat anti-rabbit immunoglobulins (Molecular Probes). Cells were washed with PBS and further incubated with a mouse monoclonal anti-vimentin antibody (RPN 1102, Amersham) diluted 1:30 for 30 min at room temperature, washed three times in PBS, and incubated for 30 min with Cy5-conjugated goat anti-mouse immunoglobulins (Zymed Laboratories Inc.) diluted 1:50, and washed three times in PBS before mounting in Mowiol 40-88 (32459-0, Aldrich-Chemie, Steinheim, Germany) supplemented with 100 mg/ml 1.4. diazabicyclo [2.2.2]-octan (2780-2, Aldrich-Chemie). Samples were analyzed in a Leitz Aristoplan fluorescence microscope and a Leica TCS40 confocal laser scanning microscope using the program SCANware 4.2a (Leitz, Heidelberg, Germany).
Nestin antiserum 6 was produced using a bacterially expressed fusion protein (plasmid kindly provided by Martha Marvin, Harvard Medical School, Boston, MA) containing the bacterial TrpE protein fused with the last 1197 amino acids of the carboxyl terminus of nestin, as antigen, according to protocol described elsewhere (43). Affinity purified rabbit anti-nestin 6 yielded results identical to those produced with rabbit nestin antibody 130 (1) in immunoblotting with purified proteins and various cellular extracts (Fig. 1A). Immunofluorescence with anti-nestin 6 and anti-nestin 130 gave identical microscopic patterns (Fig. 1B).
A phosphopeptide antibody against the synthezised nestin phosphopeptide KLEAENSRLQpTPG, anti Thr 316 nestin, was generated. The KLEAENSRLQpTPG peptide was conjugated to keyhole limpet hemocyanin via glutaraldehyde and used for immunization in rabbits. The anti-phospho-Thr 316 nestin serum was positively and further negatively affinity purified using prepacked HiTrap 1-ml columns (Amersham Pharmacia Biotech), to which the synthesized phosphopeptide and unphosphorylated peptide, respectively, were covalently coupled. For immunoblotting the double affinity purified anti-phospho-Thr 316 nestin serum was diluted 1:2000 in 5% bovine serum albumin/MOPS buffer (50 mM MOPS, 125 mM NaCl, and 12.5 mM NaOH, pH 7.4) and incubated overnight at 4°C. The blots were washed four times in MOPS buffer supplemented with 0.3% Tween, before incubation with an horseradish peroxidase-conjugated goat anti-rabbit secondary antibody. Blots were developed using the ECL Western blotting detection system (Amersham Pharmacia Biotech). To test the specificity of the antiserum, it was preincubated with excess phosphorylated peptide before incubating the blots. For immunofluorescence, the anti-Thr 316 nestin serum was diluted 1:50 in 1% bovine serum albumin/PBS and incubated overnight at 4°C, then washed three times with PBS, and further incubated with secondary antibody. Phosphorylation of vimentin on its cdc2-specific site, Ser 55 , was detected as previously described by using a phosphopeptide-specific antibody against vimentin phosphorylated on Ser 55 (44).
Determination of Nestin Phosphorylation Levels and Identification of Nestin Phosphopeptides-ST15A cells were metabolically labeled with 300 Ci/ml [ 32 P]orthophosphate for 4 h in phosphate-free Dulbecco's modified Eagle's medium supplemented with 10% dialyzed fetal calf serum in the presence of 0.3 g/ml nocodazole. This procedure blocked the cells entering mitosis. Mitotic cells were harvested by mechanical shake-off, centrifuged down at 1500 ϫ g 5 min at ϩ4°C, and washed twice by centrifugation in ice-cold PBS. Adherent cells were washed once with ice-cold PBS and harvested by scraping. Whole cell extracts were obtained by lysing the cells in SDS lysis buffer (20 mM Tris-HCl, pH 7.2, 5 mM EGTA, 5 mM EDTA, 0.4% SDS, 10 mM sodium pyrophosphate, 10 g/ml leupeptin, 10 g/ml antipain, and 10 g/ml pepstatin, respectively, and 1 mM PMSF). The cell extracts were boiled for 5 min and further sonicated for 20 s with a probe sonicator. Nestin and vimentin were immunoprecipitated by dilution of whole cell extracts with RIPA buffer (20 mM Hepes, pH 7.4, 140 mM NaCl, 10 mM pyrophosphate, 5 mM EDTA, 0.4% Nonidet P-40, 100 mM PMSF, 10 g/ml leupeptin, and 10 g/ml antipain) added to either anti-nestin antibody 130 or rabbit vimentin antibody 264, respectively, and finally recovered with protein A-Sepharose (Sigma).
