INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To initiate a successful infection, animal viruses bind to the cell surface, penetrate into the cytosol, and target their genome to the sites of viral transcription and replication. For many viruses this is the host nucleus (Whittaker et al., 2000). Particular neurotropic viruses that enter at the presynaptic plasma membrane, such as herpes simplex viruses, are transported over long distances because the site of entry is far away from the nucleus. Herpes simplex virus type 1 (HSV1) is a human pathogen that initially replicates in epithelial cells of the oral cavity. Amplified virus enters neurons and is transported to the neuronal nuclei located in the trigeminal ganglion (reviewed in Enquist et al., 1998). After lytic infection of some neurons, a latent infection is established (Wagner and Bloom, 1997).

We have calculated that it would take 231 years for a herpes virus capsid to diffuse by 10 mm in the axonal cytoplasm (Sodeik, 2000). High concentrations of protein, the cytoskeleton, and organelles cause molecular crowding in the cytoplasm, which effectively restricts free diffusion of molecules larger than 500 kDa (Luby-Phelps, 2000). Thus, virions and subviral particles are transported by active processes. Besides hijacking vesicular transport during endocytosis and secretion, viruses also exploit the host's cytoskeleton directly for their itinerary (Sodeik, 2000; Ploubidou and Way, 2001).

HSV1 virions consist of four structural components: DNA, capsid, tegument, and envelope (Steven and Spear, 1997; Zhou et al., 2000). The icosahedral capsid with a diameter of 125 nm surrounds the double-stranded viral DNA of 152 kb. The tegument, the hallmark of all herpes viruses, is an amorphous layer of ~20 proteins. It is localized between the capsid and the viral envelope that contains ~12 membrane proteins.

For cell entry the envelope of HSV1 fuses with the plasma membrane. Different molecules such as heparan sulfate proteoglycans, members of the tumor necrosis receptor family (HVEM), and the immunoglobulin family (nectins) serve as receptors for the HSV1 viral glycoproteins gB, gC, and most importantly gD (reviewed in Spear et al., 2000). The fusion of the viral envelope with the plasma membrane is mediated by the viral glycoproteins gB, gD, gH, and gL (Spear et al., 2000).

All tegument proteins and the capsid with the DNA are released into the cytosol. In epithelial cells and in axons of cultured neuronal cells, incoming cytosolic capsids are transported along microtubules (MTs) to the nucleus (Kristensson et al., 1986; Topp et al., 1994; Topp et al., 1996; Sodeik et al., 1997). Electron microscopy and careful quantification demonstrated that ~70% of cytosolic capsids bind to nuclear pores and that concomitantly these capsids have lost their electron-dense core (Sodeik et al., 1997). Using an in vitro uncoating assay, Ojala et al. (2000) demonstrated that capsid binding to the nucleus requires importin- and that the release of the viral DNA is triggered by the interaction with the nuclear pore. Transcription, viral replication, and capsid assembly take place in the nucleus (for reviews see Steven and Spear, 1997; Roizman and Knipe, 2001).

MTs are polar hollow protein cylinders of tubulin with a fast-growing and -shrinking plus end usually located toward the cell periphery and a minus-end mostly stabilized by attachment to the centrosome, the major microtubule organizing center (MTOC; Nogales, 2000). Most if not all minus-end-directed MT transport is mediated during interphase by dynein motors, whereas kinesins transport cargo toward the opposite direction (Vallee and Sheetz, 1996; Hirokawa, 1998). Cytoplasmic dynein is a 20 S MT-activated ATPase consisting of two dynein heavy chains (DHC), two intermediate chains (DIC), four light intermediate chains (DLIC) and four different classes of light chains (DLC; Karki and Holzbaur, 1999; King, 2000). Dynein is responsible for the perinuclear localization of several organelles around the MTOC and retrograde organelle transport in axons and is active during mitosis (Vallee and Sheetz, 1996; Hirokawa, 1998).

In many cases dynein is assisted by a second 20 S protein complex, called dynactin (Vallee and Sheetz, 1996; Karki and Holzbaur, 1999). It consists of 2 copies of p150^Glued, 4 molecules of dynamitin, p62, ~10 copies of Arp1 (actin-related-protein 1), possibly 1 conventional actin, Arp11, and actin capping protein (p37 and p32), p27, p25, and p24 (Holleran et al., 1998; Eckley et al., 1999). p150^Glued can bind directly to DIC and thus link dynein to dynactin (Karki and Holzbaur, 1995; Vaughan and Vallee, 1995). Dynamitin, at high concentrations after transient transfection, dissociates the dynactin complex (Echeverri et al., 1996; Eckley et al., 1999). Excess dynamitin affects all tested dynein-mediated transports in vivo and in vitro: e.g., spindle organization, chromosome transport, and the subcellular localization of several membrane organelles (Echeverri et al., 1996; Burkhardt et al., 1997; Presley et al., 1997; Valetti et al., 1999; Sharp et al., 2000).

Quantitative immunoelectron microscopy showed that DHC colocalizes with incoming herpes virus capsids (Sodeik et al., 1997). Here, we demonstrate that incoming HSV1 capsids also colocalized with DIC and the p150^Glued subunit of dynactin. To test whether HSV1 capsids use cytoplasmic dynein for their transport to the nucleus, we transiently transfected cells with dynamitin, subsequently challenged them with HSV1, and measured virus binding, internalization, capsid transport to the nucleus, and immediate-early viral gene expression. High concentrations of dynamitin and dynamitin-GFP clearly reduced the number of viral capsids transported to the nucleus compared with untransfected cells. Because fewer capsids reached the nucleus, presumably fewer viral genomes were delivered to the nucleoplasm, and the amount of viral protein synthesis was reduced. Overexpression of dynamitin did not downregulate virus receptors at the plasma membrane, because both virus binding and internalization were not reduced. We propose that incoming HSV1 capsids are propelled by dynein along MTs and that functional dynactin is required for their efficient transport.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cells and Antibodies

PtK₂ cells (ATCC CCL-56) were grown in 10%, Vero cells (ATCC CCL-81) in 7.5%, and BHK-21 cells (ATCC CCL-10) in 15% fetal calf serum. The media contained MEM, 2 mM glutamine, nonessential amino acids, and with the exception of the PtK₂ medium 100 U/ml penicillin and 100 µg/ml streptomycin.

