The Carboxyl Terminus of Tegument Protein pUL21 Contributes to Pseudorabies Virus Neuroinvasion

Herpesviruses are a group of DNA viruses that infect both humans and animals. Alphaherpesviruses are distinguished by their ability to establish latent infection in peripheral neurons. After entering neurons, the herpesvirus capsid interacts with cellular motor proteins and undergoes retrograde transport on axon microtubules. This elaborate process is vital to the herpesvirus lifecycle, but the underlying mechanism remains poorly understood. Here, we determined that pUL21 is an inner tegument protein of pseudorabies virus (PRV) and that it interacts with the cytoplasmic dynein light chain Roadblock-1. We also observed that pUL21 promotes retrograde transport of PRV in neuronal cells. Furthermore, our findings confirm that pUL21 contributes to PRV neuroinvasion in vivo. Importantly, the carboxyl terminus of pUL21 is responsible for interaction with Roadblock-1, and this domain contributes to PRV neuroinvasion. This study offers fresh insights into alphaherpesvirus neuroinvasion and the interaction between virus and host during PRV infection.

ABSTRACT

Following its entry into cells, pseudorabies virus (PRV) utilizes microtubules to deliver its nucleocapsid to the nucleus. Previous studies have shown that PRV VP1/2 is an effector of dynein-mediated capsid transport. However, the mechanism of PRV for recruiting microtubule motor proteins for successful neuroinvasion and neurovirulence is not well understood. Here, we provide evidence that PRV pUL21 is an inner tegument protein. We tested its interaction with the cytoplasmic light chains using a bimolecular fluorescence complementation (BiFC) assay and observed that PRV pUL21 interacts with Roadblock-1. This interaction was confirmed by coimmunoprecipitation (co-IP) assays. We also determined the efficiency of retrograde and anterograde axonal transport of PRV strains in explanted neurons using a microfluidic chamber system and investigated pUL21’s contribution to PRV neuroinvasion in vivo. Further data showed that the carboxyl terminus of pUL21 is essential for its interaction with Roadblock-1, and this domain contributes to PRV retrograde axonal transport in vitro and in vivo. Our findings suggest that the carboxyl terminus of pUL21 contributes to PRV neuroinvasion.

IMPORTANCE Herpesviruses are a group of DNA viruses that infect both humans and animals. Alphaherpesviruses are distinguished by their ability to establish latent infection in peripheral neurons. After entering neurons, the herpesvirus capsid interacts with cellular motor proteins and undergoes retrograde transport on axon microtubules. This elaborate process is vital to the herpesvirus lifecycle, but the underlying mechanism remains poorly understood. Here, we determined that pUL21 is an inner tegument protein of pseudorabies virus (PRV) and that it interacts with the cytoplasmic dynein light chain Roadblock-1. We also observed that pUL21 promotes retrograde transport of PRV in neuronal cells. Furthermore, our findings confirm that pUL21 contributes to PRV neuroinvasion in vivo. Importantly, the carboxyl terminus of pUL21 is responsible for interaction with Roadblock-1, and this domain contributes to PRV neuroinvasion. This study offers fresh insights into alphaherpesvirus neuroinvasion and the interaction between virus and host during PRV infection.

INTRODUCTION

Alphaherpesviruses are a group of large double-stranded DNA (dsDNA) viruses. They are able to establish a lifelong latent infection in neurons within the peripheral nervous systems of their hosts from which the viruses can later reactivate, causing disease (1–4). After entering a cell, alphaherpesvirus particles travel through the cytoplasm and replicate in the nucleus. Progeny virions then traffic back to the cell surface for egress (5). In the nervous system, alphaherpesviruses use microtubules as a means of transport to translocate viral genomes to neurons in sensory ganglia (retrograde transport). Afterward, progeny viruses reappear from the neurons, spreading to innervated peripheral tissues (anterograde transport) (6).

Herpesviruses share a virion structure composed of a linear dsDNA genome, an icosahedral capsid, a tegument layer comprising more than 20 proteins that are sometimes subclassified as components of the inner and outer tegument, and a lipid bilayer envelope studded with viral glycoproteins (7–9). Upon entering host cells, the envelope and outer tegument proteins disassociate (10). The remaining capsid-inner tegument complex retrogradely transports to the nucleus in a microtubule-dependent manner (11). An experiment on tegument protein associations with capsids categorized seven herpes simplex virus 1 (HSV-1) proteins as inner tegument proteins, pUL14, pUL16, pUL21, pUL36 (VP1/2), pUL37, pUS3, and ICP0 (12). Dynein binds directly to the inner tegument of herpesviruses during cytoplasmic transport (12). Several proteins within the capsid-tegument complex are related to retrograde transport along microtubules (6, 13, 14). Among them, the large inner tegument protein pUL36 (VP1/2) tethers to the dynein-dynactin complex and promotes retrograde capsid transport. VP1/2 interacts with the dynein-dynactin motor complex through its proline-rich region. Deletion of this region weakens virulence and retrograde axonal transport of PRV. However, the capsids of mutant PRV still can reach the nuclei of neurons (6). These results imply that other viral proteins might be involved in the retrograde transport of capsids.