The immunoprecipitated proteins were separated on SDS-polyacrylamide gels (45) and stained with either Coomassie Brilliant Blue or silver to control for equal loading of proteins. The gels were dried and autoradiographed at Ϫ70°C using Kodak X-Omat AR or Kodak Biomax MS films. Specific 32 P labeling of proteins was quantified using a phosphoimager (Bio-Rad).
For phosphopeptide or phosphoamino acid analysis, vimentin and nestin were separated on 7.5% SDS-PAGE, fixed in 50% methanol, dried, and autoradiographed. The corresponding nestin and vimentin bands were cut out from the gels digested twice (9 ϩ 3 h) with trypsin (T-8642, Sigma, 10 g/ml in 50 mM ammonium bicarbonate) at 37°C. The digested peptides were washed with double-distilled water and dried using a speed vac. Peptide maps of trypsin-digested samples (46) were obtained by two-dimensional separation on microcrystalline cellulose TLC plates (Merck). Trypsinized peptides were also subjected to phosphoamino acid analysis by acid hydrolysis and further two-dimensional TLC electrophoresis together with Ser(P), Tyr(P), and Thr(P) standards (46). Phosphoamino acids were visualized with ninhydrin staining and autoradiography.
Separation of IF Fractions Following Extraction with Triton X-100 -Triton X-100 buffer (20 mM Hepes, pH 7.6, 100 mM NaCl, 5 mM MgCl 2 , 5 mM EGTA, 1% Triton X-100, 1 mM PMSF, 10 g/ml leupeptin, and 10 g/ml antipain) was added to culture dishes, cells were detached and collected, homogenized on ice, and centrifuged for 15 min at 10,000 ϫ g (4°C) to pellet IF. The supernatant was further centrifuged at 200,000 ϫ g for 30 min at 4°C to obtain a soluble fraction (supernatant) and a pellet containing additional, most likely fragmented IF. All fractions were dissolved in 3ϫ Laemmli sample buffer (45) for further gel electrophoresis and Western blotting.
IF preparations containing 10 g of nestin and 100 g of vimentin were phosphorylated with ATP (20 M ATP and 10 Ci of [␥-32 P]ATP) using cdc2 isolated from BHK-21 fibroblasts (15) or recombinant cdc2 kinase (Calbiochem). The kinase reaction was carried out in 10 mM Hepes, pH 7.2, 60 mM NaCl, 0.5 mM CaCl 2 , 2.5 mM EGTA, and 2 mM MgCl 2 at 37°C and stopped after 30 min by addition of 3ϫ Laemmli sample buffer (45). The in vitro phosphorylated IF preparations were separated by SDS-PAGE and autoradiographed, cut out of the gel, digested with trypsin, and further separated by TLC phosphopeptide mapping as described above.
Purification of Trypsinized Nestin Phosphopeptides for Sequencing-Dried, trypsin-digested cdc2-phosphorylated nestin peptides were solubilized in 0.08% trifluoroacetic acid and fractionated by reverse phase chromatography on a C18 column (Aquapore OD300, 1 ϫ 250 mm; Applied Biosystems). The chromatography was preformed at room temperature with a flow rate of 0.1 ml/min. Peptides were eluted with a continuous gradient of 0 -80% acetonitrile containing 0.08% trifluoroacetic acid for a total of 120 min. Fractions containing 32 P-labeled peptides were collected and further fractionated on a C8 column (Aquapore RP300, 250 ϫ 1.0 mm; Applied Biosystems). Chromatography was carried out at room temperature with a flow rate of 0.1 ml/min. Fractions containing 32 P-labeled peptides were collected for microseqencing and further manual Edman degradation.