Dynamitin expression was analyzed using mAb 50.1 (Paschal et al., 1993) or a rabbit anti-myc antibody (Gee et al., 1997), DHC was detected with affinity-purified rabbit pAb (Vaisberg et al., 1993), DIC with a rabbit pAb L5 (Vaughan and Vallee, 1995), p150^Glued with rabbit pAbs D'Artagnon, Aramis, or Portos (Vaughan and Vallee, 1995), and the cation-independent mannose-6-phosphate receptor with a rabbit pAb (Griffiths et al., 1988). MTs were visualized using mouse mAb 1A2 (Kreis, 1987) and actin filaments with TRITC-Phalloidin (Sigma-Aldrich, Schnelldorf, Germany). We used preadsorbed rabbit pAbs raised against DNA-containing capsids (anti-HC) and empty capsids (anti-LC; Cohen et al., 1980) and a mouse mAb 5C10 against VP5 (Newcomb et al., 1996) to detect incoming viral capsids by immunofluorescence microscopy (Sodeik et al., 1997). Immediate-early viral gene expression was measured with a mouse mAb against ICP4 (Showalter et al., 1981). Viral glycoproteins were labeled with mouse mAb DL6 against gD and rabbit pAb R68 against gB (Eisenberg et al., 1985, 1987). Secondary antibodies were purchased from Dianova (Hamburg, Germany).

Virological Techniques

Preparation of Stock Virus Cold and radioactively labeled virus stocks of HSV1 were prepared as described (Sodeik et al., 1997). We used wild-type strain F (ATCC VR-733) and two HSV1 mutants: strain [KOS]tk12, which expresses the LacZ gene controlled by the immediate-early ICP4 promoter (Warner et al., 1998) and strain R7202, which lacks the majority of the glycoprotein E codons including the start codon and therefore does not exhibit viral Fc receptor activity (Baines and Roizman, 1993).

Plaque Assay Virus was diluted in 10-fold steps in RPMI with 0.2% wt/vol BSA, 20 mM HEPES, pH 7.0 (RPMI/BSA) and incubated in six-well dishes with just-confluent Vero cells for 1 h at room temperature on a rocking platform. The inoculum was removed, and 2 ml/well normal growth medium containing 10 µg/ml pooled human IgG (Sigma-Aldrich) was added. The cells were further cultured for 2 d at 37°C, 5% CO₂ and fixed in absolute methanol. After incubation with a mAb to gD (DL6) and a secondary anti-mouse antibody conjugated to alkaline phosphatase, the cells were washed with TSM (100 mM Tris-HCl, pH 9.5, 100 mM NaCl, 5 mM MgCl₂) and treated with 0.2 mM nitroblue tetrazolium chloride and 0.8 mM 5-bromo-4-chloro-3-indolyl phosphate until dark plaques became visible. We routinely obtained titers around 10¹⁰ PFU/ml for cold wild-type and HSV1[KOS]tk12 virus and 10⁷ PFU/ml for ³H-thymidine labeled virus.

Virus Infection Control and transfected Vero, BHK, or PtK₂ cells were inoculated with virus in RPMI/BSA for 2 h on ice to allow virus binding. After three washes with ice-cold RPMI/BSA, they were shifted to growth medium at 37°C and 5% CO₂. In those experiments analyzing the subcellular localization of incoming viral particles, 0.5 mM cycloheximide was added to prevent synthesis of new viral proteins (Sodeik et al., 1997). When nocodazole (50 µM) was used to depolymerize MTs, cells were pretreated for 1 h at 37°C, and the drug was present during all further incubation steps. For immunofluorescence microscopy in 24-well plates we used 8 × 10⁶ PFU/well HSV1 for entry experiments, 1 × 10⁷ PFU/well for virus binding experiments, and 4 × 10⁵ PFU/well for ICP4 experiments.

Light Microscopy

Cells grown on coverslips were fixed with 3% (wt/vol) paraformaldehyde (PFA in PBS) for 20 min followed by 50 mM NH₄Cl/PBS for 10 min and 0.1% Triton X-100/PBS for 5 min. For colocalization studies and visualization of MTs, cells were fixed with PHEMO-fix (3.7% [wt/vol] PFA, 0.05% [wt/vol] glutaraldehyde, 0.5% Triton X-100 in PHEMO buffer) either at room temperature or at 37°C for 10 min, and washed with PHEMO buffer (68 mM PIPES, 25 mM HEPES, pH 6.9, 15 mM EGTA, 3 mM MgCl₂, 10% [vol/vol] DMSO) followed by 50 mM NH₄Cl/PBS for 10 min.

In most experiments we used 10% (vol/vol) goat serum with 5 mg/ml BSA as blocking reagent and performed the immunolabeling essentially as described (Sodeik et al., 1997). For the experiments described in Figure 1 we used 0.2 mg/ml human immunoglobulins (IgGs, I4506; Sigma, Taufkirchen, Germany) with 5 mg/ml BSA to block nonspecific protein binding.

