Identification of FEZ1 as a protein that interacts with JC virus agnoprotein and microtubules: role of agnoprotein-induced dissociation of FEZ1 from microtubules in viral propagation.

The human polyomavirus JC virus (JCV) is the causative agent of a fatal demyelinating disease, progressive multifocal leukoencephalopathy, and encodes six major proteins, including agnoprotein. Agnoprotein colocalizes with microtubules in JCV-infected cells, but its function is not fully understood. We have now identified fasciculation and elongation protein zeta 1 (FEZ1) as a protein that interacted with JCV agnoprotein in a yeast two-hybrid screen of a human brain cDNA library. An in vitro binding assay showed that agnoprotein interacted directly with FEZ1 and microtubules. A microtubule cosedimentation assay revealed that FEZ1 also associates with microtubules and that agnoprotein induces the dissociation of FEZ1 from microtubules. Agnoprotein inhibited the promotion by FEZ1 of neurite outgrowth in PC12 cells. Conversely, overexpression of FEZ1 suppressed JCV protein expression and intracellular trafficking in JCV-infected cells. These results suggest that FEZ1 promotes neurite extension through its interaction with microtubules, and that agnoprotein facilitates JCV propagation by inducing the dissociation of FEZ1 from microtubules.

The human polyomavirus JC virus (JCV) 1 is the causative agent of progressive multifocal leukoencephalopathy, a fatal demyelinating disease. The genome of JCV comprises a double-stranded circular DNA molecule that contains three functional regions: the viral early and late genes and the noncoding regulatory sequence (1). The early region of the JCV genome encodes the large T and small t antigens, which are responsible for the initiation of viral DNA replication and activation of late gene transcription (2)(3)(4). The late coding region encodes three structural proteins (VP1, VP2, and VP3) that are components of the viral capsid (5). In addition, the leader sequence of late transcripts encodes agnoprotein, a viral auxiliary protein that contains 71 amino acids (6).
The auxiliary proteins of various eukaryotic viruses have diverse effects on different stages of infection, including transcription (7,8), viral assembly (9), and the release of viral particles (10,11). They can also affect host cell functions and thereby contribute to the pathogenesis of viral-induced disease (12). The agnoprotein of simian vacuolating virus 40 (SV40), which belongs to the same polyomavirus family as does JCV, contributes to various stages of the viral lytic cycle. Mutation of the agnoprotein of SV40 was found to result in a moderate growth defect that was attributable to impairment of the viral maturation pathway (13)(14)(15). Immunofluorescence analysis revealed that SV40 agnoprotein facilitates the localization of VP1, the major capsid protein, to the nucleus and perinuclear region of infected cells (16). Furthermore, the lack of agnoprotein led to inefficient release of mature SV40 virions from infected cells and impaired the ability of the virus to propagate in monkeys (14).
In addition to SV40, JCV is closely related to other polyomaviruses, including human BK virus. These viruses exhibit marked nucleotide sequence similarity, especially in the coding regions (1). All JCV proteins with the exception of agnoprotein are localized predominantly in the nuclei of infected cells (17,18), whereas JCV agnoprotein is largely restricted to the perinuclear region of the cytoplasm (19), as is SV40 agnoprotein (20). We previously showed that deletion of the agnogene of JCV results in a viral growth defect (21). Furthermore, a small interfering RNA (siRNA) specific for agnoprotein mRNA was found to inhibit JCV infection (22). These observations suggest that the agnoprotein of JCV, like that of SV40, plays an important role in viral propagation, although the molecular mechanism of this action remains unknown.
We have now identified fasciculation and elongation protein zeta 1 (FEZ1) as a protein that interacted with JCV agnoprotein in a yeast-two hybrid assay. FEZ1 is a brain-specific coiled-coil protein that comprises 392-amino acid residues and is expressed in neurons (23). It is related to Caenorhabditis elegans UNC-76, which is required for normal axonal bundling and outgrowth. We further demonstrate that agnoprotein inhibited the function of FEZ1 apparently by blocking the association of FEZ1 with microtubules, and overexpression of FEZ1 inhibited the intracellular spread of JCV. Our results suggest that agnoprotein promotes the intracellular translocation of viral particles on microtubules by inducing the dissociation of FEZ1.

EXPERIMENTAL PROCEDURES
Construction of Plasmids-A full-length FEZ1 cDNA was amplified by the PCR from an adult human brain cDNA library and was subcloned into pCXN2-FLAG (24), pGEX6P1 (Amersham Biosciences), and pEGFP-N1 (Clontech, Palo Alto, CA); the resulting expression vectors were designated pFLAG-FEZ1 (for expression of FEZ1 with an NH 2terminal FLAG tag), pGST-FEZ1 (for expression of FEZ1 fused at its NH 2 terminus with glutathione S-transferase), and pFEZ1-GFP (for expression of FEZ1 fused at its COOH terminus to green fluorescent protein), respectively. Complementary DNAs for deletion mutants of FEZ1 were generated by PCR from pFLAG-FEZ1 and subcloned into pGEX6P1. For expression of JCV agnoprotein in mammalian cells, the agnoprotein cDNA was amplified by PCR from a plasmid containing the JCV genome, pJC1-Ͼ4pJCV (HSRRB, Osaka, Japan), and subcloned into either pcDNA4HisMax (Invitrogen) or the bicistronic expression vector pERedNLS (kindly provided by M. Matsuda).