Amino Acid Sequence Analysis-The isolated 32 P-labeled peptides were subjected to sequence analysis and mass spectral fingerprinting. Amino-terminal amino acid sequence analysis was performed with an Applied Biosystems model 477A protein sequencer equipped with an on-line applied Biosystems model 120A phenylthiohydantoin amino acid analyzer. Phosphopeptide fractions were analyzed by mass spectrometry on PerSeptive Biosystems Elite STR mass spectrometer using ␣-cyano-cinnamic acid (10 mg/ml in 50% acetonitrile, 0.1% trifluoroacetic acid) as the matrix. Spectra were acquired in the reflector mode, and the mass spectra were internally mass calibrated. The obtained sequences were confirmed with the published amino acid sequence of nestin predicted from cDNA (1).
Manual Edman Degradation-Manual Edman degradation was carried out on the sequenced radioactively labeled nestin peptides as described (47). The peptides were subjected to manual Edman degradation for 8 -20 cycles. The material released after each cycle and the disc containing any residual undigested peptides were counted.

RESULTS
The mitotic process involves reorganization of cytoskeletal structures including the cytoplasmic and nuclear IFs. Because nestin is expressed predominantly in cells that undergo mitosis or rapidly change morphology, we examined the regulation of nestin morphology and organization in comparison with the well characterized class III IF protein vimentin in a rat immor-talized central nervous system precursor cell line, ST15A, during different stages of the cell cycle.
Morphological Changes of Nestin Organization during Interphase-Mitosis Transition-During interphase, ST15A cells contained IF networks that extended from the perinuclear region to the cell surface. Nestin and vimentin formed indistinguishable networks in these cells, as demonstrated by double label confocal microscopy (Fig. 2, A, B, K, L, and M). During the transition from late prophase to metaphase, the IF network was reorganized into cage-like structures surrounding the nucleus (Fig. 2, D and E). Nestin and vimentin networks underwent identical changes during mitosis. The colocalization of nestin and vimentin persisted throughout the cell cycle. During telophase, nestin and vimentin showed a more diffuse and much less intense fluorescence in the cleavage furrow, indicating a possible disassembly of the IF networks at this site (Fig. 2, G and H).
IF networks show variable behavior in different cell types during mitosis. In some cell types a cage of IF is maintained around the nucleus, whereas in other cell types IFs break down or disassemble with a simultaneous increase in the solubility of the IF proteins. It was, therefore, of interest to determine the solubility state of the IF proteins in mitotic ST15A cells. Detergent extractions of control and nocodazole-treated ST15A cells were centrifuged at different speeds to obtain three different fractions: an insoluble fraction after 10,000 ϫ g and soluble and insoluble fractions after 200,000 ϫ g. Immunoblot- ting of the fractions demonstrated that most of the protein stays as filaments (insoluble fraction after 10,000 ϫ g) in nocodazole-treated cells. However, there was an increase in the soluble fraction after centrifugation at 200,000 ϫ g (Fig. 3), indicating the presence of disassembled soluble subunits. Note the presence of the lower molecular mass protein bands, representing cleaved forms of nestin. The intensity of these bands increases in the soluble fractions of nocodazole-treated cells, indicating a possible processing of nestin during mitosis.
