Functional Carboxy-Terminal Fluorescent Protein Fusion to Pseudorabies Virus Small Capsid Protein VP26

ABSTRACT

Fluorescent protein fusions to herpesvirus capsids have proven to be a valuable method to study virus particle transport in living cells. Fluorescent protein fusions to the amino terminus of small capsid protein VP26 are the most widely used method to visualize pseudorabies virus (PRV) and herpes simplex virus (HSV) particles in living cells. However, these fusion proteins do not incorporate to full occupancy and have modest effects on virus replication and pathogenesis. Recent cryoelectron microscopy studies have revealed that herpesvirus small capsid proteins bind to capsids via their amino terminus, whereas the carboxy terminus is unstructured and therefore may better tolerate fluorescent protein fusions. Here, we describe a new recombinant PRV expressing a carboxy-terminal VP26-mCherry fusion. Compared to previously characterized viruses expressing amino-terminal fusions, this virus expresses more VP26 fusion protein in infected cells and incorporates more VP26 fusion protein into virus particles, and individual virus particles exhibit brighter red fluorescence. We performed single-particle tracking of fluorescent virus particles in primary neurons to measure anterograde and retrograde axonal transport, demonstrating the usefulness of this novel VP26-mCherry fusion for the study of viral intracellular transport.

IMPORTANCE Alphaherpesviruses are among the very few viruses that are adapted to invade the mammalian nervous system. Intracellular transport of virus particles in neurons is important, as this process underlies both mild peripheral nervous system infection and severe spread to the central nervous system. VP26, the small capsid protein of HSV and PRV, was one of the first herpesvirus proteins to be fused to a fluorescent protein. Since then, these capsid-tagged virus mutants have become a powerful tool to visualize and track individual virus particles. Improved capsid tags will facilitate fluorescence microscopy studies of virus particle intracellular transport, as a brighter particle will improve localization accuracy of individual particles and allow for shorter exposure times, reducing phototoxicity and improving the time resolution of particle tracking in live cells.

INTRODUCTION

Fluorescent protein fusions to herpesvirus capsids have proven to be a powerful method to visualize and track individual virus particles through most of the virus replication cycle. A fluorescent protein fusion to the N terminus of small capsid protein VP26 (xFP-VP26) was first described in herpes simplex virus 1 (HSV-1; human herpesvirus 1) more than 18 years ago (1). A similar N-terminal fusion in pseudorabies virus (PRV; suid herpesvirus 1) was first used in 2001 to resolve and track individual particles over long distances in neurons (2). Since then, N-terminal FP-VP26 fusions have become the most common method to visualize alphaherpesvirus particles in living cells.

Mature herpesvirus capsids are composed of the major capsid protein (VP5 in HSV and PRV), triplex proteins, portal protein, capsid vertex-specific components (UL25 and UL17 in HSV and PRV), and the small capsid protein (VP26 in HSV and PRV). Fluorescent protein fusions have been reported only for the more peripheral capsid proteins UL25, UL17, and VP26. Together, UL25 and UL17 decorate the capsid vertices. Five molecules of UL17 and five homodimers of UL25 bind at each of the 11 pentons (not including the unique portal vertex), giving a potential 55 molecules of UL17 and 110 molecules of UL25 per capsid. Fluorescent protein-tagged UL25 appears to incorporate to full or nearly full occupancy (3–5), and tagged UL17 incorporates approximately 30% less than full occupancy (6). The UL25 fusion seems to be tolerated by the virus particularly well, as there is no detectable impact on virus replication or neurovirulence (4); however, with only ∼100 copies of fluorescent protein per capsid, these particles are too dim for many applications, particularly high-resolution live-cell analysis.

In contrast, VP26 binds to the major capsid protein in hexons, with 6 copies on each of the 150 hexons, giving 900 potential binding sites per capsid. N-terminal xFP-VP26 fusions, however, appear to incorporate only 300 to 400 copies per capsid (4). It is, nonetheless, the brightest fluorescent protein capsid tag currently available.