View larger version (142K): [in this window] [in a new window]   Figure 1.   (A) Incoming wild-type HSV1 capsids colocalize with p150Glued and DIC. A significant fraction of incoming HSV1 capsids colocalizes with p150Glued and DIC (yellow dots in the overlays, c and f). This apparent colocalization is not due to binding of rabbit IgG to the herpes virus Fc receptor on the surface of infected cells because there is no signal from a nonimmune rabbit serum (i). Moreover, immunoblots confirmed that there is no unspecific cross-reactivity of the anti-dynactin or anti-dynein antibodies to structural viral proteins (our unpublished results). Immunofluorescence microscopy of PtK2 cells infected with wild-type HSV1 for 1 h in the presence of 0.5 mM cycloheximide. Cells were fixed with PHEMO-fix at room temperature, blocked with 0.2 mg/ml pooled human IgG and 5 mg/ml BSA and double-labeled with anti-p150Glued (pAb Portos; a), anti-DIC (pAb L5, d) or a nonimmune rabbit serum (serum, g) and anti-VP5 (mAb 5C10; b, e, and h). Note, that the anti-VP5 antibodies cross-reacted in PtK2 cells weakly with the cortical actin and stress fibers. (B) Incoming HSV1 capsids of a gE deletion mutant colocalize with p150Glued and DIC. Incoming HSV1 capsids of the mutant strain R7202 that does not express a viral Fc receptor colocalize with p150Glued and DIC (yellow dots in the overlays, c and f). Moreover, there is no colocalization (i) after double labeling with a nonimmune rabbit serum (g) and capsid antibodies (h). Immunofluorescence microscopy of PtK2 cells infected with HSV1 deleted for gE (R7202) for 1 h in the presence of 0.5 mM cycloheximide. Cells were fixed with PHEMO-fix at room temperature, blocked with 0.2 mg/ml pooled human IgG and 5 mg/ml BSA and double-labeled with anti-p150Glued (pAb Portos; a), anti-DIC (pAb L5, d), or a nonimmune rabbit serum (g) and anti-VP5 (mAb 5C10; b, e, and h).

HSV1 carries in its envelope a Fc-receptor with strong affinity for human IgGs, decreasing affinity for rabbit, sheep, and goat and no affinity for murine IgGs (reviewed in Dubin et al., 1992). The degree of the Fc-receptor-specific signal in the absence of human IgGs depended strongly on the protocol used: it was most prominent under the conditions we found optimal for the colocalization studies, namely after PHEMO fixation at room temperature, but surprisingly only weak after PHEMO fixation at 37°C or after PFA fixation followed by TX-100 (our unpublished results). Initial double-labeling experiments for dynein or dynactin with HSV1 capsids demonstrated that in the presence of human IgGs, primary rabbit and secondary goat antibodies bound only with their antigen-binding domain but not with their Fc-domain to infected cells. Without human IgGs, the membrane of incoming virions was also detected by preimmune and immune rabbit sera but not by unspecific mouse antibodies (our unpublished results).

The cells were examined with a fluorescence microscope (DM IRB/E; Leica, Wetzlar, Germany), and micrographs were taken on Kodak TMAX-400 (Eastman Kodak, Rochester, NY) or Ilford HP5 400 films (Ilford, Cheshire, United Kingdom). All digitized images (using a Nikon LS-1000 35-mm film scanner; Tokyo, Japan) were image processed using Adobe Photoshop version 5.5 (San Jose, CA). Colocalization and expression levels were analyzed using a digital interline charge coupled device camera (MicroMax-5MHz-782Y; Princeton Instruments Inc., Princeton, NJ) and the Metamorph software version 4.01 (Universal Imaging Corporation, West Chester, PA). Dynamitin-GFP and GFP-expressing cells were scored as high expressors (see Figure 8B) when their average gray values in the cytoplasm were above 400 (exposure time: 0.2 s) using the function "show region statistics" of the Metamorph software. Cells overexpressing dynamitin (mAb 50.1) were scored as high expressors when their average gray values in the cytoplasm were above 600 (exposure time: 0.5 s).

Transient Transfection

Plasmids For transient transfections we used the plasmids pEGFP-N1 (Clontech Laboratories Inc., Palo Alto, CA) expressing enhanced green fluorescent protein (GFP) and p50 expressing myc-tagged dynamitin (Echeverri et al., 1996). For the dynamitin-GFP-expressing construct (p50-GFP), the full dynamitin cDNA sequence was inserted upstream of GFP into the pEGFP-N1 vector between the EcoRI and the BamHI sites. The proper restriction sites were created by PCR, and subsequent products were sequenced to ensure that no mutation was created. All constructs are under the control of the cytomegalovirus immediate-early promotor (Clontech).

Experiments Measuring Virus Binding, Internalization, and Protein Synthesis PtK₂ cells were seeded in 10-cm diameter culture dishes at a density of 5 × 10⁵. Twenty-four hours later the cells were transfected with 30 µg DNA per dish using calcium phosphate (Sambrook et al., 1989). After 24 h, the cells were washed once with PBS, and 10 ml/dish normal growth medium was added for 19-22 h.

Experiments Analyzed by Immunofluorescence Microscopy Transfections were made with calcium phosphate or lipofectamine reagent (Life Technologies, Karlsruhe, Germany). For the latter, cells were seeded onto coverslips (12-mm diameter) in a 24-well cell culture dish at a density of 4 × 10⁴ (Vero) or 3 × 10⁴ (PtK₂) cells per well. After 16-18 h, the cells were washed twice with serum-free, antibiotic-free MEM and incubated with 300 µl/well serum-free, antibiotic-free MEM containing 1.5 µl/well lipofectamine and 150 ng DNA for 5 h. The transfection mixture was removed, and 1 ml/well normal growth medium was added for 24-28 h. For transfections with calcium phosphate, PtK₂ cells were seeded onto coverslips (12-mm diameter) in a 24-well cell culture dish at a density of 2 × 10⁴ cells per well. After 24 h cells were transfected with 1.1 µg DNA per well. Twenty-four hours later 1 ml/well normal growth medium was added for 19-22 h.

Immediate-early Viral Gene Expression

Immediate-early viral gene expression was analyzed using the mutant HSV1(KOS)tk12, which expresses the bacterial LacZ gene coding for the enzyme -galactosidase under the control of the immediate-early ICP4 promotor of HSV1 (Warner et al., 1998). PtK₂ cells transfected for 44 h with dynamitin-GFP or GFP were inoculated with 2 ml per 10-cm dish RPMI/BSA containing 4-8 × 10⁶ PFU of HSV1(KOS)tk12 for 2 h on ice and then shifted to 37°C for 3-4 h. The cells were harvested by trypsinization, and trypsin was immediately inhibited by adding trypsin inhibitor, and further protein synthesis by cycloheximide. Transfection efficiencies were determined by flow cytometry (FACSCalibur; Becton Dickinson, Heidelberg, Germany), cell densities by BCA assay (Pierce, Rockford, IL) after lysis in 1% SDS/PBS, and -galactosidase activities after lysis in 0.5% TX-100/PBS with 1 mg/ml BSA and protease inhibitors using O-nitrophenyl--D-galactopyranoside as substrate. -Galactosidase activity per cell was normalized to GFP-transfected cells.