The tegument protein pUL21 is conserved throughout the alphaherpesvirus family (15–17). In HSV-1, this protein is dispensable for replication in cultured cells (18), but in HSV-2, pUL21 is essential for virus production in both early and late stages of the virus replication cycle (16). Initially, pUL21 was considered to be a capsid protein related to capsid maturation (19); however, the HSV-1 pUL21 was proven later to be a component of the tegument, and it was demonstrated to promote the generation of long cellular processes, perhaps owing to its interaction with microtubules (20). Deletion of pUL21 in HSV-1 leads to a delay in the early stages of the viral replication cycle and to the formation of smaller plaques (21, 22). In PRV, the pUL21 protein is dispensable for replication in porcine tissues, and the deletion of pUL21 from PRV causes the formation of DNA-deprived capsids, indicating a potential role for pUL21 in viral genome packaging (23). The deletion of PRV pUL21 also results in a drastic decrease in the incorporation of the pUL46, pUL49, and pUS3 tegument components into mature virions, showing that pUL21 is important for virion assembly (24). PRV pUL21 is also involved in neuroinvasion and neurovirulence (25) and is a determinant of PRV virulence. Point mutations in the UL21 gene of PRV Bartha contribute to its attenuation (26). Analysis of PRV interstrain diversity revealed three amino acid mutations in the pUL21 of PRV Bartha (27). Furthermore, repair of the UL21 gene in PRV Bartha enhances retrograde transport and transneuronal infection (28). Recently, the crystal structures of the N-terminal and the C-terminal domains of HSV-1 pUL21 were determined (17, 29). Although pUL21 is involved in several stages of viral replication, the underlying mechanism of how pUL21 deletion leads to reduced neuroinvasion is unclear.

The cytoplasmic dynein motor transports cargo to the minus end of microtubules (30). The dynein complex is involved in the transport of herpesvirus in the early stage of infection (31). This large complex consists of two heavy chains, two intermediate chains, two light-intermediate chains, and several light chains (32–35). A dynein motor complex can hold up to three different light-chain dimers, Tctex-1/Tctex-3, Roadblock-1/Roadblock-2, and LC8-1/LC8-2 (32). Extensive analyses have demonstrated that many different cargoes and regulatory proteins can interact with a subset of components in the dynein complex (31, 35). The dynein light chain Roadblock, initially identified in Drosophila and Chlamydomonas spp., is a conserved family with roles related to axon transport, flagellar motility, and mitosis (36). Roadblock is incorporated into the dynein complex through protein-protein interactions (37, 38). This protein is ubiquitous and is expressed differentially in a diversity of cells and tissues (39, 40). Roadblock proteins reside in the cellular cytoplasm as punctate structures and accumulate outside the nucleus (40). The three-dimensional structure of Roadblock was determined in previous work; it can exist as a monomer or symmetric homodimer (41, 42). However, it has not yet been reported if Roadblock can directly bind to viral components during infection.

In this study, we aimed to determine if pUL21 is an inner tegument protein of PRV and whether it interacts with the cytoplasmic dynein chains. We also performed in vitro studies on epithelial cells and explanted neurons as well as in vivo studies in mice to determine the effect of pUL21 deletion in PRV.

RESULTS

pUL21 is an inner tegument protein of PRV.

After entering host cells, the envelope and outer tegument proteins of herpesvirus are lost (10). The incoming motor dynein directly binds to the inner tegument during cytoplasmic transport. Previous work using an assay based upon the resistance of tegument proteins to detergent and salt extraction categorized seven HSV-1 inner tegument proteins (12). Based on this finding for the pUL21 protein of HSV-1, the pUL21 protein of PRV is probably also an inner tegument protein. To confirm this possibility, extracellular virions of wild-type (WT) PRV and PRV EGFP-VP1/2 were sedimented and treated as described in Materials and Methods. KCl-treated and untreated extracellular virions were analyzed, and residual proteins remaining on the capsids were quantified by Western blotting assays (Fig. 1A and B) and densitometry analyses (Fig. 1C). The major capsid protein VP5 was set as a loading control. After KCl treatment, the amount of capsid protein VP26 was as stable as that of VP5. Because the inner tegument protein VP1/2 is tightly combined with the capsid (6, 12), its residual amount was also stable after KCl treatment, as expected. In contrast, since outer tegument proteins VP16 and VP13/14 are loosely attached to the capsid (12), their residual amounts were dramatically lower after KCl treatment, also as expected. The glycoprotein gB was removed by the treatment because it was sensitive to the detergent in the KCl treatment (12, 43). Notably, the residual amount of tegument protein pUL21 was as stable as that of the inner tegument protein VP1/2 (Fig. 1). These results suggest that pUL21 is an inner tegument protein of PRV.