In addition to reduced incorporation, N-terminal xFP-VP26 fusions also have other defects in HSV-1 and PRV. xFP-VP26 fusions are reported to cause a reduction in virus replication in vitro and a reduction in neurovirulence in mice (7). Most reports in the HSV-1 and PRV literature range from no measurable effect on virus replication (1, 4, 8, 9; see also Fig. 2) to approximately one log reduction in virus titer (7, 9, 10), similar to a VP26-null mutation (7). Higher expression levels of xFP-VP26 and fluorescent protein variants with higher dimerization affinity appear to have greater defects in virus replication, possibly because capsid proteins prematurely aggregate in the nucleus (9; reviewed in reference 10). This aggregation and sequestration of capsid proteins may also explain why N-terminal fusions to the small capsid protein of human and mouse cytomegalovirus (human herpesvirus 5 and murid herpesvirus 1) are not only nonfunctional but also dominant negative (11, 12).

FIG 2.

Design and validation of recombinant PRV expressing C-terminal fluorescent protein fusion. (A) Fluorescent protein coding sequences were inserted into the UL35 open reading frame, encoding an in-frame fluorescent protein fusion to the VP26 small capsid protein. (B) Infected cell lysates or cell-free virus particles isolated by ultracentrifugation were analyzed by SDS-PAGE and Western blotting. Blots were probed with monoclonal anti-VP5 major capsid protein, polyclonal anti-VP26 small capsid protein, and monoclonal anti-mCherry, as indicated. (C) Single-step virus replication. Parallel cell cultures were infected with the indicated viruses. Cells and supernatants were harvested at indicated times, and infectious virus titer was measured by plaque assay. (D) Plaque size analysis. Confluent cell monolayers infected with the indicated viruses were fixed and stained with methylene blue to visualize plaques. Images represent 1 cm².

Recently, a high-resolution cryoelectron microscopy (cryoEM) study of Kaposi's sarcoma herpesvirus (KSHV; human herpesvirus 8) determined the secondary structure of the small capsid protein ORF65 bound to capsid. In this 6-Å capsid structure, the alpha-helical N-terminal half of the ORF65 protein binds to a hydrophobic groove on the surface of each major capsid protein subunit, but the C-terminal half of the ORF65 protein likely is flexible and was not resolved. Using structure-guided mutagenesis, the authors further showed that the C-terminal half of ORF65 is dispensable (13). Because the N terminus of herpesvirus small capsid proteins are structurally conserved (see Fig. 1), we hypothesized that the C terminus of VP26 tolerates fluorescent protein fusions better than the N terminus.

FIG 1.

Herpesvirus small capsid proteins are structurally homologous. (A to C) CryoEM reconstructions of the apical region of capsid hexons (capsid protein hexamer, gray) with associated small capsid protein. The density representing one of the six small capsid protein molecules is colored yellow to illustrate how the small capsid protein binds to the capsid hexon. (A) KSHV hexon reconstruction (6.0 Å), rendered from EMD-6038, originally published by Dai et al. (13). (B) HSV-1 hexon reconstruction (6.8 Å), rendered from EMD-6386, originally published by Huet et al. (35). (C) PRV hexon reconstruction (7.2 Å), rendered from EMD-6387, originally published by Huet et al. (35). (D) Secondary structure prediction showing conserved N-terminal alpha-helices (highlighted and labeled to correspond to structural features indicated in panels A to C).

C-terminal fluorescent protein fusions to the small capsid protein ORF23 have been described in varicella-zoster virus (VZV; human herpesvirus 3); however, the recombinant virus was unable to replicate (14, 15) unless untagged wild-type ORF23 is coexpressed (14) (Table 1). Recently, a C-terminal fusion to VP26 also has been described in PRV (17). In this report, the authors fused the engineered dehalogenase enzyme, HaloTag (18), to the C terminus of VP26, expressed from an ectopic locus. This recombinant PRV also expresses a 2-fold greater amount of untagged wild-type VP26 from its native locus. This strategy of coexpressing a tagged protein and the untagged wild-type proteins is not ideal because herpesviruses are highly recombinogenic, so duplicated sequences have a high likelihood of recombining to produce unexpected phenotypes (10). It remains unknown whether PRV can tolerate C-terminal VP26-FP fusions without coexpressing untagged wild-type VP26.

TABLE 1.

Summary of previous small capsid protein fusions in the alphaherpesviruses

Virus	Small capsid protein	Fusion	Coexpress with untaggeda	Virus replication	Reference or source
HSV-1	VP26	N-terminal FP	−	+	1
VZV	ORF23	N-terminal FP	−	+	15, 16
		C-terminal FP	−	−	15
			+	+	14
PRV	VP26	N-terminal FP	−	+	2
		C-terminal HaloTag	+	+	17
		C-terminal FP	−	+	This report

^aRecombinant viruses express both tagged and untagged small capsid proteins.

