Ultrastructural Visualization of Individual Tegument Protein Dissociation during Entry of Herpes Simplex Virus 1 into Human and Rat Dorsal Root Ganglion Neurons

Cells and viruses.

Three recombinant viruses were used in parallel in this study: vUL37-GFP (green fluorescent protein)-labeled HSV-1 (strain 17), mRFP1-VP26/pUL36-GFP (HSV F-GS2945) (strain F), and mRFP1-VP26/GFP-pUL37 (HSV F-GS3245) (strain F). These viruses have been described previously (3, 4, 48, 61). vUL37-GFP was kindly provided by Frazer Rixon (MRC Virology Unit, Institute of Virology, United Kingdom). Greg Smith (Northwestern University Medical School) kindly provided HSV F-GS2945 and HSV F-GS3245.

Preparation of virus stocks.

Viruses were grown and passaged in Vero cells. Briefly, Vero cells in 150-cm² Falcon tissue culture flasks were infected at a multiplicity of infection (MOI) of 0.01 PFU per cell. Infected cells were frozen and thawed three times and subjected to sonication in a Branson sonicator (3 times for 20 s, 100% duty cycle). The cellular debris was removed by centrifugation at 15,800 × g for 10 min at 4°C in a Sorvall RC26 Plus ultracentrifuge. Virus was then pelleted by centrifugation through a 10% sucrose cushion at 64,000 × g for 2 h at 4°C in a Beckman Coulter Optima XL-100K ultracentrifuge. The pellet was resuspended in a small volume of Dulbecco's modified Eagle medium (DMEM) (Invitrogen) supplemented with 1% fetal calf serum (FCS) (vol/vol). Viral titers were measured by plaque assays performed on Vero cells.

Antibodies for TIEM.

Antibodies were kindly provided by the following investigators: rabbit antibody against VP5 (NC1) from Gary Cohen and Roselyn Eisenberg, University of Pennsylvania (8); rabbit antibody against VP22 from Peter O'Hare, Marie Curie Research Institute, Oxted, United Kingdom (15); mouse antibody against VP16 (LP1) from Tony Minson, University of Cambridge, United Kingdom (44); rabbit antibody against pUL37 (780 antiserum) from Frank Jenkins, Uniformed Services University of the Health Sciences, MD (62); and rabbit antibodies to pUL36 and pUL37 (35) from Thomas Mettenleiter, Freidrich Loeffler Institute, Insel Riems, Germany. The gold-conjugated antibodies were purchased from British Biocell International, United Kingdom.

Preparation of human fetal DRG explants.

DRG were prepared as described previously (61) from human fetal tissue obtained at therapeutic termination following the informed consent of the mother. Sydney West Area Health Services and the University of Sydney Human Research Ethics Committee approved these protocols. The DRG were dissected, cleansed of connective tissue, placed onto Matrigel (BD Biosciences)-coated coverslips, and cultured at 37°C with 5% CO₂ in neurobasal medium supplemented with 4 mM l-glutamine (Invitrogen), 2% B-27 (Invitrogen), and 7S nerve growth factor (100 ng/ml) (Sigma) for 5 to 7 days for axon outgrowth prior to HSV-1 infection.

Preparation of dissociated rat neuronal cultures.

DRG neurons were prepared from 4-day-old Wistar rat neonates as previously described (49). Briefly, DRG were dissociated in Hanks calcium- and magnesium-free solution (Invitrogen) plus 0.25% trypsin (Sigma) and 0.05% collagenase (Worthington Biomedical Co.) for 30 min at 37°C, followed by DNase (10 mg/ml) (Sigma) for 5 min at 37°C, washed twice by centrifugation at 80 × g, and passed through 35% Percoll (Sigma). The cell pellet was resuspended in 500 μl of neurobasal medium (Invitrogen), plated on Matrigel-coated plastic coverslips, and cultured at 37°C with 5% CO₂ in neurobasal medium supplemented with 4 mM l-glutamine (Invitrogen), 2% B-27 (Invitrogen), and 7S nerve growth factor (100 ng/ml) (Sigma) for 3 days prior to HSV-1 infection. Sydney West Area Health Services and the University of Sydney Animal Research Ethics Committees approved the use of rat neonates.