FIG 1.

Western blot analyses of residual proteins on KCl-treated PRV capsids. (A and B) Extracellular virions of WT PRV (A) and PRV EGFP-VP1/2 (B) were sedimented and treated as described in Materials and Methods. KCl-treated and untreated extracellular virions were analyzed. Residual proteins on capsids were quantified by Western blot analyses. (C) Capsid protein VP5 was set as the loading control (100%). The normalized relative protein amount was quantified by densitometry scanning (ImageJ2x).

The tegument protein pUL21 interacts with the cytoplasmic dynein light chain Roadblock-1.

Previous evidence shows that the inner tegument protein VP1/2 tethers to the dynein-dynactin complex during retrograde capsid transport (6). Therefore, we hypothesized that the inner tegument protein pUL21 might also bind to microtubule motor proteins during infection. To test this hypothesis, the PRV UL21 gene and nine cytoplasmic dynein chain genes, including light-chain genes LC8-1, LC8-2, Roadblock-1, Tctex-1, and Tctex-3, light-intermediate-chain genes DYNC1LI1, DYNC1LI2, and DYNC2LI1, and intermediate-chain gene DYNC1I2 were cloned and tested using a bimolecular fluorescence complementation (BiFC) assay. The BiFC assay has been extensively applied to investigating protein-protein interactions due to its simplicity and high sensitivity (43–46). The principle of the BiFC assay is to divide a full-length fluorescent protein into two nonfluorescent fragments that are separately fused to two proteins; if these two labeled proteins interact with each other, the two fragments of the fluorescent protein obtain close enough proximity to reconstruct an integral fluorescent protein (47, 48) (Fig. 2A).

FIG 2.

Bimolecular fluorescence complementation of Roadblock-1 and PRV pUL21. (A) Schematic diagram of the bimolecular fluorescence complementation (BiFC) assay. (B) BiFC assays were performed to assess the potential interaction of pUL21 with the cytoplasmic dynein light chains. HA, hemagglutinin; RB1, Roadblock-1. Scale bars = 20 μm. (C) Western blot analysis of cotransfected cell lysates. LC8-1, Tctex-1, Roadblock-1, pUL21, and pUL51 were expressed in cotransfected cells. **, dynein light chains (LC8-1, Tctex-1, or Roadblock-1) fused with the C terminus of RFP (LC151). *, endogenous dynein light chains.

The PRV UL21 gene and the above-described nine cytoplasmic dynein genes were cloned into the pHA-KN151 and pmyc-LC151 BiFC vectors, respectively. The PRV pUL21 was fused to the N terminus (KN151) of the red fluorescence protein (RFP), and the cytoplasmic dynein chains were each connected to the C terminus (LC151) (Fig. 2A). Recombinant plasmids were cotransfected into HeLa cells in pairs for screening protein interactions. Cells were examined under a fluorescence microscope at 24 h posttransfection. Lumin is a full-length RFP (44), and cells transfected with pLumin were set as the cytoplasm positive control (Fig. 2B). Fos and Jun are a known pair of interacting proteins (44, 49), and cells cotransfected with pJun-HA-KN151 and pFos-myc-LC151 showed red fluorescence, as expected (Fig. 2B). Roadblock is able to form homodimers (40–42). Cells cotransfected with pRoadblock-1-HA-KN151 and pRoadblock-1-myc-LC151 also displayed red fluorescence, as expected (Fig. 2B). These findings indicate that the BiFC assay was working properly. Red fluorescence was also detected in the cytoplasm of cells cotransfected with pUL21-HA-KN151 and pRoadblock-1-myc-LC151 (Fig. 2B). Compared with the level of red fluorescence produced by interaction of pUL21 with Roadblock-1, neither the cytoplasmic dynein light chain LC8-1 nor Tctex-1 interact with pUL21 (Fig. 2B). Furthermore, the tegument protein pUL51 does not interact with Roadblock-1 (Fig. 2B). Western blot analysis showed that LC8-1, Tctex-1, Roadblock-1, pUL21, and pUL51 were expressed in cotransfected cells (Fig. 2C). These results demonstrate that the tegument protein pUL21 interacts specifically with the cytoplasmic dynein light chain Roadblock-1 in the cytoplasm.