Yeast Two-hybrid Assay-The yeast two-hybrid assay was performed with a Matchmaker System 3 and an adult human brain cDNA library obtained from Clontech. Yeast AH109 cells were transformed with both the brain cDNA library and a full-length cDNA for JCV agnoprotein subcloned into the yeast shuttle vector pGBKT7. Plasmids isolated from positive colonies were introduced into Escherichia coli DH5␣ and sequenced, and the DNA sequence data were compared with sequences in the NCBI data base with the BLAST program.
Cell Culture and Virus Preparation-Human embryonic kidney 293 (HEK293) cells (HSRRB), SV40-transformed human glial SVG-A cells (kindly provided by W. J. Atwood) (25), and JCV-producing (JCI) cells (26) were maintained under an atmosphere of 5% CO 2 at 37°C in Dulbecco's minimal essential medium supplemented with 10% heatinactivated fetal bovine serum, 2 mM L-glutamine, penicillin, and streptomycin (Sigma). HEK293 cells that express JCV agnoprotein in an inducible manner (293AG cells) were established with the T-REx system (Invitrogen); agnoprotein expression was induced by exposure of the cells to doxycycline (Invitrogen) at a final concentration of 1 g/ml.
To establish 293AG cells that stably express FEZ1-GFP, we transfected 293AG cells with pFEZ1-GFP and cultured the transfectants in the presence of Geneticin (Invitrogen) at 400 g/ml. Several clones were expanded in the selection medium, after which expression of FEZ1-GFP was confirmed by immunoblot analysis. Control 293AG cell lines were similarly established by stable transfection with pEGFP-N1. To establish SVG-A cells that stably express FLAG-FEZ1, we transfected SVG-A cells with pFLAG-FEZ1 and cultured the transfectants in the presence of Geneticin (100 g/ml). Expression of FLAG-FEZ1 in selected cells was confirmed by immunoblot analysis. Again, control SVG-A cell lines were similarly established by stable transfection with pCXN2-FLAG. PC12 cells stably expressing either FEZ1-GFP or GFP were kindly provided by T. Fujita (27) and were maintained in Dulbecco's minimal essential medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, penicillin, streptomycin, and Geneticin (400 g/ml). For virus preparation, JC virus-infected SVG-A cells were cultured for 2 weeks, harvested, and suspended in 10 mM Tris-HCl (pH 7.5) containing 0.2% bovine serum albumin. The cells were frozen and thawed three times and then treated for 16 h at 37°C with neuraminidase type V (Sigma) at 0.05 unit/ml. After an additional incubation for 30 min at 56°C, the cell lysate was centrifuged at 1000 ϫ g for 10 min, and the resulting supernatant was assayed for JCV by a hemagglutination assay (28) and stored at Ϫ80°C until use.
Protein Preparation-Recombinant baculovirus bearing Agno was constructed using the Gateway system (Invitrogen). Briefly, the coding region of Agno was amplified by PCR using a specific primer set 5Ј-G-GGGACAAGTTTGTACAAAAAAGCAGGCTGGATGGTTCTTCCGCC-AGCTGTC-3Ј and 5Ј-GGGGACCACTTTGTACAAGAAAGCTGGGTCT-ATGTAGCTTTTGGTTCA-3Ј and pBR-Mad1 (30) as a template. PCR products were subcloned into pDONR201 (Invitrogen), following cloning into pDEST10 (Invitrogen). Recombinant baculoviruses were pre-pared according to manufacturer's instruction. The histidine-tagged agnoprotein (His-Agno) was purified as follows: a 500-ml culture of Sf9 cells infected with the recombinant baculovirus for 3 days was harvested by centrifugation at 500 ϫ g for 10 min, and the cell pellet was resuspended in 5 volumes of the lysis buffer (600 mM NaCl, 20 mM Tris-HCl, pH 7.5, 0.5% (v/v) Triton X-100, 1 mM PMSF, and 0.05 mg/ml DNase) using the Teflon homogenizer. After rotation for 30 min at 4°C, the extract was clarified by centrifugation at 24,000 ϫ g for 20 min at 4°C. The supernatant was loaded onto the nickel-chelated Cellulofine (Seikagaku Corp., Tokyo, Japan) column pre-equilibrated with the lysis buffer without PMSF and DNase, and the column was washed with same buffer. After extensively washing with the solution (600 mM NaCl, 20 mM imidazole-HCl, pH 7.5), the bound proteins were eluted with the buffer (600 mM NaCl and 200 mM imidazole-HCl, pH 7.5). The anti-Agno antibody-conjugated Sepharose resin was added into the eluate, and the mixture was rotated overnight at 4°C. The resin was transferred into an empty column and washed with TBS containing 0.05% (w/v) Tween 20 (TBST). The bound proteins were eluted with the solution (0.1 M glycine-HCl, pH 2.8), and the elute was immediately neutralized by addition of 0.1 volume of 1 M Tris-base. The fractions containing His-Agno were pooled and dialyzed against TBS. The purified His-Agno was stored at 4°C until use.