In this study, we describe a new PRV recombinant, PRV 1028, expressing a C-terminal VP26-mCherry fusion. VP26-mCherry is expressed from the native locus, and this virus does not coexpress untagged wild-type VP26. Compared to previously characterized recombinants expressing N-terminal fusions, PRV 1028 expresses more VP26 fusion protein and incorporates more VP26 fusion protein into virus particles, and individual virus particles exhibit brighter red fluorescence. We performed single-particle tracking of fluorescent virus particles in primary neurons to measure anterograde and retrograde axonal transport, demonstrating the utility of this new C-terminal VP26-mCherry fusion for the study of viral intracellular transport.

RESULTS AND DISCUSSION

Small capsid protein secondary structure and capsid binding are conserved among the Herpesviridae.

The KSHV small capsid protein ORF65 binds to a hydrophobic groove on the major capsid protein of capsid hexons, adopting an extended conformation. A fold-back helix positions the N terminus of the protein between two adjacent major capsid protein monomers, an extended stem helix spans each major capsid protein monomer, and a bridge helix cross-links two adjacent major capsid protein monomers (Fig. 1A). The C-terminal half of ORF65 is flexible and is not resolved in this structure (13). Based on a secondary structure prediction and lower-resolution cryoEM structures, the small capsid proteins of the alphaherpesviruses are structurally conserved, with similar fold-back, stem, and bridge helices (Fig. 1B to D). Note that fluorescent protein fusions to the N terminus of the small capsid proteins would place the fluorescent protein moiety between two adjacent major capsid protein monomers, perhaps leading to steric conflicts in incorporation.

Validation of fluorescent constructs.

A novel PRV recombinant, PRV 1028, expressing an mCherry fluorescent protein fusion to the C terminus of VP26, was generated by de novo gene synthesis and homologous recombination (Fig. 2A). To validate expression of VP26 fusions, we infected cells with PRV Becker (wild-type), PRV 180 (mRFP1-VP26), PRV 960 (mCherry-VP26), and PRV 1028 (VP26-mCherry) and performed Western blotting on cell lysates or cell-free virus particles with antibodies to detect the major capsid proteins VP5, VP26, and mCherry. The VP5 signal serves as a control to ensure equal infection, viral gene expression, and virus particle production. Compared to untagged VP26, the major xFP-VP26 bands are shifted by about 30 kDa, consistent with the added molecular mass of the fluorescent protein fusions. Both mCherry VP26 fusion proteins (PRV 960 and PRV 1028) are detected by VP26 and mCherry antibodies. Based on band densities, PRV 1028 appears to express more VP26 fusion protein in both infected cell lysates and cell-free virus particles than PRV 180 and PRV 960 (Fig. 2B). Note that smaller bands, less than ∼37 kDa, are consistent with previously described (19) hydrolysis at the acylimine group in the red fluorescent protein chromophore during SDS-PAGE sample preparation (the site of this cleavage is indicated with an asterisk in Fig. 2A). Recombinant viruses expressing fluorescent protein fusions exhibit very little reduction in single-step replication (Fig. 2C) and, as previously reported (10), slight but statistically significant reductions in plaque size (Fig. 2D).

Fluorescence expression and intranuclear capsid protein aggregation.

We next imaged infected cell monolayers and individual virus particles to assess fluorescence intensity. The red fluorescent protein mCherry is reported to be between 27 and 37% brighter than its predecessor, mRFP1 (20, 21). Accordingly, PRV 960-infected cells and virus particles, expressing mCherry-VP26, are brighter than those of PRV 180, expressing mRFP1-VP26 (Fig. 3A and B). PRV 1028-infected cells, expressing the C-terminal fusion VP26-mCherry, are ∼30% brighter than PRV 960-infected cells (Fig. 3A), consistent with the higher protein expression observed by Western blotting (Fig. 2B). More of the VP26-mCherry fusion protein is also incorporated into virus particles (Fig. 2B), producing particles that are ∼30% brighter (Fig. 3B).

FIG 3.