HSV-1 infection of DRG cultures.

For all time points, neuronal cultures in 24-well plates were infected with recombinant viruses at 2.0 × 10⁵ PFU/coverslip. The cells were incubated at 4°C for 30 min to promote virus attachment. The cultures were then returned to 37°C with 5% CO₂ and incubated for 30 min postinfection (mpi) or for 2, 4, or 24 h postinfection (hpi). The shift in temperature from 4°C to 37°C allowed the bound virus particles to enter the cells and give a synchronous infection. For incubation times longer than 2 h, the virus inoculum was removed at 2 hpi and the cells were washed with neurobasal medium. The neurons were always incubated at 37°C with 5% CO₂. Mock-infected cultures were incubated in the same culture medium and fixed at the same time points. HSV F-GS2945 and HSV F-GS3245 were used for fluorescence microscopy studies, and vGFP-UL37 was used for TIEM studies.

Fluorescence microscopy and image analysis.

DRG cultures were processed for fluorescence studies as previously described (47). Briefly, neuronal cultures on coverslips were fixed in 3% formaldehyde for 30 min at room temperature. This was followed by six washes in phosphate-buffered saline (PBS) before the coverslips were mounted in Prolong Gold with 4′,6′-diamidino-2-phenylindole (DAPI) on glass slides. The slides were examined using a Deltavision Core image restoration system. Images were acquired using a Photometrics CoolSnap QE camera with sequential exposures to monomeric red fluorescent protein (mRFP) (600 ms) and GFP (1 s) and were deconvolved using Sedat-Agard algorithms (24) available through the Deltavision SoftWoRx software, version 3.0.0. Images were acquired as z-stack series and represented as volume projections (minimum number of stacks was 30). Background was subtracted after capture using SoftWoRx, through a single adjustment of the levels histogram. The images were then cropped using Adobe Photoshop CS5.

The proportion of mRFP1 capsids colocalizing with GFP during entry at 2 and 4 hpi were determined by manual counting of the total number of mRFP particles in the cytosol and at the nuclear rim and then counting the number of these particles colocalizing with GFP. The results were expressed as percentage of capsids that colocalized with GFP. For these counts, an average of 10 cells were used for each virus and time point. Both original (raw) and deconvolved images were compared to ensure no artifact (loss or addition of mRFP1 or GFP fluorescent puncta) during the counting process.

Imaging of extracellular virus particles.

To characterize the extracellular virus particles of fluorescence-tagged viruses (GS2945 and GS3245), supernatants from Vero cells infected with either virus were added to glass coverslips coated with Cell Tak adhesive (BD Pharmingen). The coverslips were spun at 1,200 × g for 30 min at 4°C in a benchtop centrifuge. The coverslips were rinsed once with PBS and fixed in 3% formaldehyde for 20 min at room temperature. After two rinses in PBS, the coverslips were mounted with Prolong gold and imaged using deconvolution microscopy as described above. The proportion of mRFP capsids emitting GFP fluorescence were manually counted, and results were expressed as averages from five separate counts.

TIEM.

DRG cultures were processed by modified freeze substitution as previously described (49, 61). Briefly, coverslips with DRG cultures in situ were fixed in 4% formaldehyde and 0.1% glutaraldehyde for 1 h and then washed in PBS. Coverslips were then trimmed, dipped in 10% gelatin, and placed in cryoprotectant (2.3 M sucrose) overnight. Preparations were then frozen by rapid plunging into liquid nitrogen followed by transfer to dry methanol at −90°C, freeze substituted from methanol, embedded in Lowicryl HM20, and polymerized with UV light at −45°C for 48 h in a Reichert AFS freeze substitution system (Leica Microsystems, Austria) as previously described (48).

Immunolabeling.