Primary Antibodies-Mouse monoclonal antibodies to large T antigen (Ab-2), to ␣-tubulin, to pan-actin (MAB1501R), and to MAP2 (clone AP-20) were obtained from Oncogene Research Products (Uniondale, NY), Sigma, Chemicon International (Temecula, CA), and Roche Diagnostics, respectively. Goat polyclonal antibodies to GST were from Amersham Biosciences. Rabbit polyclonal antibodies to JCV agnoprotein and to JCV VP1 were produced as described previously (19,29). Rabbit polyclonal antibodies to enhanced GFP were kindly provided by N. Mochizuki. Mouse monoclonal antibodies to FLAG (M2) and horseradish peroxidase-conjugated mouse monoclonal antibodies to FLAG (M2) were from Sigma.
Immunoprecipitation was performed by incubation of cell lysates at 4°C first for 1 h with protein G-Sepharose FF beads (Amersham Biosciences) and then, after removal of the beads, for 4 h with antibodycoupled protein G-Sepharose FF beads. After washing with cell lysis buffer, the bead-bound proteins were subjected to immunoblot analysis.
Immunofluorescence Analysis-Cells were fixed with 3% paraformaldehyde in phosphate-buffered saline (PBS), permeabilized with 0.1% Triton X-100 in PBS, and incubated at room temperature with 5% dried skim milk in PBS (31). For detergent extraction, cells were washed once with PBS and twice with PHEM buffer (60 mM PIPES, 25 mM HEPES, pH 6.9; 10 mM EGTA, 2 mM MgCl 2 ) before extraction with 0.2% Saponin in PHEM buffer for 3 min on ice. The detergentinsoluble cell components remaining on the coverslips were then washed with PHEM buffer and fixed with methanol for 4 min at Ϫ20°C. The cells were then incubated first with primary antibodies and then with Alexa 488-or Alexa 594-labeled goat antibodies to rabbit immunoglobulin G or with Alexa 594-labeled goat antibodies to mouse immunoglobulin G (Molecular Probes, Eugene, OR). Nuclei were counterstained with propidium iodide (0.2 g/ml), and the cells were then observed with a confocal laser-scanning microscope (Olympus, Tokyo, Japan).
GST Precipitation Assay-GST fusion proteins of FEZ1 or its deletion mutants were expressed in Escherichia coli AD494 DE3 and purified with the use of glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech). For in vitro GST precipitation assays, GST or GST fusion proteins (50 pmol) were mixed with 10 l of the histidine-tagged agnoprotein in a final volume of 500 l with binding buffer (10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 0.5% Nonidet P-40, and 0.5 mM PMSF) and Complete protease inhibitor mixture and then incubated for 1 h at 4°C. After the addition of 10 l of 50% (v/v) glutathione-Sepharose 4B, the mixture was incubated for another 3 h at 4°C. The beads were separated by centrifugation and washed with binding buffer, and the bound proteins were subjected to immunoblot analysis with antibodies to agnoprotein.
Microtubule Cosedimentation Assay-Microtubule binding assays were performed as previously described (32)(33)(34)(35). In brief, GST fusion proteins of FEZ1 or its deletion mutants (0.5 g) in 50 l of a MES-based buffer (100 mM MES-NaOH (pH 6.8), 1 mM EGTA, 0.1 mM EDTA, 0.5 mM MgCl 2 , 1 mM dithiothreitol, 0.1 mM GTP) supplemented with 50 M Taxol (Sigma) were centrifuged at 100,000 ϫ g for 1 h at 25°C. Purified tubulin (Sigma) at 5 mg/ml in the MES-based buffer was polymerized by incubation for 30 min at 37°C in the presence of 50 M Taxol and GFP at a final concentration of 2.5 mM, and the polymerized microtubules were separated by centrifugation at 100,000 ϫ g for 30 min at 25°C. The supernatants containing the GST fusion proteins were then incubated with the polymerized microtubules (50 g) for 30 min at 37°C, after which the mixtures were each layered on top of 100 l of 30% sucrose in the MES-based buffer and centrifuged at 30,000 ϫ g for 30 min at 25°C. The resulting supernatants and pellets (washed once with the MES-based buffer) were subjected to immunoblot analysis.
For examination of the interaction between agnoprotein and microtubules, JCI cells were lysed in PTN buffer supplemented with Complete protease inhibitor mixture, and the cell lysate was centrifuged at 100,000 ϫ g for 30 min at 25°C. The resulting supernatant (60 l containing 100 g of protein) was added to a pellet of polymerized microtubules (50 g) and incubated for 30 min at 37°C in the presence of 50 M Taxol. Samples were then layered over 100 l of 30% sucrose in the MES-based buffer and centrifuged at 30,000 ϫ g for 30 min at 25°C. The resulting supernatants and pellets (washed once with the MES buffer) were subjected to immunoblot analysis.
Assay of Neurite Extension in PC12 Cells-PC12 cells stably expressing either FEZ1-GFP or GFP were transfected with pERedNLS-agno or the empty vector. At 24 h after transfection, the cells were incubated for 4 h in serum-free medium and then maintained for 48 h in serum-free medium supplemented with nerve growth factor (NGF) (Sigma) at a concentration of 50 ng/ml. The cells were washed with PBS, fixed with 3% paraformaldehyde at room temperature for 10 min, and then washed three times with PBS. Cells expressing agnoprotein were identified on the basis of their DsRed-positive nuclei. A change in cell morphology characterized by neurite outgrowth, cell flattening, and an increase in the size of the cell body was defined previously (27,36,37).