Infected cell and virus particle fluorescence intensity and intranuclear aggregate formation. (A) Confluent cell monolayers were infected with the indicated viruses and imaged at 12 h postinfection by fluorescence microscopy at ×10 magnification. Fluorescence intensity was quantified over 10 random fields of view. Error bars represent standard deviations. (B) Infected cell supernatants were harvested at 12 to 14 h postinfection and imaged by fluorescence microscopy at ×100 magnification. Fluorescence intensity of individual virus particles was quantified over 18 random fields of view. Error bars represent standard deviations. (C to E) Quantification of intranuclear capsids and aggregates. (C) Representative infected cell nuclei imaged at 6 h postinfection by fluorescence microscopy at ×100 magnification. The nucleus boundary is indicated by the dashed line. Scale bars represent 10 μm. (D and E) At least 100 infected cell nuclei were manually scored for severity of intranuclear capsid protein aggregates and the quantity of individual intranuclear capsids, using previously described (10) scoring criteria. (D) The following intranuclear aggregate scores were used: +++, very large aggregates; ++, midsized aggregates, as in panel C; +, small aggregates; −, no aggregates. (E) The following scores for intranuclear capsid concentration were used: +++, dense; ++, many individual capsids, as in panel C; +, few individual capsids.

Previously, we discussed the propensity of VP26 fluorescent protein fusions to exacerbate the formation of intranuclear capsid protein aggregates, which inhibit virus particle production (10). Higher expression levels of VP26 fusions and use of fluorescent protein variants retaining a residual multimerization affinity lead to particularly severe intranuclear aggregate formation (9, 10). To determine whether the C-terminal VP26-mCherry fusion affects capsid assembly or formation of intranuclear aggregates, we imaged infected cell nuclei at 6 h postinfection and blindly scored the extent of large intranuclear aggregates versus density of small individual capsid puncta. Our scoring criteria were previously developed to capture the broad range of phenotypes we observed using a variety of other PRV recombinants expressing VP26 capsid fusions (10). Consistent with previous observations of N-terminal fusions in PRV (10), each of the viruses expressing a red fluorescent protein VP26 fusion produced an intermediate level of intranuclear capsid protein aggregates and an acceptable density of individual intranuclear capsid (Fig. 3C to E). The C-terminal VP26-mCherry fusion expressed by PRV 1028 appears to cause a modest increase in the number of cells with large intranuclear aggregates (Fig. 3D) but does not cause the severe intranuclear aggregates we previously observed with other fusions, e.g., EGFP-VP26 (10), despite higher protein expression in infected cells.

Competition for incorporation of N-terminal versus C-terminal VP26 fusions.

We initially hypothesized that fluorescent protein fusion to the unstructured C terminus of VP26 reduces steric hindrances, allowing more incorporation of the VP26-mCherry fusion protein. Alternatively, since we observed more VP26-mCherry is expressed in infected cells (Fig. 2B and 3A), greater incorporation of VP26-mCherry into virus particles may be due to this higher protein expression. To determine which of these alternative hypotheses best explains the higher incorporation of VP26-mCherry into virions, we performed a competition assay in which virus particles can incorporate various amounts of two different fluorescent fusion proteins.

First, as a control condition under which there is no competition for incorporation of different proteins, we reanalyzed previously described microscopy data (22) in which virus particles incorporate a red VP26 fusion into capsids and a green glycoprotein fusion into the virion envelope (mRFP1-VP26 and gM-EGFP). Since incorporation of capsid proteins and membrane glycoproteins into virus particles occurs at spatially and temporally distinct sites in an infected cell, we expect there should be no evidence of competition (i.e., green and red fluorescence intensities should be independently distributed). As expected, we observe no significant correlation between red and green fluorescence (correlation P value of >0.01) (Fig. 4A).

FIG 4.

Competition for incorporation into virus particles. (A to C) Infected cell supernatants were imaged by fluorescence microscopy at ×100 magnification, and the fluorescence intensity of individual virus particles is displayed as a scatter plot. (A) Negative-control condition with no evidence of competition. PRV recombinants express mRFP1-VP26 fusion and/or a green fluorescent protein fusion to viral glycoprotein M. These fluorescent fusion proteins are incorporated into virus particles independently, and red and green fluorescence intensities are not correlated. (B) Positive-control condition with competition. A cell line expressing N-terminal mNeonGreen-VP26 (PK-mNG-VP26 cells) was infected with PRV 180 expressing N-terminal mRFP1-VP26, and virus particles were harvested at the indicated time points. Because red and green N-terminal fluorescent protein fusions are competing for the same capsid binding sites, red and green fluorescence intensities are inversely correlated. (C) Competition of N-terminal versus C-terminal VP26 fusions. Cells were coinfected with PRV recombinants expressing N-terminal mNeonGreen-VP26 and either N-terminal mCherry-VP26 or C-terminal VP26-mCherry. Red and green fluorescence intensities are inversely correlated in both cases and fall along approximately the same trendline, suggesting that the N- and C-terminal fusions are competing for the same capsid binding sites.