Tissue sections were collected on Formvar- and Pioloform-coated gilded nickel grids and immunolabeled as described previously (48). Briefly, tissue sections were incubated with 50 mM glycine for 15 min and in blocking buffer with 10% normal serum and acetylated bovine serum albumin (BSA) for 30 min. Primary antibodies were incubated overnight at 4°C, followed by incubation with secondary antibodies conjugated to 5- to 10-nm gold particles. Steps following primary antibody incubation, including washing steps and secondary-antibody incubation, were performed using a Leica EM IGL Immunostainer (Leica Microsystems, Austria). After immunolabeling, the sections were stained using 1% uranyl acetate (in 50% ethanol) followed by Reynolds lead citrate and examined with either a Philips CM10 or CM120 BioTWIN transmission electron microscope at 80 kV or 100 kV, respectively. Images were recorded on Kodak 4489 electron microscope film or collected using a SIS Morada digital camera.

Tegument composition of viral capsids after entry into neurons.

TIEM was used to examine the tegument composition of incoming HSV-1 virions in cultured human fetal explant and rat DRG neurons. Axons in DRG explant cultures are approximately 4 to 5 mm in length (61) and hence are optimal for visualizing viral particles during entry and retrograde transport in axons. As rat neonates are more readily available, cultures of dissociated rat DRG neurons, which yielded pure neuronal cultures, were chosen to examine virus particles in the neuronal cell body during virus entry.

Both rat and human DRG cultures were infected with vUL37-GFP, and the cultures were fixed at 30 mpi and 2, 4, and 24 hpi. Coverslips with DRG cultures in situ were fixed and processed for electron microscopy. Serial ultrathin sections were cut parallel to the growth plane (longitudinally) in order to examine proximal, middle, and distal regions of axons in situ and to ultrastructurally examine neuronal cell bodies from the cell surface to the nucleus without disrupting the arrangement of cells and axons.

Immunogold labeling (5- or 10-nm gold particles) with antibodies to the tegument proteins VP16, VP22, pUL36, and pUL37 and capsid VP5 was performed to examine the presence of these proteins on viral particles in the axons and cell body during the initial 4 hpi. The presence and intensity of immunolabeling of viral particles were classified semiquantitatively: the presence of 3 or fewer gold particles was considered moderate to weak labeling, and the presence of 4 or more gold particles was considered strong labeling. Using extracellular virions as controls, gold label was considered to be “on” a virus particle if it was located directly on the surface of the viral particle or within two gold particles of the target membrane (in the case of enveloped capsids or extracellular particles) or the edge of cytosolic capsids (with no envelope).

Extracellular virions and cytoplasmic enveloped virions were identified both on the basis of morphology, including the presence of an electron dense core or capsid, a well-developed tegument layer, and a distinct membrane representing viral envelope, and on the basis of size (170 to 220 nm in diameter) (20). Unenveloped capsids were distinguished from axonal vesicles on the basis of their size (approximately 125 nm in diameter) (74), thicker structure of the viral capsid compared to vesicle walls, and electron-dense DNA cores (26, 27, 48). In addition, viral capsids were further identified by immunolabeling with antibodies to the major capsid protein VP5. Electron micrographs presented here in support of the findings are representative of multiple observations.

Detection of unenveloped capsids in axons and in the cytoplasm of the cell body during virus entry.

Extracellular virions bound to the plasma membrane of axons and neuronal cell bodies, as well as unenveloped capsids within axons and cell body, were observed at all time points up to 4 hpi (see Fig. 1 to 5). However, the number of extracellular virions decreased progressively with time, and a reduction of almost 90% in the number of extracellular virions was seen from 30 mpi to 4 hpi (Tables 1 and 2), indicating efficient uptake of the virus by 4 hpi. While extracellular virions were readily observed, capsids within axons and cell bodies were considerably harder to find. Hence, few such capsids could be examined and counted. An average of 2,500 axonal processes and 200 cell body profiles were examined for each time point and each label (Tables 1 and 2). Unenveloped capsids could be detected within axons as early as 30 mpi and could also be seen in the cytosol of infected neurons at the same time point. By 4 hpi, capsids could be seen at the nuclear membrane and in the cytosol of the cell body. These unenveloped capsids were similar in morphology to those previously reported in axons of HSV-1-infected neurons (27, 47–49, 61) and were not present in uninfected cultures (data not shown). Furthermore, no enveloped capsids within vesicles or partially enveloped capsids were observed at any time between 30 mpi and 4 hpi (Tables 1 and 2).