RT-PCR Analysis-For RNA extraction, cells grown in 6-well plates were washed once with PBS and harvested with trypsin-EDTA. Total RNA was isolated with the use of an RNeasy Mini kit (Qiagen, Valencia, CA), and portions (500 ng) were treated with 2 units of DNase I (Invitrogen) for 1 h at 37°C in a 10-l reaction mixture before incubation with 2.5 mM EDTA for 15 min at 65°C. The treated RNA (4 l) was subjected to RT with a Superscript first-strand synthesis system (Invitrogen). The absence of contamination with genomic DNA was verified by the addition of RNase-free water instead of Superscript II reverse transcriptase as a negative control for each sample. PCR was performed in a final volume of 50 l containing cDNA, Gene Taq universal buffer, 2.5 mM of each deoxynucleoside triphosphate, Gene Taq polymerase (Nippon Gene, Tokyo, Japan), and primers specific either for FEZ1 (5Ј-CACTGGTGAGTCTGGA-TG-3Ј and 5Ј-CGAGGTCCTCCATGGACTTGAAG-3Ј) or for ␤-actin (5Ј-TTGCCGACAGGATGCAGAA-3Ј and 5Ј-GCCGATCCACACGGA-GTACT-3Ј). The reaction mixtures were incubated at 94°C for 1 min and then subjected to 38 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s with a GeneAmp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA). All reactions were confirmed in at least three independent experiments.
Virus Inoculation and siRNAs Transfection-SVG-A sells were grown to 50% confluence on a 6-well plate and incubated with 1000 hemagglutination units of JCV in Dulbecco's minimal essential medium containing 2% fetal bovine serum for 24 h, At 24 h post inoculation of JCV, transfection of siRNA (100 pmol) were performed with Lipofectamine 2000 (Invitrogen) to the JCV-infected SVG-A cells. The cells were harvested and analyzed by Western blot analysis at 96 h post transfection. The results were confirmed by at least three independent experiments.

Identification of FEZ1 as an Agnoprotein-binding Protein-
We first performed a yeast two-hybrid assay with full-length JCV agnoprotein as the bait to identify proteins that interact with agnoprotein. Several positive clones were isolated from a human brain cDNA library and were found to encode a portion of FEZ1 that contains the three poly-glutamic acid regions and the coiled-coil (CC) domain (Fig. 1A). The largest cDNA clone encoded amino acids 32-392 of human FEZ1.
To examine the interaction between agnoprotein and FEZ1, we established a cell line, designated 293AG, that was derived from HEK293 cells and that expresses JCV agnoprotein (tagged with the hexahistidine and Myc epitopes at its NH 2 terminus) under the control of a tetracycline-responsive promoter. All 293AG cells expressed the recombinant agnoprotein (molecular size, ϳ14 kDa) within 3 h of exposure to doxycycline (Dox). In the absence of Dox, we failed to detect agnoprotein in the cells by immunoblot or immunocytofluorescence analysis (data not shown). Immunoprecipitation and immunoblot analysis of 293AG cells transfected with an expression vector for FLAG epitope-tagged FEZ1 revealed that FLAG-FEZ1 coprecipitated with agnoprotein (Fig. 1B).
The subcellular localization of exogenously expressed agnoprotein and FEZ1 in transfected HEK293 cells was examined by immunocytofluorescence analysis. Confocal microscopy revealed that agnoprotein immunoreactivity was present in the perinuclear region and extended into the cytoplasm in a meshlike pattern (Fig. 1C). In most cells, FEZ1 was detected throughout the cytoplasm and colocalized with agnoprotein only in the perinuclear region (Fig. 1C, upper panels). However, in some cells the localization of FEZ1 is mirrored by that of agnoprotein and well colocalized with agnoprotein (Fig. 1C, lower panels). This suggests that FEZ1 was recruited to the same location as agnoprotein as a result of interactions between FEZ1 and agnoprotein.
Association of Agnoprotein with Microtubules-Agnoprotein colocalized with microtubules in the perinuclear region of JCVinfected SVG-A cells ( Fig. 2A), consistent with our previous observations (38). To confirm the association between agnoprotein and microtubules, we performed a microtubule cosedimentation assay with lysates of JCI cells and microtubules polymerized in the presence of Taxol. Agnoprotein was indeed detected in the sedimented fraction only in the presence of microtubules (Fig. 2B). In contrast, the large T antigen of JCV, which was detected in the nucleus of JCV-infected cells (19), was present exclusively in the supernatant fraction even in the presence of microtubules (Fig. 2B), suggesting that the association of agnoprotein with microtubules is specific. However, it was not clear if agnoprotein binds directly to microtubules, because the cosedimentation assays were performed on cell extracts. To confirm the direct binding of agnoprotein to microtubules, we performed the assay using recombinant histidinetagged agnoprotein (His-Agno) from the insect cells infected with the recombinant baculovirus encoding His-Agno cDNA (Fig. 2C). We incubated the His-Agno in the MES-based buffer in the absence (Fig. 2D, lanes 1 and 6) 5 and 10). In the absence of polymerized microtubules, agnoprotein remains in the supernatant fraction (lanes 1-3). Only small amounts of agnoprotein were detected in the sedimented fraction in the presence of nocodazole-treated microtubules (lane 9), whereas a significant amount of agnoprotein associated with microtubule pellets in the presence of Taxol-polymerized microtubules (lane 10). These results demonstrate that agnoprotein binds directly to microtubules. To examine if agnoprotein colocalizes specifically with microtubules, JCV-infected cells (JCI cells) were treated with either cytochalasin D, which specifically depolymerizes actin fibers, or nocodazole, which depolymerizes microtubules (Fig. 2E). In cells treated with Me 2 SO or cytochalasin D, agnoprotein was still localized in the perinuclear region and cyto-plasm. In cells treated with nocodazole, however, agnoprotein was found in aggregates dispersed throughout the remainder of the cell. These observations suggest that agnoprotein colocalizes specifically with microtubules.