As a positive control demonstrating competition between two different fluorescent fusion proteins, next we infected a cell line expressing an N-terminal green VP26 fusion with a virus expressing an N-terminal red VP26 fusion (mNeonGreen-VP26, mRFP1-VP26). Over the course of infection, expression of the mRFP1-VP26 fusion protein from the viral genome gradually increases, while expression of the mNeonGreen-VP26 transgene in the cellular genome is gradually downregulated. Consistent with previous results (23), virus particles released at 8 hpi contain mostly mNeonGreen-VP26, virus particles released at 14 hpi are beginning to incorporate more mRFP1-VP26, and virus particles released at 20 hpi contain a mixture of mNeonGreen-VP26 and mRFP1-VP26. Because mNeonGreen-VP26 and mRFP1-VP26 are competing for the same capsid binding sites, we observe a trade-off between green and red fluorescence (correlation P value of <0.01) (Fig. 4B).

Finally, to assess whether our initial hypothesis, that C-terminal VP26-mCherry fusion reduces steric hindrances to allow more incorporation, we performed coinfections with equal concentrations of PRV expressing the N-terminal fusion mNeonGreen-VP26 (PRV 959) and either N- or C-terminal red VP26 fusions (PRV 960 or PRV 1028, respectively). Under the condition where we coinfect with viruses expressing N-terminal fusions, mNeonGreen-VP26 and mCherry-VP26 are competing for the same capsid binding sites, so we again observe a trade-off between green and red fluorescence (correlation P value of <0.01) (Fig. 4C).

If C-terminal VP26-mCherry reduces steric hindrance, we expect to see less evidence of competition. If, for example, mNeonGreen-VP26 incorporates 300 to 400 copies of fusion protein (the approximate observed maximum for N-terminal fusions) out of 900 potential VP26 binding sites on each capsid, the remaining majority of binding sites might remain accessible to the C-terminal VP26-mCherry fusion. Thus, we might expect to observe increased incorporation of VP26-mCherry without a corresponding decrease in mNeonGreen-VP26 fluorescence due to competition. Under the condition where we coinfect with viruses expressing N-terminal mNeonGreen-VP26 and C-terminal VP26-mCherry, we observe a trade-off between green and red fluorescence (correlation P value of <0.01) that falls along the same trendline as that of N-terminal mCherry-VP26 competition (Fig. 4C). Therefore, it is likely that the C-terminal VP26-mCherry fusion does not relieve the hypothetical defect that causes N-terminal fusions to incorporate less than expected. Rather, the C-terminal fusion may incorporate more simply because of its higher expression level in infected cells.

Retrograde and anterograde axonal transport.

To measure postentry retrograde transport and postreplication anterograde transport of progeny in neurons, we took advantage of the well-established modified Campenot trichamber neuronal culture system (Fig. 5A and D). The trichamber provides fluidic separation between neuronal cell bodies and axons that penetrate under the chamber wall. To measure postentry retrograde transport, we added virus stock to the rightmost axonal compartment, incubated between 15 and 90 min to allow particle entry and retrograde transport to begin, and then imaged by fluorescence microscopy (Fig. 5A). PRV 960 and PRV 1028 particles had nearly identical velocity distributions (Fig. 5B and C), indicating that the new C-terminal VP26-mCherry fusion does not affect motor recruitment and retrograde axonal transport. The velocity distribution was fit well by a Gaussian distribution with an average velocity of 1.2 μm/s (Fig. 5B and C), consistent with previous observations of PRV particle transport as well as published measurement of cytoplasmic dynein microtubule motors.

FIG 5.