Mitotic Phosphorylation of Nestin Is Mediated by cdc2 Kinase-The mitotic reorganization of the cytoplasmic IF networks in many cell types is known to be regulated by phosphorylation of the IF proteins. Therefore, the phosphorylation state of nestin in ST15A cells during interphase-mitosis transition was examined. The mitotic reorganization of the cytoplasmic IF networks in ST15A cells was accompanied by an increase in the phosphorylation levels of the IF proteins, nestin and vimentin (Fig. 4A). In accordance with previous studies (24,25), vimentin showed a low level of constitutive phosphorylation, which was elevated 6-fold in mitotic cells (Fig. 4A, left  panel). In comparison with vimentin, nestin had a higher level of constitutive phosphorylation in interphase cells and showed a 3-fold increase in the phosphorylation level during mitosis (Fig. 4A, right panel). In addition to the ϳ220-kDa nestin band, immunoprecipitation with the nestin antibody also yielded lower molecular mass bands. These bands were weakly phosphorylated in control cells and showed increased phosphorylation in nocodazole-treated cells. They most likely represent cleaved forms of nestin, because immunoblotting of IF preparations from ST15A cells (Fig. 1) and whole cell extracts of ST15A cells (data not shown) using antibodies against nestin shows the presence of corresponding bands. Phosphoamino acid analysis of nestin immunoprecipitated from in vivo 32 Plabeled mitotic ST15A cells showed phosphorylation of nestin at serine and threonine residues (Fig. 4C).
cdc2 kinase is involved in the mitotic phosphorylation and consequent reorganization of various IF networks in different cell lines (13,25,47). When IF preparations from ST15A cells were treated with cdc2 kinase, nestin, vimentin, and the nuclear lamins were phosphorylated (Fig. 4B) (also see 25,30). The in vitro phosphorylation of nestin occurred on serine and threonine residues (Fig. 4C).
Comparative phosphopeptide mapping of nestin immunopre-cipitated from metabolically labeled interphase and mitotic ST15A cells (Fig. 5) showed five mitosis-specific phosphopeptides (phosphopeptides 1, 2, 3, 4, and 6). There were also interphase-specific phosphopeptides (phosphopeptides 7, 8, and 9), and one phosphopeptide that was present in both interphase and mitotic preparations (5). Although not obvious from Fig. 5, which does not give a quantititative presentation of the relative labeling intensity on interphase versus mitotic peptides, peptide 5 showed consistently elevated phosphorylation in mitotic preparations as compared with interphase samples. According to quantification by phosphoimager image analysis, the relative increase (mitosis/interphase) was 1.9fold Ϯ 0.3 (mean Ϯ range of values; measurements made on peptide maps obtained from five different experiments). The mitosis-specific phosphopeptides 1, 2, 3, 4, and 6 corresponded to cdc2-specific tryptic phosphopeptides (Fig. 5). The identity of the individual peptides was confirmed by parallel chromatography of mixed samples (data not shown). The relative dominance of the mitotic peptides that correlated to peptides phosphorylated by cdc2 in vitro, indicated that cdc2-mediated phosphorylation sites are likely to regulate nestin organization during mitosis. These results identify nestin as a physiological substrate for cdc2 kinase and suggest that this kinase is involved in the mitotic reorganization of nestin.
Identification of cdc2-specific Phosphorylation Sites on Nestin-To determine cdc2-specific phosphoamino acid residues on nestin, ST15A IF preparations phosphorylated by cdc2 in vitro were run on SDS gels, from which the phosphorylated nestin was cut out and further digested with trypsin. The resulting 32 P-labeled tryptic peptides were fractionated by reverse phase HPLC. Radioactive fractions (radioactive peaks 1-7) were sequenced and subjected to manual Edman degradation to determine the cycle at which the radiolabel is released. The nestinpeptide sequence 316 LQTPGR was obtained by successful sequencing of radioactive peak 1. Manual Edman degradation of the fraction confirmed phosphorylation of the peptide on the third amino acid, corresponding to the threonine residue (Thr 316 ) of the sequence obtained (Fig. 6). Sequencing of the other 32 P peaks observed in the graph were not successful, because of the impurity of the fractions. Nestin is a very large protein (Ͼ1800 amino acids), and cleaving nestin with trypsin generated too many peptides to allow for sufficient purification of all phosphopeptides by the techniques we used.