Agnoprotein-induced Dissociation of FEZ1 from Microtubules-We next examined whether a GST fusion protein of full-length FEZ1 interacted with microtubules in the microtubule cosedimentation assay. GST-FEZ1, but not GST, was detected in the sedimented fraction in the presence of polymerized microtubules (Fig. 3A). In addition, both ␣-tubulin and agnoprotein were detected in immunoprecipitates prepared with antibodies to GFP from lysates of Dox-treated 293AG cells expressing a FEZ1-GFP fusion protein (Fig. 3B). Together, these data suggested that FEZ1 directly interacts with microtubules.
To examine whether agnoprotein affects the interaction of FEZ1 with microtubules, we performed the microtubule cosedimentation assay with lysates of 293AG cells stably expressing FEZ1-GFP. The expression level of agnoprotein in the cells was varied by their incubation with Dox for 0, 3, 6, 12, or 24 h. The amount of FEZ1-GFP that cosedimented with microtubules decreased as the expression level of agnoprotein increased (Fig.  3C). Conversely, the amount of agnoprotein that precipitated with microtubules increased in a concentration-dependent manner. However, the amount of MAP2 that cosedimented with microtubules did not change. These observations suggested that agnoprotein induces the dissociation of FEZ1 from microtubules, and that competition is specific between agnoprotein and FEZ1.
Regions of FEZ1 That Mediate Interaction with Microtubules and Agnoprotein-To identify the region of FEZ1 that mediates its interaction with agnoprotein, we constructed a series of deletion mutants of FEZ1 as GST fusion proteins (Fig. 4, A and  B) and subjected them to GST precipitation assays. The deletion mutants of FEZ1 fused to GST were incubated with purified the histidine-tagged agnoprotein (His-Agno) expressed in insect cells and then precipitated with glutathione-Sepharose beads. Immunoblot analysis of bead-bound proteins revealed that wild-type FEZ1 and the deletion mutant of FEZ1 containing residues 192-392 (C192) interacted with agnoprotein to similar extents, whereas the deletion mutant of FEZ1 containing residues 163-296 (N296) precipitated progressively smaller amounts of agnoprotein, and no binding was observed with the COOH-terminal deletion mutant containing residues 1-191 (N191) and an NH 2 -terminal deletion mutant containing residues 297-392 (C297) (Fig. 4C). These observations suggested that agnoprotein binds directly to FEZ1, and the CC domain of FEZ1 is important for the association with agnoprotein.
We next attempted to delineate the region of FEZ1 responsible for association with microtubules. The deletion mutants of FEZ1 fused to GST were subjected to a microtubules cosedimentation assay. The FEZ1 C192 and FEZ1 C297 mutants retained the ability to bind to microtubules (Fig. 4D). In contrast, the FEZ1 N191 mutant failed to interact with microtubules. The FEZ1 N296 mutant detected in sedimented fraction in the absence of microtubules as well as presence of microtubules, showing that the FEZ1 N296 mutant could sediment by itself independently with microtubules. These results suggested that the COOH-terminal region (residues 297-392) of FEZ1 contribute to the interaction with microtubules.
Thus, we have demonstrated that the binding regions of agnoprotein and microtubules with FEZ1 did not overlap each other. To further investigate the mechanism how agnoprotein disrupts the interaction between FEZ1 and microtubules, we performed a GST precipitation assay using GST-fused FEZ1 proteins and cell lysates from the agnoprotein-inducible cell line (293AG cells). This demonstrated that precipitated ␣-tu-  3 and 6), and were subsequently incubated in the absence (lanes 2 and 5) or presence (lanes 1, 3, 4, and 6) of Dox for 24 h. Cell lysates were then subjected to immunoprecipitation (IP) with antibodies to agnoprotein (anti-agno), and the resulting precipitates were subjected to immunoblot analysis (IB) with antibodies to FLAG or to agnoprotein, as indicated (lanes 1-3). A portion of cell lysates corresponding to 2% of the input for immunoprecipitation was also subjected directly to immunoblot analysis with the same antibodies (lanes 4 -6). C, intracellular localization of recombinant agnoprotein and FLAG-FEZ1. HEK293 cells were transfected with pFLAG-FEZ1 and pcDNA4HisMax-agnoprotein. Forty-eight hours after transfection, the cells were stained with antibodies to agnoprotein (red, left) and to FLAG (green, center). The merged fluorescence images (Overlay, right) are also shown. Scale bars, 10 m. bulin with wild type FEZ1 or FEZ1 C192 mutant were remarkably attenuated in the presence of agnoprotein; however, precipitated ␣-tubulin with FEZ1 C297 mutant lacking the agnoprotein binding site was not altered even in the presence of  ). B, microtubule cosedimentation assay with a lysate of JCVinfected SVG-A cells. Cell lysate was incubated in the absence (Ϫ) or presence (ϩ) of microtubules (Mt) and Taxol (Taxol), after which the precipitates (Ppt) and supernatants (Sup) obtained by centrifugation of the incubation mixtures were subjected to immunoblot analysis with antibodies to agnoprotein and to JCV large T antigen. A portion of the cell lysate corresponding to 10% of the input to the sedimentation assay was also subjected directly to immunoblot analysis with the same antibodies. C, the purified histidine-tagged agnoprotein (arrowhead).