Fully enveloped and partially enveloped capsids enclosed within vesicles were, however, observed in varicosities and growth cones at 24 hpi, in accordance with our previous published findings (48, 61).

In addition to morphology, enveloped and unenveloped viral capsids in axons and neuronal cell bodies were identified by immunolabeling with antibodies to capsid protein VP5.

At early time points postinfection, strong label for capsid VP5 was present on extracellular virions that were either free or bound to the plasma membrane of axons and neuronal cell bodies (Fig. 1A). Label for VP5 could also be seen on unenveloped capsids within axons as previously published (47) and on unenveloped capsids in the cytosol at all time points postinfection, including 24 hpi (Fig. 1B and C). The density of label for VP5 (4 or 5 gold particles) on unenveloped capsids remained consistent and was similar to that on extracellular virions at early and late time points of infection (Fig. 1). Labeling for VP5 could not be performed for all experiments in order to allow labeling for other proteins. However, capsids could be readily identified by their size and morphology.

Dissociation of outer tegument VP16 from viral particles during entry.

At all time points examined during infection, about 75% of extracellular viral particles observed carried strong label for VP16 (Fig. 2A) (Table 1). These particles were usually found adjacent to cell bodies of infected neurons as well as to axonal processes. At 30 mpi, unenveloped capsids (75%; 3/4) (Table 1) carrying weak or no label for VP16 (Fig. 2B and C) were observed adjacent to the inner aspect of the axonal membrane. Label for VP16 was also present along the plasma membrane and in the cytosol of axons, in close proximity to the incoming capsids (Fig. 2B and C).

At 2 and 4 hpi, no label for VP16 could be seen on unenveloped capsids within axons. Similarly, no label for VP16 could be seen on capsids in the cytosol of infected neurons at these time points (4/4 and 7/7, respectively) (Table 1; Fig. 2D). Free label (not associated with capsid) for VP16 was, however, seen throughout the cytosols and nuclei of neuronal cell bodies (Fig. 2D). This labeling for VP16 was specific to infected neurons and was not detected in uninfected controls (data not shown).

Infected cultures fixed at 24 hpi served as a positive control for our experiments. At 24 hpi, strong label for VP16 was seen on extracellular virions (83.3%; 40/48) (Table 1; Fig. 2E) and unenveloped capsids (80%; 20/25) (Table 1; Fig. 2F) in the cytosol of the neuronal cell body. A direct comparison of VP16 label on unenveloped capsids during entry and egress revealed that while most capsids had 4 or more gold particles at 24 hpi (egress), only weak to moderate label (2 or 3 gold particles) for VP16 was on the capsids at 30 mpi (Table 1). By 4 hpi, capsids were devoid of any label for VP16 suggesting that there was then complete dissociation of the protein from incoming capsids (Table 1).

Dissociation of outer tegument protein VP22 from viral particles during entry.

As for VP16, strong label for VP22 was observed on almost 80% of extracellular virions at all time points postinfection (Table 1; Fig. 3A) and also on viral particles bound to the plasma membrane (Fig. 3B). At 30 mpi, weak or no label for VP22 was seen on unenveloped capsids (Table 1) lying close to the inner aspect of the axonal membrane (Fig. 3C). Label for VP22 was frequently detected beneath the axonal membrane in close proximity to incoming capsids (data not shown). This most likely represented residual dissociated protein from the incoming capsid. At 2 and 4 hpi, almost all the unenveloped capsids observed inside the cytosol of the neuronal cell body carried little or no label for VP22 (Table 1; Fig. 3D), suggesting that, like VP16, VP22 dissociated completely during transport of viral capsids toward the nucleus.

At 24 hpi, label for VP22 was present on extracellular virions lying outside the cell membrane (80%; 40/50) (Table 1; Fig. 3E). In addition, strong specific label for VP22 was observed on unenveloped (33%; 10/30) and enveloped (80%; 24/30) capsids (Table 1) in the cytosol (Fig. 3F). As with VP16, there was a marked difference in the density of VP22 label between unenveloped capsids at 24 hpi and those at 30 mpi and at 2 and 4 hpi, suggesting that there was a gradual but complete loss of VP22 from the incoming capsids by 4 hpi.

Presence of inner tegument protein pUL36 on viral particles during entry.