To confirm our finding that pUL21 interacts with Roadblock-1, we carried out immunoprecipitation (IP) and Western blot assays. PK-15 cells were infected with WT PRV at a multiplicity of infection (MOI) of 10 and harvested at 2 h postinfection (hpi). Cell lysates were analyzed by Western blotting directly or were incubated with the indicated antibodies (Fig. 3A and B). Precipitated proteins were then analyzed by Western blotting. VP16, pUL21, Roadblock-1, LC8-1, and Tctex-1 were all detected in the input samples. Following the IP, Roadblock-1 was observed to have coprecipitated with pUL21. The coprecipitation of Roadblock-1 with VP16 was examined as a negative control; as expected, these proteins did not coprecipitate (Fig. 3A). In agreement with our above-described results from the BiFC assays that the cytoplasmic dynein light chain LC8-1 and Tctex-1 were not able to interact with pUL21 (Fig. 2B), LC8-1 or Tctex-1 did not coprecipitate with pUL21 (Fig. 3B). Next, the PRV UL51 gene was cloned into pcDNA3.1(+). A hemagglutinin (HA) tag was introduced at the N terminus of pUL51. PK-15 cells were transfected and lysed at 24 h posttransfection. Cell lysates were incubated with corresponding antibodies for IP or analyzed by Western blotting directly (Fig. 3B). Consistent with the results in BiFC assays, pUL51 was not able to coprecipitate with Roadblock-1 (Fig. 3B). Previous work reported that pUL21 forms a complex with pUL16 (50–52). Therefore, we also tested the interaction between pUL21 and pUL16 by IP. The result indicates that pUL21 coprecipitated with pUL16, as expected (Fig. 3B). These results further confirm that the PRV tegument protein pUL21 interacts with cytoplasmic dynein light chain Roadblock-1.

FIG 3.

Immunoprecipitation and immunofluorescence assays of Roadblock-1 and PRV pUL21. (A) Immunoprecipitation (IP) and Western blot assays of Roadblock-1 and pUL21. (B) IP and Western blot assays of other protein pairs. HA, hemagglutinin; His, histidine; RB1, Roadblock-1. (C) PK-15 cells were transfected with pcDNA3.1(+)-HA-UL21 and subsequently fixed and permeabilized at 12 h posttransfection for an immunofluorescence assay. Anti-HA tag monoclonal antibody and anti-Roadblock-1 polyclonal antibody were used. DAPI staining indicates the nuclei. Arrow indicates the perinuclear region. Scale bar = 20 μm. (D) PK-15 cells were infected with corresponding PRV strains or mock infected. Anti-pUL21 and anti-Roadblock-1 rabbit polyclonal antibodies, as well as anti-α-tubulin monoclonal antibody, were used for immunofluorescence assays. DAPI staining indicates the nuclei. Scale bar = 25 μm.

According to previous reports, pUL21 localizes not only in the cytoplasm but also in the nucleus (20, 31), specifically at the nuclear rim (16, 18, 30). In contrast, Roadblock is generally located in the cytoplasm as punctate structure and accumulates outside the nucleus (40). Here, we examined the subcellular localization of pUL21 and Roadblock-1 within the same cells. The PRV UL21 gene was cloned into pcDNA3.1(+), and an HA tag was introduced at the N terminus of pUL21. PK-15 cells were transfected with pcDNA3.1(+)-HA-UL21. Cells were fixed and permeabilized at 12 h posttransfection for an immunofluorescence assay. The results showed that Roadblock-1 and pUL21 colocalized in the perinuclear region (Fig. 3C). Next, we tested the colocalization of pUL21 and Roadblock-1 with microtubules. PK-15 cells were infected with corresponding PRV strains or mock-infected. Cells were fixed and permeabilized for an immunofluorescence assay. The results showed that both pUL21 and Roadblock-1 colocalized with microtubules in the perinuclear region (Fig. 3D).

The tegument protein pUL21 promotes PRV retrograde axonal transport in explanted neurons.

To study the contribution of pUL21 to PRV infection, the ΔUL21 and repaired (R) mutant strains were constructed (Fig. 4A). The efficiency of retrograde axonal transport of PRV ΔUL21 in neurons was determined. A previously described microfluidic chamber system (53–55) was used to investigate the axonal transport of PRV strains. The chamber consists of mirror-image compartments connected by microgrooves (Fig. 4B). Chicken dorsal root ganglia (DRG) neurons were plated in the somal compartments, and axon growth was guided into the axonal compartments through the microgrooves of the chambers (Fig. 4B). A total of 2 × 10⁶ PFU of virus was inoculated into each axonal compartment. The entire contents of the axonal and somal compartments were harvested separately at 24 hpi. Virus yields were determined by titration on MDBK cells. In the axonal compartments, the resulting titers of the three viruses were similar (Fig. 4C). In the somal compartments, titers of approximately 1 × 10³ PFU/ml were detected for the WT and repaired strains, but no PRV ΔUL21 was detected (Fig. 4C). These data suggest that the deletion of pUL21 abrogates PRV retrograde axonal transport in explanted DRG neurons. However, this abolishment at 24 hpi may be caused by a delayed retrograde transport of PRV ΔUL21. To test this possibility, we performed the same experiment, but we harvested the virus later, at 48 hpi. Compared with the WT strain, the levels of infectious PRV ΔUL21 detected in somal compartments were dramatically lower (Fig. 4D). These data suggest that pUL21 promotes PRV retrograde axonal transport in explanted DRG neurons.