FIG. 3. Agnoprotein-sensitive interaction of FEZ1 with microtubules.
A, microtubule cosedimentation assay performed with GST or GST-FEZ1. The recombinant proteins were incubated with microtubules, after which the precipitates and supernatants of the incubation mixtures were subjected to immunoblot analysis with antibodies to GST. The positions of GST-FEZ1 (arrowhead) and GST (arrow) are indicated. B, coprecipitation of ␣-tubulin and agnoprotein with FEZ1-GFP. 293AG cells stably expressing FEZ1-GFP or GFP were incubated with Dox (for 24 h), lysed, and subjected to immunoprecipitation with antibodies to GFP, and the resulting precipitates were subjected to immunoblot analysis with antibodies to GFP, to ␣-tubulin, and to agnoprotein. A portion of the cell lysates corresponding to 5% of the input for immunoprecipitation was also subjected directly to immunoblot analysis with the same antibodies. C, agnoprotein-induced dissociation of FEZ1 from microtubules. 293AG cells stably expressing FEZ1-GFP (lanes 1-5) or GFP (lanes 6) were incubated with Dox for 0 h (lanes 1), 3 h (lanes 2), 6 h (lanes 3), 12 h (lanes 4), or 24 h (lanes 5 and 6) and were then subjected to the microtubule cosedimentation assay. The microtubule precipitates (left panel) as well as a portion of the cell lysates corresponding to 10% of the input to the sedimentation assay (right panel) were subjected to immunoblot analysis with antibodies to GFP, to agnoprotein, to MAP2, and either to ␣-tubulin or to actin. agnoprotein (Fig. 4E). These results suggested that the interaction of agnoprotein with FEZ1 at the CC domain is essential for disruption of interaction between microtubules and FEZ1 at the COOH-terminal region.

Inhibition by Agnoprotein of the Promotion of Neurite Outgrowth by FEZ1 in PC12
Cells-FEZ1 has been implicated in axonal outgrowth and fasciculation (23) and promotes neurite extension in rat pheochromocytoma PC12 cells (37). We have now shown that FEZ1 associates with microtubules, and cytoskeletal filaments, including microtubules, are the final common target of various signaling cascades that influence development of the growth cone and neurite extension (39). To examine the possible effect of agnoprotein on the promotion of neurite outgrowth by FEZ1 in PC12 cells, which cease proliferation and begin to extend neurites in response to stimulation with NGF, we transfected PC12 cells that stably express FEZ1-GFP or GFP with pERedNLS-agno or the empty vector. Agnoprotein-expressing cells were thus labeled by expression of DsRed in the nucleus. Agnoprotein significantly inhibited neurite extension in NGF-treated cells that expressed FEZ1-GFP, but it had no effect on neurite outgrowth in NGF-treated cells expressing GFP (Fig. 5). These results suggested that agnoprotein inhibits the promotion of neurite outgrowth by FEZ1 in PC12 cells.

FIG. 5. Agnoprotein-induced inhibition of the promotion of neurite outgrowth by FEZ1 in PC12 cells.
A and B, representative morphology of NGF-treated PC12 cells expressing FEZ1-GFP in the absence or presence of agnoprotein, respectively. The nuclei of transfected cells are colored red by DsRed. Scale bars, 50 m. C, PC12 cells stably expressing GFP or FEZ1-GFP were transfected with the bicistronic expression vector pERedNLS-agno (agno) or pERedNLS (vector) and subsequently exposed to NGF. The cells were examined by fluorescence microscopy, and those that had been successfully transfected were identified by their DsRed-positive nuclei. The percentage of cells that underwent a morphological change, characterized in part by flattening of the cell body and extension of neurites with a length at least twice the diameter of the soma, was determined by evaluation of Ͼ300 cells with DsRed-positive nuclei. Data are means Ϯ S.D. of values from three independent experiments. *, p Ͻ 0.05 for the indicated comparison (Student's t test). FIG. 4. Delineation of the regions of FEZ1 that mediate interaction with microtubules and agnoprotein. A, schematic representation of GST fusion constructs of wild-type (WT) FEZ1 and various FEZ1 deletion mutants. B, immunoblot analysis with antibodies to GST of the GST fusion proteins of FEZ1 deletion mutants purified from bacteria. Asterisks indicate the mature recombinant proteins. C, in vitro precipitation assay with the GST-FEZ1 deletion mutants and His-Agno. After incubation with His-Agno, the GST fusion proteins were precipitated with glutathione-Sepharose and bead-bound proteins were subjected to immunoblot analysis with antibodies to agnoprotein (upper panel) or to GST (lower panel). A portion of His-Agno corresponding to 10% of the input to the binding assay was also subjected directly to immunoblot analysis. D, microtubule cosedimentation assay performed with the GST fusion proteins of FEZ1 deletion mutants. The supernatant and precipitate fractions were subjected to immunoblot analysis with antibodies to GST. E, GST precipitation assay using GST-fused WT and mutant FEZ1 proteins and cell lysates from the agnoprotein-inducible cell line (293AG cells). After incubation with cell lysates, the GST fusion proteins were precipitated with glutathione-Sepharose, and bead-bound proteins were subjected to immunoblot analysis with antibodies to ␣-tubulin.