Quantification of Virus Binding and Internalization

We assayed for virus binding and internalization essentially as described using a protease protection assay (Sodeik et al., 1997). PtK₂ cells transfected for 44 h with dynamitin-GFP or GFP were inoculated with 2 ml per 10-cm dish RPMI/BSA containing ³H-thymidine labeled HSV1 (5-10 kBq and 8 × 10⁶ to 1.6 × 10⁷ PFU). Cell-associated radioactivityrepresenting the amount of bound virionswas determined by scintillation counting. Virus binding was expressed as radioactivity per cell and normalized to GFP-transfected cells. To assay for virus internalization, ³H-thymidine-labeled virus was bound to transfected cells at 4°C for 2 h. The cells were washed to remove unbound virus, shifted to 37°C for 30 min, transferred back to ice, and washed with ice-cold RPMI. Cell associated radioactivity after proteinase K treatment was determined by scintillation counting (Sodeik et al., 1997). In these virus binding and internalization experiments, the transfection efficiencies were measured by flow cytometry and cell densities using a hemocytometer.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Colocalization of Incoming HSV1 Capsids with Cytoplasmic Dynein and Dynactin

We have shown previously by quantitative immune electron microscopy that in Vero cells incoming cytosolic HSV1 capsids colocalize with DHC (Sodeik et al., 1997). Here, we used PtK₂ cells that are extremely flat in the periphery to analyze the subcellular distribution of HSV1 relative to dynein and dynactin. In uninfected cells, anti-DHC, anti-DIC, and anti p150^Glued revealed a weak diffuse labeling throughout the cytoplasm that was strongest around the nucleus; moreover, centrosomes and numerous tubules and vesicles concentrated around the nucleus were labeled (our unpublished results). These structures might represent host organelles that use dynein for transport along MTs (Burkhardt et al., 1997; Sodeik et al., 1997; Harada et al., 1998; Valetti et al., 1999; Habermann et al., 2001).

After 1 h of infection, a mouse mAb to the capsid protein VP5 labeled numerous small fluorescent spots that represented individual capsids distributed over the entire cytoplasm (Figure 1, A and B, b, e, and h). After 2 h and more so after 3 h, the majority of capsids had accumulated at the nucleus (see Figure 7). Numerous small dots of dynactin (Figure 1Ac), DHC (our unpublished results), and DIC (Figure 1Af) colocalized with viral capsids, whereas there was no colocalization after double labeling with the mouse anti-VP5 and a preimmune rabbit serum (Figure 1Ai).

HSV1 carries in its envelope the viral protein complex gE/gI that has a strong Fc-receptor binding activity for human IgGs, decreasing affinity for rabbit, sheep, and goat, but does not bind to murine IgGs (reviewed in Dubin et al., 1992). However, none of eight mouse monoclonal antibodies generated against different subunits of dynein (DHC, DIC, DLC) or dynactin (p150^Glued, dynamitin) showed any colocalization with viral capsids despite testing several fixation and permeabilization protocols. In the presence of rabbit or goat sera, we therefore blocked the Fc-receptor with human IgGs (cf. MATERIALS AND METHODS). Moreover, we used the HSV1 mutant R7202, which is deleted for gE and does not contain a Fc-binding activity (Baines and Roizman, 1993). After 1 h of infection with R7202, numerous viral capsids were also labeled for dynactin and dynein, in the presence (Figure 1B) and also absence (our unpublished results) of human IgGs.

Approximately 15-20% of incoming capsids from wild-type HSV1 and the mutant R7202 were labeled with antibodies to dynactin (20% for p150^Glued; n = 130) and dynein (15% for DIC; n = 100). The lacking reactivity of monoclonal antibodies and the incomplete colocalization using polyclonal rabbit sera directed against dynein or dynactin subunits might suggest that there was steric hindrance and thus only limited epitope access in a putative ternary complex of capsids, dynein, and dynactin. Alternatively, it is possible that only a subset of viral capsids binds to dynein and/or dynactin at a given time point. Because most subunits of dynein and dynactin only exist in 20 S complexes and not as soluble proteins, these data showed that both protein complexes, dynein and dynactin, were at least transiently present on incoming viral capsids.

Immediate-early Viral Gene Expression after Overexpression of Dynamitin

To test whether functional dynein and dynactin were required during HSV1 entry, we transfected PtK₂ cells with dynamitin, which inhibits many dynein-mediated transport processes (Echeverri et al., 1996; Burkhardt et al., 1997; Presley et al., 1997; Valetti et al., 1999). In cells overexpressing dynamitin or dynamitin-GFP, mannose-6-phosphate receptor containing organelles were scattered over the entire cytoplasm rather than concentrated in the perinuclear region (our unpublished results), indicating that the function of dynein was disrupted (Burkhardt et al., 1997; Valetti et al., 1999). We routinely obtained transfection efficiencies of 65-75% for dynamitin-GFP and 70-85% for GFP as measured by flow cytometry (our unpublished results).