Because in vitro phosphorylated sites may not reflect the real phosphorylation sites in vivo, a polyclonal phosphospecific nestin antibody, anti-phospho-Thr 316 nestin, was generated to characterize phosphorylation of nestin at this site in vivo. The antibody showed high specificity against the phosphopeptide used for immunization, as compared with unphosphorylated peptide (Fig. 7A). The affinity purified antibody recognized the full-length nestin protein (ϳ220 kDa) (Fig. 7B, first panel). In addition, the antibody showed strong immunoreactivity against a ϳ175-kDa band and weak reactivity against several bands of lower molecular mass (Fig. 7B, first panel). This ladder pattern can be seen on Western blots using several different nestin antibodies at high concentration, Most likely the pattern of fragmented protein represents processing products of nestin. This kind of fragmentation is seen in nestin in vivo preparations from many different cell types. 2 Western blot of whole cell extracts of nocodazole-treated and control cells using a polyclonal nestin antibody raised against the carboxyl terminus of nestin (rabbit anti-nestin 6) showed a similar pattern (Fig. 7B, third panel). The polyclonal antibody recognized the ϳ175-kDa band in addition to prominent immunoreactivity against the ϳ220-kDa band (Fig. 7B, third panel). Immunoblotting of whole cell extract of ST15A cells with a monoclonal nestin antibody also gives immunoreactivity at ϳ175 kDa (Fig.  7B, fourth panel).
The results by using the phosphopeptide antibody indicated that nestin was constitutively phosphorylated at Thr 316 and that phosphorylation at this site increased during mitosis, as shown by the increased intensity of the ϳ220and the ϳ175-kDa bands in nocodazole-treated cell samples (Fig. 7B, first  panel). The phosphorylation of the ϳ175-kDa band was especially pronounced. Experiments on competitive binding inhibition, where the antibody was preincubated with phosphopeptide in excess, almost completely abolished the antibody reactivity of the ϳ220-kDa band and the band at ϳ175-kDa (Fig. 7B, second panel from the left). Immunoblotting of in vitro cdc2-phosphorylated and unphosphorylated IF preparations showed that cdc2 kinase increases the phosphorylation of nestin (Fig. 7C).
Immunoblotting with a monoclonal nestin antibody of insoluble and soluble fractions of control and nocodazole-treated cells after centrifugation at 10,000 ϫ g and 200,000 ϫ g also revealed the presence of the ϳ175-kDa band (Fig. 3). This band seems to transfer to the soluble pool after centrifugation at 200,000 ϫ g during mitosis, because the intensity of the band increases in the soluble fraction of nocodazole-treated cells. Another possible explanation for the increased intensity is that nestin is cleaved during mitosis to yield an increased amount of a soluble ϳ175-kDa peptide.
The constitutive phosphorylation of nestin at Thr 316 in interphase cells and the increase in phosphorylation at this site during mitosis was further supported by immunolabeling of interphase and mitotic cells using the phosphospecific antibody (Fig. 7D). Confocal images showed weak immunoreactivity on nestin IFs in interphase cells with an elevated reactivity in mitotic cells. In interphase cells there is a punctate staining along the filaments. Toward the cell periphery the staining becomes more diffuse. The immunoreactivity is strongest around the nucleus. In mitotic cells the staining is more intense, but no clear filaments are seen, compared with the monoclonal nestin staining, where filaments can be visualized. In comparison, the one identified cdc2-target on vimentin, i.e. Ser 55 , showed complete mitotic specificity with no phosphorylation in interphase cells, as revealed with a Ser(P) 55 -specific antibody (Fig. 7D). DISCUSSION During interphase, nestin and vimentin form indistinguishable cytoplasmic networks in ST15A cells. This indicates that these two IF proteins form copolymers. Copolymerization between nestin and class III IF proteins has been discussed in earlier studies (1,8) and is supported by the findings that neither nestin nor vimentin can form filaments in GFAP-deficient astrocytes without the presence of a coexpressed IF network (9). The ability of nestin to coassemble with vimentin in vitro has recently been shown (10) our results, there is a dramatic reorganization of the IF networks into cage-like structures in normal and nocodazole-arrested mitotic ST15A cells. Coincident with the morphological changes seen in the IF networks during mitosis in ST15A cells, there is an increase in the phosphorylation levels of nestin and vimentin. Nestin shows