Inhibition by FEZ1 of JCV Protein Expression in SVG-A
Cells Inoculated with JCV-To examine the possible effects of FEZ1 on JCV infection in SVG-A cells, we established SVG-A cells that stably express FLAG-FEZ1. Given the lack of antibodies to FEZ1, we examined total (endogenous plus exogenous) FEZ1 expression by RT-PCR. The amount of FEZ1 mRNA in cells stably expressing FLAG-FEZ1 (FEZ1#1 and FEZ1#3 lines) was higher than that apparent in parental SVG-A cells or in cells stably transfected with the empty vector (Fig. 6A). The abundance of JCV agnoprotein and VP1 was markedly reduced in the cells expressing FLAG-FEZ1 compared with that apparent in control cells after inoculation with JCV (Fig. 6, B and C). To confirm if the effect for JCV infection in FLAG-FEZ1-expressing SVG-A cells depends on the expression of FEZ1, we used stealth siRNA (Invitrogen) to reduce the expression of FEZ1. At 24-h post inoculation of JCV, siRNAs (siFEZ1-301, siFEZ1-352, and Scramble) were transfected. The cells were harvested and analyzed by immunoblot at 96-h post transfection. The amount of FEZ1 mRNA in cells transfected with siFEZ1-301 or siFEZ1-352 siRNA was lower than that transfected with Scramble siRNAs (Fig. 6D). In the cells expressing FLAG-FEZ1, the abundance of JCV agnoprotein and VP1 was markedly increased in the cells transfected with siRNA to FEZ1 compared with Scramble siRNA. On the other hand, no difference was shown with cells transfected with siRNA to FEZ1 and Scramble in the cells stably transfected with the empty vector (Fig. 6E). FEZ1 thus suppressed the expression of JCV proteins in SVG-A cells. We also examined the transcriptional activity of JCV promoters in these various cell lines with a luciferase reporter assay but found no effect of FLAG-FEZ1 (data not shown).
Inhibition by FEZ1 of JCV Propagation in SVG-A Cells-Finally, we examined JCV-inoculated SVG-A cell lines by immunocytofluorescence analysis with antibodies to VP1. The number of VP1-positive cells did not differ significantly between either FEZ1#1 or FEZ1#3 cells and mock-transfected cells 3 days after inoculation (Fig. 7A). In contrast, the proportion of VP1-positive cells was significantly smaller for FEZ1#1 or FEZ1#3 cells than for mock-transfected cells 7 days after inoculation (Fig. 7), suggesting that FEZ1 suppressed the propagation of JCV in SVG-A cells. VP1 immunoreactivity was detected in the nucleus and cytoplasm of both parental and mock-transfected cells but was restricted to the nucleus in FEZ1#1 and FEZ1#3 cells 7 days after inoculation (Fig. 7B). These data suggested that FEZ1 suppressed the translocation of VP1 from the nucleus to the cytoplasm. We also analyzed the JCV inoculated cells by electron microscopy. This revealed that virion formation was intact in FEZ1-overexpressing cells (data not shown).

DISCUSSION
In the present study, we have identified FEZ1 as a binding partner of JCV agnoprotein. FEZ1 is a brain-specific protein and a mammalian homolog of C. elegans UNC-76 (23). Agno- protein interacted with FEZ1 in both yeast and mammalian cells and colocalized with both FEZ1 and microtubules in the perinuclear region of mammalian cells. We previously showed that agnoprotein colocalizes with microtubules (38), and we here confirmed the binding of agnoprotein to microtubules with a microtubule cosedimentation assay. Microtubules are a major component of the cytoskeleton in growing axons, and, together with their associated molecules, they play an important role in axon outgrowth (39). We have now shown that FEZ1, which promotes axon outgrowth in PC12 cells (37), also binds to microtubules in the microtubule cosedimentation assay and that the interaction between FEZ1 and microtubules is disrupted by agnoprotein. Regions of FEZ1 that mediate interaction with microtubules and agnoprotein are not the same, and the binding of agnoprotein with the CC domain of FEZ1 plays a pivotal role in disruption of interaction between microtubules and FEZ1. These results suggest that the interaction of agnoprotein with FEZ1 might lead to some conformational changes in the COOH-terminal region of FEZ1, resulting in lost of its ability to combine with microtubules.