FIG 4.

The efficiency of retrograde and anterograde axonal transport of PRV ΔUL21 in explanted neurons. (A) Verification of the ΔUL21 and repaired virus strains by Western blot analysis. (B) Chicken DRG neurons in a microfluidic chamber system (53). A, axonal compartments; S, somal compartments. Arrows indicate the somata, and arrowheads indicate the axons. Scale bar = 200 μm. (C and D) Quantification of the efficiency of retrograde axonal spread. Virus was inoculated into axonal compartments. Virus yields of the axonal and somal compartments at 24 hpi (C) or 48 hpi (D) were determined. (E) Quantification of the efficiency of anterograde axonal transport. Virus was inoculated into somal compartments. Virus yields of the axonal and somal compartments at 24 hpi were determined.

Upon reactivation, PRV undergoes axonal sorting by pUS9 and fast axonal transport in the “married model” (56–58). Hence, as an inner tegument protein, pUL21 is probably irrelevant to anterograde axonal transport. To confirm this likely possibility, a total of 2 × 10⁶ PFU of virus was inoculated into each somal compartment. The entire contents of the axonal and somal compartments were harvested separately at 24 hpi, and virus yields were determined by titration on MDBK cells. The harvested titers of each virus strain were fairly similar between the somal compartments and axonal compartments (Fig. 4E). These data show that the deletion of pUL21 has no influence on PRV anterograde spread in explanted DRG neurons.

The tegument protein pUL21 enhances PRV retrograde axonal transport and neuroinvasion in mice.

Next, we investigated the contribution of pUL21 to retrograde axonal transport and neuroinvasion by PRV in mice. The animals were infected by intraocular injection or intranasal instillation. An intraocular injection model, in which a total of 1 × 10⁵ PFU of virus was injected unilaterally into the anterior chamber of the eye, was used for studying retrograde transport (Fig. 5A) (6, 59).

FIG 5.

PRV ΔUL21 virulence and neuroinvasion in mice. (A) Schematic diagram of the intraocular injection model (6, 59). (B) Survival time of mice (n = 8) that were intraocularly infected. Each dot represents one mouse. Means are shown as black lines. ***, P < 0.001. (C) Survival time of mice (n = 6) that were intranasally inoculated. (D) Viral loads in the SCG of PRV-infected mice in intraocular injection group (n = 8). ****, P < 0.0001; n.s., not significant.

In the absence of pUL21, the mean survival time of the intraocularly infected mice was significantly prolonged. Mice infected with WT PRV died at 87 hpi, whereas those infected by the PRV ΔUL21 strain died at 103 hpi (Fig. 5B). Mice infected with the repaired strain died at 88 hpi. In contrast, the deletion of pUL21 had a slight influence (not statistically significant) on the mean survival time of intranasally infected animals. Mice infected with WT PRV died at 108 hpi, those infected by the PRV ΔUL21 strain died at 117 hpi, and those infected with the repaired strain died at 115 hpi (Fig. 5C).

Viruses injected into the anterior chamber can infect the iris (6, 59), and superior cervical ganglia (SCG) contain neurons that project axons into the iris (Fig. 5A). For the intraocular injection groups, mouse SCG were dissociated immediately after death, and their viral loads were determined. Over 50,000 PFU of virus per animal were detected in the SCG from all mice infected with WT PRV; however, only about 3,000 PFU of virus per animal were detected in the SCG from all mice infected with the PRV ΔUL21 strain (Fig. 5D). These data suggest that pUL21 enhances PRV retrograde axonal transport and neuroinvasion in mice.

The carboxyl terminus of pUL21 is essential for its interaction with Roadblock-1.

Then, we mapped the domain of pUL21 that is required for its interaction with Roadblock-1. Full-length/truncated PRV UL21 genes and the Roadblock-1 gene were each cloned into pcDNA3.1(+). An HA tag was introduced at the N terminus of pUL21, while a 6× His tag was added at the N terminus of Roadblock-1 (Fig. 6A). HEK293 cells were transfected and lysed at 24 h posttransfection. Cell lysates were incubated with an anti-6× His tag monoclonal antibody for IP or analyzed by Western blotting directly. Only the full-length, NΔ100, and NΔ200 pUL21 proteins coprecipitated with Roadblock-1. The CΔ300, CΔ200, and CΔ100 pUL21 proteins failed to coprecipitate with Roadblock-1 (Fig. 6B). These results demonstrate that the carboxyl terminus (amino acids [aa] 301 to 533) of pUL21 is essential for the interaction of pUL21 with Roadblock-1.