Virus particle transport in axons of primary neurons. Primary superior cervical ganglion neurons in Campenot trichamber cultures were inoculated with the indicated virus. Individual virus particles transporting in axons were tracked in live-cell fluorescence microscopy. (A to C) Virus was added to the right axonal chamber and imaged soon after to measure retrograde postentry axonal transport. (B and C) Both PRV 960 and PRV 1028 exhibited retrograde transport with an average velocity of 1.2 μm/s and nearly identical velocity distributions. The red line represents Gaussian fit. (D to F) Virus was added to the left cell body chamber, incubated 12 to 14 h to allow a single replication cycle, and imaged to measure anterograde axonal transport of progeny virus particles. (E and F) Both PRV 960 and PRV 1028 exhibited anterograde transport with an average velocity of 1.5 μm/s and very similar velocity distributions. The red line represents Gaussian fit. Note that velocity distributions are broader and less well fit by a Gaussian distribution.

To measure postreplication transport of progeny particles, we inoculated the leftmost cell body compartment, incubated it 12 to 14 h to allow a single round of replication, and then measured the velocities of progeny particles trafficking in the anterograde direction (Fig. 5D). PRV 960 and PRV 1028 particles had very similar velocity distributions (Fig. 5E and F), indicating that the new C-terminal VP26-mCherry fusion does not affect motor recruitment during anterograde axonal transport. Again consistent with previous reports (23), the velocity distributions of anterograde transport were broader and fit less well by a Gaussian distribution, suggesting that anterograde transport of virus particles is mediated by the ensemble activity of multiple kinesin motors (24, 25). Anterograde-directed particles exhibited instantaneous velocities greater than 2 μm/s, consistent with faster kinesin-3 motors (published typical velocities of 2.6 μm/s [26]) but an average velocity of 1.5 μm/s, which is also consistent with a role for kinesin-1 (published typical velocity of 1.2 μm/s [26]) (Fig. 5E and F).

In conclusion, we have demonstrated that PRV can tolerate fluorescent protein fusions to the C terminus of small capsid protein VP26 without the need to coexpress wild-type unfused VP26. As the C terminus of herpesvirus small capsid proteins is not conserved phylogenetically and unstructured by cryoEM, tagging the C terminus of these proteins may be applicable to other herpesviruses. The recombinant PRV 1028, expressing the C-terminal VP26-mCherry fusion, expressed approximately 30% more of the fusion protein and produced virus particles that are ∼30% brighter. In single-particle tracking experiments, particle localization accuracy is a function of the inverse square root of the number of photons acquired (27). Thus, a particle that is 30% brighter will allow for a 30% reduction in acquisition time while maintaining the same localization accuracy, or a 14% improvement in localization accuracy given the same imaging parameters. Interestingly, although we initially hypothesized that tagging the C terminus of VP26 would relieve steric hindrances to allow greater incorporation, we observed a modest ∼30% increase in C-terminal VP26-mCherry particle fluorescence, which is still much less than we would expect if VP26-mCherry molecules were occupying all 900 potential VP26 binding sites. In competition experiments, C-terminal VP26-mCherry appears to be competing with N-terminal mNeonGreen-VP26 for the same capsid binding sites, suggesting additional, unresolved reasons why fewer than expected tagged VP26 molecules incorporate into each virus particle.

MATERIALS AND METHODS

Cells and viruses.

PK15 cells, from a porcine kidney epithelial cell line, were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% penicillin-streptomycin and 10% fetal bovine serum (FBS). All PRV variants were propagated on PK15 cells in DMEM supplemented with 1% penicillin-streptomycin and 2% FBS. PK-mNG-VP26 cells, a derivative of PK15 cells expressing an mNeonGreen-VP26 fusion protein, were previously described (23). Parental PRV Becker, PRV 180, which expresses an mRFP1-VP26 fusion, PRV 959, which expresses an mNeonGreen-VP26 fusion, and PRV 960, which expresses an mCherry-VP26 fusion, were previously described (10, 28). PRV 1028, which expresses a C-terminal VP26-mCherry fusion, was constructed as follows. A plasmid, pUC57-PRV-VP26-mCherry(C-term), was synthesized (GenScript), encoding the VP26-mCherry fusion protein flanked by homologous sequences to facilitate the insertion of this transgene into the PRV genome by homologous recombination. The fusion junction is the following (VP26 sequence is indicated in boldface type, the linker peptide in plain type, and mCherry sequence in italics): …KRTFCPRPPPGGSGVSKGEEDNMA…. This plasmid was linearized by restriction digestion and cotransfected with purified PRV Becker nucleocapsid DNA into PK15 cells using Lipofectamine 2000 (Invitrogen). Recombinant virus was selected through three rounds of plaque purification, and a high-titer virus stock (1.7 × 10⁸ PFU/ml) was grown on PK15 cells.