We next infected transfected PtK₂ cells with a HSV1 mutant expressing -galactosidase under the control of an immediate-early HSV1 promoter (HSV1[KOS]tk12; Warner et al., 1998), and the enzyme activity was quantified as an indicator for immediate-early viral gene expression. -galactosidase activity was highest in untransfected cells and lowest in cells transfected with dynamitin-GFP (Figure 2). Overexpressing dynamitin-GFP reduced the amount of -galactosidase by 25% compared with overexpression of GFP alone.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Quantification of immediate-early viral gene expression in transfected cells. -Galactosidase activity per cell was reduced in dynamitin-GFP-overexpressing cells compared with untransfected or GFP-transfected cells. Two-sided Student's t test confirmed that viral protein synthesis is significantly lower in dynamitin-GFP-transfected cells compared with either GFP-transfected cells (p = 1.13 × 104) or to untransfected control cells (p = 2.1 × 105). There was no significant difference in -galactosidase expression between control and GFP-transfected cells (p = 1.58 × 101). The mean values for five independent experiments each performed in duplicates are shown.

To analyze single cells, we infected PtK₂ cells with wild-type HSV1 and double-labeled them with antibodies to the transiently expressed proteins and ICP4, an immediate-early, nuclear herpes virus protein (Everett, 2000). After overexpression of dynamitin and dynamitin-GFP there were about half as many cells labeled for ICP4 compared with GFP expressing or untransfected cells (Figure 3, A and B). Thus, the expression of ICP4 and -galactosidase, both under the control of the ICP4 promotor, were reduced after overexpressing dynamitin or dynamitin-GFP compared with controls. Inhibition of immediate-early viral gene expression might be due to changes in 1) the MT-network, 2) virus binding to the cell surface, 3) virus internalization, or 4) a reduced cytosolic transport of incoming capsids to the nucleus.

View larger version (57K): [in this window] [in a new window]   Figure 3.   Overexpression of dynamitin reduces immediate-early viral gene expression. (A) Overexpression of dynamitin (a) or dynamitin-GFP (b) reduced the immediate-early viral gene expression compared with control (a-c) or GFP-transfected cells (c). Immunofluorescence microscopy of PtK2 cells transfected with dynamitin or dynamitin-GFP and 30 h later infected with HSV1 for 3 h. Cells were fixed with PFA and either double-labeled with anti-myc (a; top panel) to detect transfected cells and anti-ICP4, an immediate-early protein of HSV1 (a-c; bottom panels) or single-labeled with anti-ICP4 (b and c; bottom panels), and the transfected proteins were detected by their intrinsic GFP fluorescence (b and c; top panels). (B) Quantification of viral ICP4 synthesis after transfection. The overexpression of dynamitin reduced immediate-early viral gene expression. Two-sided Student's t test confirmed that ICP4 expression is significantly lower in dynamitin (p = 2.52 × 103) or dynamitin-GFP (p = 3.57 × 103) transfected cells compared with GFP-transfected cells. There was no significant difference in ICP4 expression between untransfected and GFP-transfected cells (p = 1). The experiment described in A was quantitated (three independent experiments; altogether >500 cells analyzed for each condition). Cells overexpressing dynamitin, dyna-mitin-GFP, or GFP and untransfected cells were divided into two classes: cells displaying a nuclear ICP4 signal and cells not displaying a nuclear ICP4 signal.

The Cytoskeleton after Transient Transfection

Dynactin contains conventional actin, Arp1 and Arp11 (Schafer et al., 1994; Eckley et al., 1999). We therefore tested whether dynamitin overexpression affected the actin cytoskeleton in PtK₂ cells but detected no changes in filamentous actin upon transfection (our unpublished results; Burkhardt et al., 1997). The overexpression of dynamitin can affect the organization of MTs in fibroblastic cells (Burkhardt et al., 1997; Quintyne et al., 1999). Because nuclear targeting of capsids is MT dependent (Sodeik et al., 1997), we analyzed the MT-network in PtK₂ cells by immunofluorescence microscopy. In many cells, the MTs emanated to the peripheral cytoplasm from one location in the perinuclear region that most likely represents the position of the MTOC or centrosome (arrows in Figure 4). However, there were also cells in which there were several MT organizing zones around the nucleus rather than a single, well-defined MTOC or in which MTs emanated more broadly from the nuclear surface (asterisks in Figure 4). The number of cells without an apparent MTOC increased after overexpression of dynamitin (a) or dynamitin-GFP (b), whereas GFP (c) did not have such an effect. However, also in untransfected cells without an obvious MTOC and unfocussed MTs many capsids reached the nucleus (Figure 4d, asterisks), suggesting that MTs but not focused MTs are required for nuclear targeting of HSV1 capsids.

View larger version (90K): [in this window] [in a new window]   Figure 4.   Organization of the microtubule network in transfected PtK2 cells. Immunofluorescence microscopy showing the MT network of PtK2 cells after overexpression of dynamitin or dynamitin-GFP. In almost all cells the MTs emanated from the perinuclear region of the cell. After overexpression of dynamitin (a) or dynamitin-GFP (b), there seemed to be fewer cells with a well-defined MTOC. In untransfected cells (b and d) and in cells overexpressing GFP (c), the MTs were focused at the MTOC in many cells. In untransfected cells with an apparent MTOC (arrows in d) and in cells with unfocussed MTs (asterisks in d) numerous capsids reached the nucleus (d, capsids). Twenty-eight hours after transfection, cells were fixed with PHEMO-fix and double-labeled with anti-myc (pAb, a; top panel) and anti-tubulin (mAb 1A2, a; bottom panel). Cells transfected with GFP proteins were only labeled with anti-tubulin (mAb 1A2, b and c; bottom panels). Note that because of the PHEMO-fixation, GFP was relocalized to the nucleus. In unfixed cells, most of the transiently expressed GFP was localized to the cytoplasm. In d, untransfected PtK2 cells were infected with HSV1 for 3 h and fixed with PHEMO-fix at 37°C. Capsids were labeled with anti-LC (d; right panel) and MTs with 1A2 (d; left panel).