FIG 6.

Immunoprecipitation assays of Roadblock-1 and truncated versions of PRV pUL21. (A) Schematic diagram of full-length PRV pUL21 and truncations of this protein. Gray dashed lines represent the deleted parts, while gray solid lines represent the remaining portions. Black solid lines represent the HA tags. Numbers indicate the amino acids positions. – or +, negative or positive interaction with Roadblock-1, respectively; HA, hemagglutinin; RB1, Roadblock-1. (B) Immunoprecipitation assays of Roadblock-1 and truncated pUL21. Arrowheads indicate the full-length and truncated versions of pUL21. Asterisks indicate heavy chains (∼55 kDa) and light chains (∼25 kDa) of immunoglobulin.

The carboxyl terminus of pUL21 contributes to virus spread in epithelial cells.

To study the contribution of the carboxyl terminus (aa 301 to 533) of pUL21 to PRV infection, PRV UL21 CΔ300 and the repaired (R) strain were constructed by a two-step Red-mediated recombination using PRV bacterial artificial chromosome (BAC). To verify the deletion of the carboxyl terminus of pUL21 and the repaired strain, PK-15 cells were infected with different strains and harvested. Cell lysates were analyzed by Western blotting. Truncated pUL21 was detected in PRV UL21 CΔ300-infected cells, and full-length pUL21 was restored in repaired strain-infected cells (Fig. 7A). One-step growth curves of these viruses show that PRV UL21 CΔ300 propagated to titers close to those of the WT and repaired strains with normal kinetics (Fig. 7B). Multistep growth curves on PK-15 cells show that PRV UL21 CΔ300 as well as PRV ΔUL21 propagated to titers much lower than those of the WT and repaired strains (Fig. 7C and D). PRV UL21 CΔ300 generates smaller plaques than do the WT and repaired strains, indicating that this strain has a weakened spread in epithelial cells (Fig. 7E). Furthermore, pUL21 CΔ300 was not able to coprecipitate with Roadblock-1, showing that pUL21 CΔ300 failed to interact with Roadblock-1 in infected cells (Fig. 7F). These findings suggest that the carboxyl terminus (aa 301 to 533) of pUL21 does not contribute to productive infection in epithelial cells, and deletion of the carboxyl terminus of pUL21 impairs virus spread in epithelial cells.

FIG 7.

Characterization of PRV UL21 CΔ300. (A) Verification of the PRV UL21 CΔ300 and repaired virus strains by Western blot analysis. (B) One-step growth curve on PK-15 cells. Each data point represents the average of triplicate samples obtained from separate infections. (C and D) Multistep growth curve on PK-15 cells. Intracellular (C) and extracellular (D) virus titers were determined by titration in MDBK cells. Each data point represents the average of triplicate samples obtained from separate infections. (E) Plaque morphology and plaque size of WT PRV, PRV UL21 CΔ300, and the repaired strain on MDBK cells. Diameters of 80 plaques of each virus strain were measured using Adobe Acrobat XI Pro. ****, P < 0.0001. Scale bars = 500 μm. (F) Immunoprecipitation (IP) and Western blot assays of Roadblock-1 and pUL21 CΔ300. The pUL21 CΔ300 was not able to coprecipitate with Roadblock-1.

The carboxyl terminus of pUL21 promotes PRV retrograde axonal transport in vitro and in vivo.

Next, the efficiency of retrograde axonal transport of PRV UL21 CΔ300 in neurons was determined. A total of 2 × 10⁶ PFU of virus was inoculated into each axonal compartment. The entire contents of the axonal and somal compartments were harvested separately at 0, 12, 24, 36, and 48 hpi. Virus yields were determined by titration on MDBK cells. In the axonal compartments, the resulting titers of the three viruses were similar (Fig. 8A). In the somal compartments, the WT and repaired strains were detected as early as 12 hpi, but PRV UL21 CΔ300 was not detected until 24 hpi (Fig. 8A). Compared with the growth curves of WT and repaired strains, the growth curve of PRV UL21 CΔ300 in the somal compartments shows an approximately 24-h delay (Fig. 8A). These data suggest that the carboxyl terminus (aa 301 to 533) of pUL21 promotes PRV retrograde axonal transport in explanted DRG neurons. The carboxyl terminus of pUL21 is not involved in anterograde transport (Fig. 8B), since pUL21 is irrelevant to anterograde axonal transport (Fig. 4E).

FIG 8.

The efficiency of retrograde and anterograde axonal transport of PRV UL21 CΔ300 in explanted neurons. (A) Quantification of the efficiency of retrograde axonal spread. Virus was inoculated into axonal compartments. Virus yields of the axonal and somal compartments were determined. (B) Quantification of the efficiency of anterograde axonal transport. Virus was inoculated into somal compartments. Virus yields of the axonal and somal compartments at 0, 12, and 24 hpi were determined.