In addition to VP16 and VP22, the presence of inner tegument proteins pUL36 and pUL37 on virus particles was also examined in axons and cell bodies of DRG neurons. Moderate label for pUL36 was seen on extracellular virions at all time points postinfection (Fig. 4A). The density of label for pUL36 on extracellular virions was consistently less than that for outer tegument proteins VP16 and VP22 (Fig. 2 and 3; Table 2). Both VP16 and VP22 are relatively abundant, being present at 1,000 to 2,000 copies per virion (22). In contrast, fewer than 500 copies of pUL36 and pUL37 are present in assembled virions (45, 62).

Unenveloped capsids carrying weak to moderate label for pUL36 were seen within axonal processes at 30 mpi (55%; 5/9) (Table 2). By 2 and 4 hpi, unenveloped capsids, carrying weak to moderate residual label for pUL36, were detected in the cytosol of the infected neurons (4/8 and 5/8, respectively) (Table 2; Fig. 4C and D). The density of label on these unenveloped capsids was consistently less than that observed on extracellular adherent virions, suggesting a partial loss of the protein during virus entry. Residual dissociated label for pUL36 from incoming capsids was observed close to the plasma membrane (Fig. 4B) and in the cytosol of infected neurons. This label was specific only to infected neurons and was not detected in uninfected neurons.

At 24 hpi, extracellular virions (Fig. 4E) and cytoplasmic enveloped capsids carrying moderate to strong label for pUL36 were observed in infected controls (Table 2). In addition, unenveloped capsids, moderately labeled for pUL36 (68%; 40/59), could be observed throughout the neuronal cytosol (Fig. 4F). The density of pUL36 label on these unenveloped capsids was more than that observed on capsids at 4 hpi suggesting, that the reduced labeling for pUL36 at 4 hpi was not because of possible steric hindrance but due to a partial loss of pUL36 from the capsids upon entry.

Presence of inner tegument protein pUL37 on viral particles during entry.

At all time points examined during infection, majority of extracellular virions labeled densely for pUL37 (Table 2; Fig. 5A). At 30 mpi, approximately 57% of unenveloped capsids within axonal processes carried weak to moderate label for pUL37 (Fig. 5B; Table 2). As with pUL36, the density of pUL37 label on these unenveloped capsids was lower than that on extracellular virions, suggesting a partial loss of the protein upon entry (Table 2). At 2 hpi, half of incoming capsids, in close proximity to the plasma membrane, carried weak to moderate label for pUL37, albeit less than that present on extracellular virions (Table 2; Fig. 5C). However, by 4 hpi, most capsids deep within the cytosol had little or no label for pUL37 on them (Table 2; Fig. 5D), suggesting that there was a progressive loss of this protein as the capsids were translocated toward the nucleus. Free label for pUL37 was also seen in the cytosol of infected neurons but not in uninfected cells (data not shown).

At 24 hpi, extracellular virions surrounding the cell body carried strong label for pUL37 (Table 2; Fig. 5E). Unenveloped (Fig. 5F) and enveloped capsids labeled for pUL37 were also readily observed within the cytosol of infected neurons (Table 2). A comparison of pUL37 label on unenveloped capsids during early and late stages of infection revealed that, as with VP16 and VP22, there was a gradual loss of pUL37 from incoming capsids by 4 hpi.

Distribution of viral capsids and inner tegument proteins pUL36 and pUL37 during virus entry into rat DRG neurons.

To further investigate the association and/or dissociation of inner tegument proteins pUL36 and pUL37 from capsids during virus entry, their distribution and colocalization with viral capsids following virus entry into axons and cell body was examined using dually fluorescence-tagged viruses, HSV F-GS2945 (mRFP1-VP26 and pUL36-GFP) and GS3245 (mRFP1-VP26 and GFP-pUL37). Visualization of two different viral proteins was achieved by detection of endogenous mRFP1 (VP26) and GFP (pUL37 or pUL36) using a Deltavision Core image restoration system. Colocalization between capsids and pUL37 or pUL36 was quantified by manual counting of the total number of mRFP particles in the cytosol and at the nuclear rim and then counting the number of these particles colocalizing with GFP. The results were expressed as percentage of capsids that colocalized with GFP. For these counts, an average of 10 cells were used for each virus and time point.