constitutive phosphorylation in interphase cells with a 3-fold increase in mitotic cells. In comparison, the elevated phosphorylation of mitotic vimentin is more dramatic, about 6-fold. The constitutive phosphorylation of nestin may correspond to specific phosphorylation sites involved in the tissue-specific functions of nestin. Other possible functions for phosphorylation of nestin are regulation of the spatial organization of IF proteins, as has been shown for neurofilaments, or regulation of protein-protein interactions with other IF proteins, such as vimentin. Assuming that nestin coassembles with type III IF proteins in vivo, it is possible that the long carboxyl-terminal domain of nestin functions as a spacer or cross-linking element between IFs and other cytoskeletal components, thereby regulating their supramolecular organization as proposed in Refs. 10 and 49. Phosphorylation could modulate the configuration of the tail domain, affecting the formation of protein-protein interactions and thereby the organization of the cytoskeletal network. In this study, the data suggest a role for cdc2 kinase in the phosphorylation of nestin and its reorganization during mitosis. Furthermore, we identified Thr 316 as a cdc2-specific in vitro phosphorylation site of nestin. Interestingly, Thr 316 is located in the carboxyl terminus near the end of the rod domain. The sequences at the carboxyl ends of the rod domain are highly conserved among the IF proteins. This part of the rod domain is important for proper IF assembly, and many of the mutations that have been identified in IF-related diseases, such as severe cases of epidermolysis bullosa simplex, are located in these regions (4). Phosphorylation of nestin at Thr 316 could, therefore, have an important impact on the assembly proper-ties of nestin and the organization of the IF network in the cell. A peptide nestin antibody directed against a nestin peptide phosphorylated at Thr 316 was generated to characterize phosphorylation of nestin at this site in vivo. We were able to show that nestin is constitutively phosphorylated at Thr 316 and that phosphorylation at this site is elevated during mitosis. In addition to the ϳ220-kDa nestin band, the phospho-specific nestin peptide antibody showed very strong immunoreactivity to a ϳ175-kDa band. The intensity of the ϳ175-kDa band was markedly increased in nocodazole-treated samples. Interestingly, a corresponding ϳ175-kDa band could be seen in autoradiographs of nestin immunoprecipitated from in vivo labeled mitotic ST15A cells. Phosphorylation of this band was increased in nocodazole-treated cells (Fig. 2). There is a difference between nestin and vimentin when it comes to prolinedirected phosphorylation. The only SP site on vimentin (Ser 55 ) shows no constitutive phosphorylation, whereas the TP site on nestin (Thr 316 ) shows significant phosphorylation in interphase cells.
The kinase responsible for the constitutive phosphorylation at Thr 316 is unknown, but it could be another cyclin-dependent kinase. For example, cdk5, which is required for normal neuronal differentiation and muscular development (50 -52), might be a candidate for interphase-specific phosphorylation of nestin at Thr 316 . cdk5 together with p35, the protein activator of cdk5, plays a role in the regulation of the structure of the neuronal cytoskeleton structure because it has been shown to phosphorylate cytoskeletal components such as the microtubule-associated protein tau and the neuronal IF protein, neurofilament H (53,54). Considering the requirement for cdk5 activity for proper neuronal and muscular development and the effect of cdk5 phosphorylation on cytoskeletal organization, we tested whether this kinase can phosphorylate nestin in vitro, and our preliminary data shows that nestin is a bona fide FIG. 6. Identification of nestin Thr 316 as an in vitro cdc2-specific phosphorylation site. Nestin phosphorylated in vitro by cdc2 was trypsinized, and phosphopeptides were fractionated by reverse phase HPLC. 32 P-containing fractions were collected separately and further purified by a second run. Radioactive fractions were collected, and the amino acid sequence of the tryptic peptide was determined. Peak 1 yielded the peptide sequence LQTPGR. This sequence correspond to the nestin sequence starting at amino acid 314 at the end of the rod domain. The fraction was further subjected to manual Edman degradation to confirm phosphorylation of the threonine residue. substrate for cdk5. 3 cdk5-mediated phosphorylation could explain the phosphorylation observed in interphase cells, but further studies are required to elucidate this.