Virus Binding and Internalization after Transfection

Overexpression of dynamitin affects both, secretory and endocytic membrane traffic (Burkhardt et al., 1997; Presley et al., 1997; Valetti et al., 1999), and the concentration of HSV1 cell surface receptors present at the plasma membrane might therefore have been changed by the transfected proteins. Because several different molecules can serve as HSV receptors (Spear et al., 2000), we decided to measure virus binding and internalization directly rather than trying to determine the subcellular localization of all potential viral surface receptors. To this end, cells overexpressing dynamitin, dynamitin-GFP, or GFP were infected with HSV1 for 15 min, fixed, and then labeled for the overexpressed proteins and viral glycoproteins. Compared with untransfected cells and with cells overexpressing GFP, glycoprotein labeling was unchanged by overexpression of dynamitin or dynamitin-GFP (Figure 5). Thus, the sum of surface-bound and internalized virus was similar under all conditions tested. Next we determined virus binding by measuring the amount of cell-bound virus and internalization using a protease protection assay (Sodeik et al., 1997). After 2 h binding on ice ~50% of ³H-thymidine-labeled HSV1 resist washing with buffer, and 95% of the bound virus can be detached from cells by proteinase treatment. If, however, the cells are warmed up, 70% of the bound HSV1 enters the cells with a half time of ~8 min (Sodeik et al., 1997; our unpublished results).

View larger version (75K): [in this window] [in a new window]   Figure 5.   HSV1 binding and internalization in transfected cells. In cells overexpressing dynamitin (a) or dynamitin-GFP (b) virus binding was not reduced compared with untransfected cells or cells overexpressing GFP (c). Immunofluorescence microscopy of PtK2 cells transfected with dynamitin or dynamitin-GFP and 20 h later inoculated with HSV1 for 2 h at 4°C. After 2 h at 4°C cells were washed and shifted to 37°C for 15 min to reduce unspecific binding, fixed in PFA and permeabilized with TX-100. They were double-labeled with anti-dynamitin (a; top panel) and anti-gB (a; bottom panel), and cells transfected with GFP proteins were only labeled with anti-gD (mAb DL6, b and c; bottom panels).

Using these assays we determined that compared with untransfected or GFP-transfected PtK₂ cells, neither binding (Figure 6A) nor internalization (Figure 6B) of ³H-thymidine-labeled HSV1 were reduced after overexpression of dynamitin-GFP. There seemed to be a slight increase of virus internalization in dynamitin-GFP- and GFP-transfected cells compared with control cells (Figure 6B). The reasons for that are unclear. However, the reduced viral protein synthesis after overexpression of dynamitin or dynamitin-GFP could neither result from reduced virus binding nor from reduced internalization.

View larger version (10K): [in this window] [in a new window]   Figure 6.   Quantification of HSV1 virus binding and internalization in transfected cells. (A) Virus binding after overexpression of dynamitin-GFP or GFP was identical to virus binding to mock-transfected cells (n = 2). Two-sided Student's t test confirmed that viral binding was not significantly different in dynamitin-GFP-transfected cells compared with either GFP-transfected cells (p = 4.21 × 101) or to untransfected control cells (p = 4.03 × 101). There was also no significant difference in viral binding between control and GFP-transfected cells (p = 9.42 × 101). Virus binding was calculated as radioactivity per cell and expressed as percent of radioactivity per cell compared with GFP-transfected cells. (B) Virus internalization after overexpression of dynamitin-GFP was not reduced compared with GFP-transfected cells (n = 6). Two-sided Student's t test confirmed that viral internalization was not significantly different in dynamitin-GFP-transfected cells compared with GFP-transfected cells (p = 1.45 × 101). However, compared with untransfected cells, virus internalization was higher in cells transfected with GFP (p = 2.46 × 103) or dynamitin-GFP (p = 7.11 × 103). Virus internalization was calculated as radioactivity per cell and expressed as percent of radioactivity per cell compared with GFP-transfected cells.

Dynamitin Reduces Cytosolic Viral Capsid Transport to the Nucleus

The reduced amount of viral protein synthesis after overexpression of dynamitin could be due to impaired cytosolic transport of incoming capsids from the periphery to the nucleus. We therefore transiently transfected PtK₂ cells with dynamitin or dynamitin-GFP and infected them with HSV1 in the presence of cycloheximide.

Most capsids had reached the nucleus in untransfected cells as well as in cells overexpressing GFP 3 h after infection (Figure 7). In cells overexpressing dynamitin or dynamitin-GFP, fewer capsids were present at the nucleus. Interestingly, in many cells overexpressing dynamitin or dynamitin-GFP the capsids were not randomly distributed over the entire cytoplasm as they are early in the infection (Sodeik et al., 1997), but rather the capsids had accumulated in the cell margins (arrows in Figure 7, a and b). Similar results were obtained using BHK and Vero cells. However, BHK and Vero cells sometimes showed dramatic changes in cell morphology, whereas the morphology of PtK₂ cells was largely not affected by overexpressed dynamitin.

View larger version (59K): [in this window] [in a new window]   Figure 7.   Overexpression of dynamitin reduces HSV1 capsid transport to the nucleus. Three hours after infection most capsids were bound to the nucleus in untransfected cells (a-c) and in cells overexpressing GFP (c). In cells overexpressing dynamitin (a) or dynamitin-GFP (b), only a few capsids reached the nucleus, and several capsids were seen in the periphery of the cells (arrows in a and b, bottom panels). Immunofluorescence microscopy of PtK2 cells transfected with dynamitin or dynamitin-GFP and 46 h later infected with HSV1 for 3 h in the presence of cycloheximide. Cells were fixed with PFA, permeabilized with TX-100, and double-labeled with anti-dynamitin (mAb p50; a, top panel) and anti-HC (pAb; a, bottom panel). Cells transfected with GFP proteins were labeled with anti-HC (pAb; b and c, bottom panels).