Mock-infected neurons at the same time points postinfection were processed similarly to infected neurons and served as controls. Each experiment was repeated three times in replicate cultures.

Distribution of viral proteins at 30 mpi.

At 30 mpi, all the viral proteins examined, capsid VP26 and tegument proteins pUL36 and pUL37, were present as numerous distinct fluorescent puncta on the periphery of neuronal cell bodies and along axons from proximal to distal ends (Fig. 6 and 7). Colocalization of mRFP1 capsid VP26 with either pUL36-GFP or GFP-pUL37 was seen in puncta on the periphery of the cell body and along axons but most frequently in the proximal axons (Fig. 6A to D and Fig. 7A to D).

The presence of capsid and tegument proteins on the periphery of the cell body, as well as axons, probably represents assembled input virions that had not yet been internalized. However, the virus inoculum could also include defective viral particles that remained bound to the plasma membrane of the axons and neuronal cell body. The viruses used for these experiments were subjected to high-speed centrifugation to remove membranous debris containing viral proteins. The virus preparation was further checked by negative staining and examined by electron microscopy (data not shown).

Distribution of viral proteins at 2 hpi.

At 2 hpi, fluorescent puncta representing mRFP1 capsids were detected in the cytosol of the cell body and at the nuclear rim (Fig. 6E to H and Fig. 7E to L). The majority of mRFP1 capsids (78.9% of a total of 71 capsids) in the cytosol contained pUL36-GFP whereas half of these mRFP1 capsids (55.2% of a total of 67 capsids) in the cytosol contained GFP-pUL37. By 2 hpi, mRFP1 capsids also reached the nuclear rim, and 71.2% (total, 59 capsids) of these contained pUL36-GFP and about half (51.3%; total, 76 capsids) contained GFP-pUL37.

GFP puncta for both pUL36 and pUL37 were also present in the cytosol and free of mRFP1 capsids, suggesting dissociation from incoming capsids.

Distribution of viral proteins at 4 hpi.

By 4 hpi, numerous fluorescent puncta for mRFP1 capsids could be seen in the cytosol and at the nuclear rim (Fig. 6I to L and Fig. 7M to P). In addition, a few discrete fluorescent puncta for mRFP1 capsids also remained on the cell periphery (Fig. 6 and 7). A proportion of fluorescent puncta for mRFP1 capsids (71.3%; total, 237 capsids) similar to that detected in the cytosol at 2 hpi were associated with pUL36-GFP, whereas 38.3% of fluorescent puncta for mRFP1 capsids (total, 162 capsids) in the cytosol at this time remained associated with GFP-pUL37.

The majority of mRFP1 capsids (70.3%; total, 202 capsids) at the nuclear rim contained pUL36-GFP, whereas only 37.2% of these mRFP1 capsids (total, 180 capsids) contained GFP-pUL37. In addition, GFP puncta for both pUL36 and pUL37 were also detected in the cytosol free (independent of mRFP1 capsids) as at 2 hpi. However, the number of GFP puncta for pUL37 markedly decreased at 4 hpi compared to 2 hpi, suggesting possible degradation of free pUL37.

In order to confirm that the lack of colocalization of pUL37 with mRFP1 capsids during travel to the nucleus was due to dissociation and not to artifacts (from fixation and/or deconvolution processing), the virus inoculum was subjected to the fixation and detection processes performed with the infected neuronal cultures. For this, droplets of input virus suspensions were placed directly onto glass coverslips, fixed in 3% formaldehyde, and examined by deconvolution microscopy to determine the proportion of mRFP1 capsids expressing GFP. As shown in Fig. 8, 88.5% of mRFP-VP26 capsids were found to express pUL36-GFP (total, 590 capsids), while 86.2% of mRFP1 capsids expressed GFP-pUL37 (total, 880 capsids), showing no loss of GFP.

These results, obtained with dually fluorescence-tagged HSV-1, are consistent with our observations using TIEM in that there is some progressive loss of tegument pUL37 from capsids as capsids are transported from the cell periphery to the nuclear rim from 2 to 4 hpi.