Our data indicate that cdc2 kinase is involved in the regulation of nestin organization during mitosis. It remains to be determined whether phosphorylation by cdc2 is sufficient to account for the reorganization of nestin during mitosis or whether some other kinase is required. There are other kinases in addition to cdc2 kinase that have been shown to phosphorylate IF in vivo during mitosis, including p37 kinase, which phosphorylates vimentin at Ser 458 and Thr 457 in mitotic BHK-21 cells (55); protein kinase C, which shows mitosisspecific phosphorylation of lamin B and vimentin (56,57); and Rho kinase, which phosphorylates GFAP during late mitosis (57,58). Recently, Rho kinase was shown to phosphorylate vimentin at Ser 71 at the cleavage furrow in late mitotic cells (60).
In summary, our interest in IF dynamics has led us to study 3  A phosphopeptide-specific antibody against phosphorylated Thr 316 was generated. A, immunoblotting of 0.5 g of the Thr 316 phosphorylated and unphosphorylated peptide spotted on nitrocellulose membrane revealed specificity of the generated antibody. The antibody recognizes the phophorylated peptide but shows no immunoreactivity against the unphosphorylated peptide. B, based on immunoblotting of control and nocodazole-treated cells, nestin is constitutively phosphorylated at this site during interphase, but the phosphorylation level is increased during mitosis. The phospho-specific peptide antibody shows strong immunoreactivity against a ϳ175-kDa nestin peptide, in addition to reactivity against the full-length ϳ220-kDa nestin (P-thr-316 nestin). The same samples were immunoblotted with a polyclonal nestin antibody (anti-nestin 6) to confirm the identity of the ϳ175-kDa band (polyclonal anti-nestin). Immunoblotting of whole cell extracts of ST15A cells with a monoclonal nestin antibody also showed the presence of the ϳ175-kDa peptide (monoclonal anti-nestin). Preincubating the antibody with the phosphorylated peptide abolished immunoreactivity against the ϳ220-kDa band and the lower molecular mass band of ϳ 175 kDa (P-thr-316 nestin ϩ p-peptide). The additional low molecular bands seen on the blot were still visible when the antibody was preincubated with excess of the phosphorylated peptide. Equal numbers of control and nocodazole-treated cells were harvested, and equal loading of proteins was further controlled by Coomassie Brilliant Blue staining (data not shown). C, immunoblotting of in vitro cdc2-phosphorylated and unphosphorylated IF preparations shows that cdc2 kinase increases phosphorylation at Thr 316 . D, double immunofluorescence labeling of ST15A cells with the Thr 316 -nestin phosphopeptide-specific antibody and monoclonal nestin antibody. Immunofluorescence labeling of interphase ST15A cells with the nestin phosphopeptide specific antibody (panel a) shows a fragmented pattern along the filaments. The intensity of the staining is increased in mitotic cells. As comparison, panel b shows labeling with the monoclonal nestin antibody, where the labeling in interphase versus mitotic cells is of similar intensity. Panel c shows labeling of ST15A cells with a phosphopeptide antibody which recognizes vimentin when phosphorylated on the major cdc2-specific site, Ser 55 . Panel d shows the whole vimentin IF network as revealed by a monoclonal vimentin antibody. Immunoreactivity of the vimentin phosphopeptide-specific antibody is absent in interphase cells.
the previously uncharacterized regulation of the IF protein nestin during the cell cycle. We have shown that reorganization of nestin during mitosis is coupled to increased phosphorylation of nestin and that cdc2 kinase seems to play a major role in this phosphorylation-mediated structural modification of nestin.