The promotion of neurite outgrowth by FEZ1 in PC12 cells was also inhibited by agnoprotein. These results suggest that FEZ1 might promote neurite extension through interaction with microtubules and that this interaction is sensitive to agnoprotein. Overexpression of FEZ1 in human glial cells inhibited the production of JCV agnoprotein and VP1 without affecting transcription from JCV promoters. Although the number of JCV-infected cells did not differ between SVG-A cells stably expressing FEZ1 and control cells at 3 days after inoculation with JCV, the proportion of VP1-positive cells was markedly reduced by FEZ1 overexpression at 7 days. These results suggest that overexpression of FEZ1 influenced the late phase of JCV infection but did not affect the early phase, including the attachment of JCV to cells, its entry into the cytoplasm, and transcription of the JCV genome. Whereas VP1 was present in both the nucleus and cytoplasm of control cells at 7 days after inoculation with JCV, it was restricted to the nucleus of cells overexpressing FEZ1. Given that the antibodies to VP1 used for these experiments recognize both VP1 monomers and mature virus particles, our results suggest that both VP1 monomers and mature JCV virions were restricted to the nucleus of cells stably expressing FEZ1. A mutational analysis of SV40 agnoprotein similarly showed that this protein plays an important role in the release of progeny virions from infected cells and in viral propagation (14 -16). The transport of progeny virions of SV40 from the nucleus to the cell surface depends on intracytoplasmic vesicular transport (40). We have detected JCV virions both in the cytoplasm of infected cells and in the surrounding extracellular space in the apparent absence of disruption of the cell membrane both in progressive multifocal leukoencephalopathy lesions and in cell cultures (19). These observations indicate that JCV virions might be released from host cells by a specific mechanism and that FEZ1 might inhibit JCV release from infected human cells.
UNC-76 of Drosophila, which is a homolog of FEZ1, binds specifically to the tail domain of kinesin heavy chain in the yeast two-hybrid system and copurification assays. Furthermore, immunostaining and genetic analyses have demonstrated that UNC-76 function is required for axonal transport in the Drosophila nervous system (41). Interestingly, we observed that FEZ1 interacted with KIF3A that is a member of KIF3 family by immunoprecipitation assay (data not shown). These observations thus suggest that UNC-76 and FEZ1 play an essential role in kinesin-mediated transport pathways. Kinesin is a plus end-directed microtubule motor that facilitates the movement of vesicles, messenger ribonucleoproteins, and organelles. It was first identified in squid axoplasm as a protein that facilitates ATP-dependent vesicle movement along microtubules (42,43). Subsequent molecular, genetic, and biochemical studies have shown that kinesin is required for intracellular transport in eukaryotes in many cellular contexts (44 -46).
The requirement for kinesin-based transport is especially prominent in polarized cells such as neurons. The polarity complex comprised of PAR3, PAR6, and atypical protein kinase C (PKC) contributes to polarity determination in many tissues and cells (47,48). FEZ1 was also identified in a yeast twohybrid assay for proteins that bind the regulatory domain of the rat atypical PKC isoform PKC (37). Whereas nonphosphorylated FEZ1 is associated with both cytosolic and membrane fractions of COS-7 cells, its phosphorylation by PKC induces the redistribution of membrane-bound FEZ1 to the cytosol. In addition, the phosphorylation of FEZ1 by PKC stimulates  's t test). B, immunostaining with antibodies to VP1 (green) of the indicated SVG-A cell lines at 7 days after inoculation with JCV. Cell nuclei were counterstained red. Scale bars, 50 m. neurite outgrowth in PC12 cells. These observations suggest that FEZ1, like UNC-76, might associate with kinesin and is essential for PKC-dependent neuronal differentiation.
FEZ1 also interacts in neural cells with the protein Disrupted-In-Schizophrenia 1, which was identified as the product of a gene disrupted by a (1;11)(q42.1;q14.3) chromosomal translocation that segregates with schizophrenia in a Scottish family. Furthermore, interaction of Disrupted-In-Schizophrenia 1 and FEZ1 appears to take place on or near the actin cytoskeleton and is up-regulated during neurite outgrowth (49).
The actin and tubulin cytoskeletons are a final common target of many signaling cascades that influence neuronal development (39), and cytoskeletal dynamics and polarized axonal transport appear to be closely related (50,51). Our observation that FEZ1 associates with microtubules, combined with its possible interaction with kinesin and the previous finding that it promotes neurite outgrowth, suggest that FEZ1 may play a key role downstream of atypical PKC and the polarity complex in the regulation of vesicle trafficking that contributes to neurite extension.
Although the release of mature progeny virions of SV40 and JCV assembled in the nucleus has been thought to occur through disintegration or rupture of infected cells, these viruses do not encode an enzyme that is able to cause cell lysis. In addition, progeny virions of SV40 have been shown to be transported from the nucleus to the cell surface in a manner dependent on vesicular transport (40). The tegument protein US11 of herpes simplex virus, another DNA virus, interacts with conventional kinesin heavy chain, and this interaction has been thought to be important for anterograde transport of nonenveloped nucleocapsids in axons (52). The transport and release of large numbers of virus particles are undesirable for host cells. It is thus possible that FEZ1 serves a potentially protective function in JCV-infected cells by inhibiting the release of progeny virions. Interestingly, expression of FEZ1 in JCV-permissible cells such as SVG-A cells or IMR-32 cells is much less than nonpermissible neuronal cells (53) such as SH-SY5Y cells (data not shown). This suggests that expression levels of FEZ1 might be a factor that determines JCV tissue specificity. Conversely, the viral agnoprotein may inhibit the physiological function of FEZ1 by inducing its dissociation from microtubules, thereby promoting the effective transmission of progeny viruses.