Then, we investigated the contribution of the carboxyl terminus (aa 301 to 533) of pUL21 to retrograde axonal transport and neuroinvasion by PRV in mice. The animals were infected by intraocular injection or intranasal instillation in the same manner as before (Fig. 5A). In the absence of the carboxyl terminus of pUL21, the mean survival time of the intraocularly infected mice was significantly prolonged (Fig. 9A). In contrast, the deletion of the carboxyl terminus of pUL21 had a slight influence (not statistically significant) on the mean survival time of intranasally infected animals (Fig. 9B). For the intraocular injection groups, mouse SCG were dissociated immediately after death, and their viral loads were determined. Virus loads of the SCG from mice infected with WT and repaired PRV were more than 10-fold higher than those of PRV UL21 CΔ300 (Fig. 9C). These data suggest that the carboxyl terminus (aa 301 to 533) of pUL21 enhances PRV retrograde axonal transport and neuroinvasion in mice.

FIG 9.

PRV UL21 CΔ300 virulence and neuroinvasion in mice. (A) Survival time of mice (n = 6) that were intraocularly infected. Each dot represents one mouse. Means are shown as black lines. **, P < 0.01. (B) Survival time of mice (n = 6) that were intranasally inoculated. (C) Viral loads in the SCG of PRV-infected mice in intraocular injection group (n = 6). ****, P < 0.0001.

DISCUSSION

In this study, PRV pUL21 exhibited a tight association with capsids, similar to that of the inner tegument protein VP1/2; thus, pUL21 is also an inner tegument protein of PRV (Fig. 1). This result is consistent with the previous findings for pUL21 in HSV-1 (12).

As an inner tegument protein, pUL21 has a potential to interact with microtubule motors during retrograde transports to the nucleus (11, 12). Our findings suggest that pUL21 interacts with the cytoplasmic dynein light chain Roadblock-1 based on the result of BiFC assays (Fig. 2B); this interaction was confirmed by co-IP assays (Fig. 3A). The subcellular localizations of pUL21 and Roadblock-1 have been reported independently (16, 18, 20, 30, 31, 40) but not together in the same cells, so we performed immunofluorescence assays to investigate their potential colocalization. The results revealed that pUL21 and Roadblock-1 colocalized in the perinuclear region (Fig. 3C).

A pUL21 deletion mutant of the PRV Kaplan strain has been already characterized by slightly reduced viral titers and plaque sizes (60). In our study, PRV ΔUL21 propagates similarly to the WT and repaired strains with normal kinetics, but it produces smaller plaques than do the WT and repaired strains (data not shown). Consistent with prior reports, these results indicate that pUL21 is dispensable for PRV replication in epithelial cells (18, 21–23, 25, 60). The formation of smaller plaques suggests that the pUL21 deletion impairs virus spread in epithelial cells. Glycoprotein gE plays an important role in the intercellular spread and virus-induced cell fusion of herpesviruses, and tegument proteins pUL11, pUL16, and pUL21 cooperate on the cytoplasmic tail of gE, contributing to the full function of gE (51). Thus, the deletion of pUL21 would impair the function of gE and affect the spread of virus. The formation of smaller plaques by PRV ΔUL21 in epithelial cells is probably due to gE dysfunction.

The interaction domain of pUL21 was mapped for further research. The carboxyl terminus (aa 301 to 533) of pUL21 was required for its interaction with Roadblock-1 (Fig. 6). Recent studies on the pUL21 of HSV-1 determined the crystal structures of its N-terminal (aa 1 to 216) and C-terminal (aa 281 to 535) domains (17, 29). These data may aid in exploring the functions of pUL21. Dynein light chains are capable of functioning independently of the dynein complex (32–35). Therefore, the interaction between pUL21 and Roadblock-1 found in our study is not necessarily indicative of a motor recruitment mechanism. For example, although initial research on HSV-1 showed that VP26 interacts with dynein light chains RP3 and Tctex1 and plays a role in retrograde cellular transport (13), this finding was contradicted by later work (61, 62). The potential interaction between pUL21 and the dynein complex will be examined in future studies.