Occasionally, there seemed to be fewer capsids visible in cells overexpressing dynamitin or dynamitin-GFP than in control cells. This was due to the fact that in control cells most capsids were concentrated at the nuclear envelope and were thus visualized in one focus plane. In contrast in cells overexpressing dynamitin or dynamitin-GFP, capsids were distributed throughout the entire cytoplasm and therefore were not visualized in one focus plane. For quantification, PtK₂ cells were randomly selected and grouped into three different classes (cf. Figure 7b): 1) cells with many capsids at the nuclear envelope typically forming a nuclear crescent, 2) cells with a reduced amount of nuclear capsids, and 3) cells with very few capsids at the nucleus. The overexpression of dynamitin or dynamitin-GFP clearly reduced the number of cells that showed many capsids at the nucleus by ~50% compared with control or GFP-transfected cells (Figure 8A). Similar results were obtained using Vero cells overexpressing dynamitin (our unpublished results). As in other systems (Burkhardt et al., 1997; Suomalainen et al., 1999), we noticed that the degree of inhibition on capsid transport was dependent on the dose of the transfected protein. We therefore estimated expression levels using a digital camera and specifically analyzed cells expressing high amounts of the transfected proteins (Figure 8B). In these cells, dynamitin-GFP and dynamitin inhibited viral capsid transport by 85% compared with GFP-transfected or control cells.

View larger version (16K): [in this window] [in a new window]   Figure 8.   Quantification of viral capsid transport in transfected cells. For quantification, untransfected and transfected cells were randomly selected and classified into three different groups: cells with many capsids at the nucleus, typically forming a nuclear rim (Figure 7b, 1), cells with a reduced amount of capsids at the nucleus (Figure 7b, 2), and cells with <10 capsids at the nucleus (Figure 7b, 3). Either all (A) or only high expressing (B) cells were analyzed. Expression levels in the cytoplasm were determined using a digital camera. One can directly compare the expression levels for GFP with dynamitin-GFP (cf. MATERIALS AND METHODS). Dynamitin and dynamitin-GFP strongly inhibited HSV1 entry, whereas GFP had no effect on viral infection. Forty-five or 46 h after transfection with dynamitin, dynamitin-GFP, or GFP, PtK2 cells were infected with HSV1 for 3 or 2 h (A and B, respectively), in the presence of cycloheximide. Cells were fixed with PFA, permeabilized with TX-100, and double-labeled with anti-capsid (anti-HC or anti-LC) and anti-dynamitin (cf. Figure 8). Cells with aggregated protein or abnormal morphology were excluded from analysis.

At early time points, incoming capsids are distributed randomly over the entire cytoplasm (Sodeik et al., 1997). However, in the cells in which the function of dynein was inhibited by the overexpression of dynamitin, the capsids were often concentrated in peripheral parts of the cells (Figure 9a). If dynamitin-transfected cells were infected in the presence of nocodazole, which disrupts the MT network, the capsids were also distributed randomly over the entire cytoplasm (Figure 9b). This experiment suggested that without the minus-end-directed MT motor dynein, the capsids might be transported by a plus-end-directed MT motor to the cell periphery.

View larger version (46K): [in this window] [in a new window]   Figure 9.   The peripheral accumulation in dynamitin expressing cells requires an intact MT network. In cells overexpressing dynamitin-GFP (a) or dynamitin, few incoming capsids reach the nucleus, and many accumulate in the peripheral cytoplasm 3 h after infection. If such cells are infected in the absence of MTs, incoming capsids are distributed randomly over the entire cytoplasm, and there is no peripheral accumulation (b). Immunofluorescence microscopy of PtK2 cells transfected with dynamitin-GFP or GFP and 46 h later infected with HSV1 for 3 h in the presence of cycloheximide and in the absence (a) or presence (b and c) of nocodazole (ND). Cells were fixed with PFA, permeabilized with TX-100, and labeled with anti-VP5 (mAb 5C10; bottom panels).

In summary our data show that overexpressing dynamitin reduced the number of cytosolic viral capsids transported to the nucleus. Because under this condition fewer viral genomes reached the nucleoplasm, immediate-early viral gene expression was also reduced. The inhibition of virus entry by high amounts of dynamitin was not due to impaired virus binding, virus internalization, or the slight changes in the MT network but to a block of dynein function in cytosolic HSV1 capsid transport along MTs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Incoming cytosolic HSV1 capsids are efficiently targeted along MTs from the plasma membrane to the nucleus (Hammonds et al., 1996; Sodeik et al., 1997; Mabit, Nakano, Prank, Saam, Döhner, Sodeik, Greber, unpublished data). The typical organization and polarity of MTs in cultured cells and axons require the use of a minus-end-directed MT motor for transport to the nucleus (Kristensson et al., 1986; Topp et al., 1994; Sodeik et al., 1997; Bearer et al., 2000). Cytoplasmic dynein is responsible for most minus-end-directed MT transport during interphase and thus a prime candidate for viral trafficking (Sodeik, 2000; Ploubidou and Way, 2001).

The dynein subunits DHC, DIC, and p150^Glued, a subunit of the dynein cofactor dynactin, colocalized with incoming HSV1 (Sodeik et al., 1997; Figure 1). Moreover, EHNA (erythro-9-3-[2-hydroxynonyl]adenine), an inhibitor of axonemal and cytoplasmic dynein, blocks HSV1 infection in neuronal cells (Kristensson et al., 1986). However, EHNA also affects adenosine deaminase, cGMP-stimulated phosphodiesterase, and the actin cytoskeleton (Penningroth, 1986; Mery et al., 1995), and it is unclear which phase of the viral life cycle was inhibited. We therefore asked whether incoming HSV1 capsids use the host factor dynein for riding along MTs to the nucleus.

Inhibition of Dynein and Dynactin Does not Affect HSV1 Internalization

The function of dynactin and dynein can be inhibited by transiently overexpressing one dynactin subunit, the 50-kDa protein dynamitin (Echeverri et al., 1996). Because endocytic and exocytic membrane traffic involve dynein-mediated vesicular transport steps (Burkhardt et al., 1997; Presley et al., 1997; Valetti et al., 1999), the subcellular localization of viral receptors might have been changed upon such treatment. However, none of the transfected proteins we tested reduced the efficiency of HSV1 binding or particle uptake. Thus, any influence on virus infection was not simply due to reduced surface expression of viral receptors and less efficient fusion of virions with the cells.