Western blotting.

PK15 cells were infected with PRV at a multiplicity of infection (MOI) of 10 PFU/cell. At 12 h postinfection (hpi), infected cells were collected and directly lysed in Laemmli sample buffer for SDS-PAGE. Viral supernatants were clarified by low-speed centrifugation and passed through a 0.45-μm filter to remove cell debris, and virus particles were pelleted by ultracentrifugation (23.5 krpm in a Beckman SW28 rotor) through a 20% (wt/vol) sorbitol cushion in 50 mM Tris buffer. Virus particle pellets then were lysed in Laemmli sample buffer. SDS-PAGE and Western blotting were performed as previously described (29). Blots were probed with the following primary antibodies: mouse monoclonal anti-VP5, rabbit polyclonal anti-VP26, and mouse monoclonal anti-mCherry (Clontech).

Single-step replication and plaque size measurements.

A confluent monolayer of PK15 cells was inoculated with PRV at an MOI of 5 PFU/cell. Cultures were incubated at 37°C for 1 h and then washed three times with phosphate-buffered saline (PBS). Cells and supernatants were harvested at the indicated times postinfection, and titers were determined by serial dilution plaque assay. To measure plaque diameters, PK15 cell monolayers with PRV plaques at ∼3 days postinfection were fixed and stained with 0.5% (wt/vol) methylene blue in 70% (vol/vol) methanol, rinsed, air dried, and imaged using a commercial photo scanner. Plaque diameters were recorded using the measurements tool in Fiji/ImageJ (v. 1.48) (30).

Fluorescence microscopy.

PK15 cells in 35-mm glass-bottom Mat-Tek dishes were infected with PRV at an MOI of 5 PFU/cell. At 6 hpi, infected cells were imaged using a previously described epifluorescence microscope (31) using a 100× objective. Infected cell nuclei containing red capsid protein fluorescence (n = 100) were blindly scored based on the density of individual capsids in the nucleus and extent of intranuclear aggregate formation, as previously described (10). At 12 hpi, infected cell monolayers were imaged using a 10× objective, and mean fluorescence intensity was measured from at least 10 random fields of view using the measurements tool in Fiji/ImageJ. At 12 to 14 hpi, infected cell supernatants were collected and spotted onto glass-bottom Mat-Tek dishes and incubated briefly to allow particles to nonspecifically adhere to the glass without drying. Individual virus particles in at least 18 random fields of view were imaged using a 100× objective, and particle fluorescence intensities were measured using the analyze particles function in Fiji/ImageJ.

Axonal transport.

Animal work was approved by the Princeton University Institutional Animal Care and Use Committee (protocol 1947-16) and performed in accordance with the American Association for the Accreditation of Laboratory Animal Care and the Animal Welfare Act. Superior cervical ganglia (SCG) were dissected from embryonic day 16.5 (E16.5) to E17.5 rat embryos and stored at 4°C in Hibernate-E medium (Invitrogen) for up to 1 week. SCG neurons were dissociated and plated in modified Campenot trichamber devices, as previously described (31, 32). SCG neurons were cultured in Neurobasal medium supplemented with 1% penicillin-streptomycin–glutamine, B27, and 50 ng/ml nerve growth factor (Invitrogen) for at least 14 days to establish a fully polarized and mature neuronal phenotype. To measure postentry retrograde axonal transport, ∼10⁷ PFU of recombinant PRV was inoculated into the axonal compartment, and transporting particles were imaged in the axonal and middle compartments within an hour of inoculation. To measure postreplication anterograde axonal transport of newly assembled progeny, ∼10⁷ PFU of recombinant PRV was inoculated into the cell body compartment, and transporting particles were imaged in the axonal and middle compartments at 12 to 14 hpi. Virus particle transport was imaged using a 100× objective on a previously described epifluorescence microscope (31) at 10 frames/s. Single-particle tracking was performed in Fiji/ImageJ using the Mosaic Particle Tracker 2D/3D plugin (33).

Protein structure modeling.

Protein structure prediction based on amino acid sequence was performed using I-TASSER (34). Previously published cryoEM density maps were downloaded from the Electron Microscopy DataBank (www.emdatabank.org), accession numbers EMD-6038 (13), EMD-6386, and EMD-6387 (35). Surface renderings of cryoEM density maps were prepared using UCSF Chimera (36).