Inhibition of Dynein and Dynactin Blocks Cytosolic HSV1 Capsid Transport along MTs

Although virus entered transfected cells normally, overexpressed dynamitin inhibited the transport of cytosolic HSV1 capsids from the cell periphery to the nucleus by ~50% and in cells expressing high levels of dynamitin even by 85%. As a consequence viral immediate-early gene expression was reduced, too.

Two factors may have prevented a complete inhibition of virus infection in our set up. First, about a third of all cells were not transfected, and immunofluorescence microscopy confirmed that those were infected normally. Second, single cell analysis strongly suggested that the degree of capsid transport inhibition correlated with the level of dynamitin expression (Figure 8). This is consistent with biochemical experiments, which show that excess dynamitin blasts dynactin into two subcomplexes of 9 and 18 S, presumably by saturating dynamitin binding sites on them, thus destroying the architecture of dynactin (Echeverri et al., 1996; Eckley et al., 1999). Because this is likely to be a dose-dependent effect, it was expected that dynactin-dependent transport and virus infection were only completely inhibited in cells expressing high levels of dynamitin.

As in the absence of MTs (Sodeik et al., 1997; Mabit, Nakano, Prank, Saam, Döhner, Sodeik, Greber, unpublished data), we detected a few capsids at the nuclear membrane in cells with high concentrations of dynamitin. This was no surprise, because virions can bind to the "apical" plasma membrane just on top of the nucleus. Capsids derived from these virions most likely reach the nuclear pores without MTs or dynein and dynactin. Dynein and dynactin-mediated MT transport is therefore not essential for infecting nonpolarized cells in culture. However, it is likely to be required during pathogenesis when HSV1 infects highly polarized epithelial and elongated neuronal cells (Enquist et al., 1998), but also in less polarized cells, as described here, dynactin and the molecular motor dynein transport HSV1 capsids efficiently along MTs.

Functions of Dynein and Dynactin during HSV1 Capsid Transport?

ATP hydrolysis induces conformational changes in the DHC head domain, which produce a power stroke toward the MT minus end, whereas the smaller subunits DIC, DLIC, and DLC are attached to the stem domain (Habura et al., 1999; Tynan et al., 2000a), which is involved in cargo binding (Vaughan and Vallee, 1995; King, 2000; Tynan et al., 2000b). Dynactin is needed for dynein-mediated vesicle transport (Gill et al., 1991; Schroer and Sheetz, 1991) and transport of nonmembranous cargo such as NuMA aggregates, aggresomes, adenovirus capsids, neurofilaments, chromosomes, and pericentrin particles (Merdes and Cleveland, 1997; Garcia-Mata et al., 1999; Suomalainen et al., 1999; Shah et al., 2000; Sharp et al., 2000; Young et al., 2000).

Earlier electron microscopy data suggested that dynein binds with its stem domain to HSV1 capsids (Sodeik et al., 1997). Our immunofluorescence microscopy data are consistent with either an indirect interaction via dynactin or a direct binding of HSV1 capsids to dynein. Dynactin might serve as an initial or permanent anchor for dynein on the capsid (Echeverri et al., 1996) or tether between MT and capsid while dynein is detaching from the MT to make its next step (King and Schroer, 2000). The ATPase activity of dynein has also been reported to be regulated by dynactin-dependent phosphorylation (Kumar et al., 2000).

Bidirectional HSV1 Capsid Transport along Microtubules?

Many subcellular structures and progeny GFP-tagged alphaherpes viruses (Smith et al., 2001; Willard, 2002) can move bidirectionally along MTs. Interestingly, overexpression of dynamitin did not lead to a random capsid distribution but to their MT-mediated accumulation in the cell margins (Figures 7 and 9). This suggested that besides dynein, HSV1 capsids might also use a plus-end-directed MT motor. Plus-end-directed capsid motility could be involved in further transport from the MTOC to the nucleus, as would be required in cells where the MTOC is not directly neighboring the nucleus. MT-mediated, plus-end-directed capsid transport could also be involved in apical entry of polarized epithelial cells (Topp et al., 1996). Moreover, during HSV1 egress from neurons, capsids are transported anterogradely to the presynapse (Miranda-Saksena et al., 2000; Ohara et al., 2000). Because of the uniform polarity of MTs in axons, this transport has to be catalyzed by a plus-end-directed motor.

If capsids were indeed able to travel along MTs in both directions, to the minus and plus ends, specific signals must regulate which motor the capsid is supposed to use during the different steps of the viral life cycle. Thus, the direction of capsid motility must be tightly controlled. The main transport direction during virus entry must be to MT minus-ends to ensure net movement to the cell center and the nucleus. However, if minus-end-directed, dynein-mediated transport was inhibited by overexpressing dynamitin, these putative plus-end-directed motors might have taken over and transported capsids to the cell margins.

Dynein or Kinesin Receptors Encoded by HSV1

In contrast to other cargo transported along MTs, the protein composition of HSV1 is known. There are ~20 tegument and capsid proteins that could function in motor or dynactin binding. The HSV1 gene product of UL34 interacts with DIC in GST pull down assays (Ye et al., 2000). However, because UL34 has properties of a type-II membrane protein and in pseudorabies virus is not present in purified virions (Klupp et al., 2000; Reynolds et al., 2001), it remains to be seen how its interaction with DIC subunit could participate in viral capsid transport. The HSV1 tegument protein US11 was shown recently to interact with the heavy chain of conventional kinesin (Diefenbach et al., 2002). Additional candidates for motor receptors include VP22 (UL49) and UL25, that both, upon transient transfection, seem to localize to MTs (Elliott and O'Hare, 1998; Kaelin et al., 2000). In contrast to VP22 that dissociates from the capsid upon virus entry, UL25 and also VP1-3 (UL36) remain on the capsid until it reaches the nucleus (Morrison et al., 1998; Kaelin et al., 2000; Sodeik, Szmak and Prank, unpublished results).

Reconstitution of MT capsid motility in vitro (Wolfstein, Döhner, Allan and Sodeik, unpublished results) and HSV1 mutants can now be used to identify structural viral proteins required for capsid targeting to the nucleus.