The efficiency of retrograde and anterograde axonal transport of PRV strains in explanted neurons using a microfluidic chamber system was determined (Fig. 4 and 8) (53). Our data show that the deletion of the carboxyl terminus (aa 301 to 533) or the full length of pUL21 affects retrograde axonal transport but has no effect on the anterograde spread of PRV in explanted DRG neurons (Fig. 4 and 8). These results demonstrate that pUL21 is involved in retrograde axonal transport of capsids in neurons. Previous work reported that PRV pUS9 is required for the anterograde spread of PRV infection (54, 63) and that pUS9 promotes anterograde transport in axonal sorting but not in fast axonal transport of particles (64). Our results suggest that, unlike pUS9, pUL21 is dispensable for PRV anterograde axonal transport (Fig. 4E). However, the deletion of the carboxyl terminus or the full length of pUL21 did not completely block the retrograde transport of PRV in neurons. Specifically, the deletion of the carboxyl terminus of pUL21 caused a prominent delay in this process (Fig. 8A). A prior study on the retrograde transport of capsids in neurons showed that PRV lacking pUL21 was transported more slowly than with parental strain in dorsal root sensory neurons. However, the amount of transport was not grossly affected by the absence of pUL21 (62). There are several possible causes for the delay of retrograde transport we observed in PRV ΔUL21 and PRV UL21 CΔ300, as follows: (i) the dynein motor recruitment procedure was affected, (ii) the transport velocity was decelerated in neurons, (iii) the transport distance was shortened, or (iv) the replication in neurons was affected. Based on the result that the carboxyl terminus of pUL21 is dispensable for virus replication in epithelial cells and neurons (Fig. 7 and 8), we speculate that the delay of retrograde transport in PRV ΔUL21 and PRV UL21 CΔ300 is due to defects in retrograde transport.

It was previously reported that the pUL21 of alphaherpesvirus is involved in neuroinvasion and neurovirulence (25, 26, 28). Here, we used an intraocular injection model for studying retrograde transport (Fig. 5A) (6, 59). The mean survival time of mice infected by PRV ΔUL21 or PRV UL21 CΔ300 was significantly longer than that of mice infected by the WT or repaired PRV strains (Fig. 5 and 9). However, the deletion of pUL21 had a minor influence on the mean survival time of intranasally PRV-infected mice (Fig. 5C). These findings indicate that the virulence of PRV ΔUL21 and PRV UL21 CΔ300 is slightly attenuated in mice, even though the retrograde transport of these viruses is dramatically affected. In addition, virus loads of PRV ΔUL21 and PRV UL21 CΔ300 in SCG were significantly lower than those of the WT and repaired PRV strains (Fig. 5E and 9C). Since the SCG consists of only first-order neurons projecting axons to the iris (6), we suppose that the lower titers of PRV ΔUL21 and PRV UL21 CΔ300 in SCG is due to defects in retrograde transport rather than defects in replication. This finding suggests that the carboxyl terminus of pUL21 contributes to PRV retrograde axonal transport and neuroinvasion in mice.

pUL21 is composed of the N-terminal (aa 1 to 216) domain and the C-terminal (aa 281 to 535) domain connected by a flexible linker in HSV-1 (17, 29). Compared with the carboxyl terminus of HSV-1 pUL21, the C terminus of PRV pUL21 contains a 17-amino-acid insertion between residues 385 and 386 (29). Three amino acid mutations, H37R, E355D, and V375A, within pUL21 contribute to the attenuated neurovirulence of PRV Bartha (26, 27). When these mutations were rescued, defects in retrograde and transneuronal infection were also repaired (28). Among the three mutations, E355D and V375A locate within the carboxyl domain of pUL21. The V375A mutation might destabilize the hydrophobic core and disrupt the binding of a potential functional partner (29). In our study, the carboxyl terminus (aa 301 to 533) of PRV pUL21, covering E355 and V375, was deleted. Hence, defects in retrograde transport and neuroinvasion of PRV UL21 CΔ300 were probably due to the failure of interaction between pUL21 CΔ300 and Roadblock-1. Future work should determine the minimal domains necessary for interaction of pUL21 and Roadblock-1.

Previous studies on the inner tegument proteins VP1/2 (pUL36) and pUL37 revealed that both proteins are critical for alphaherpesvirus retrograde axonal transport and neuroinvasion (6, 65, 66). Our findings in this study provide evidence that pUL21, as an inner tegument protein like VP1/2 and pUL37, also has a potential role in PRV retrograde axonal transport and neuroinvasion.

MATERIALS AND METHODS

Cells, viruses, and antibodies.

HeLa (Henrietta Lacks) cells (ATCC CCL-2; China Center for Type Culture Collection, Wuhan, China) were grown in Dulbecco’s modified Eagle medium (DMEM, catalog no. 12800-017; Gibco) supplemented with 10% fetal bovine serum (FBS, catalog no. 10099-141; Gibco) and 100 U/ml of penicillin and streptomycin (catalog no. GNM-15140; Genom). HEK293 (human embryonic kidney) cells (ATCC CRL-3249; China Center for Type Culture Collection) were maintained in RPMI 1640 medium (catalog no. SH30809.01B; HyClone) supplemented with 10% FBS and 100 U/ml of penicillin and streptomycin. PK-15 (pig kidney) cells (ATCC CCL-33; China Center for Type Culture Collection) were grown in DMEM supplemented with 10% newborn calf serum (NBCS, catalog no. 16010-142; Gibco) and 100 U/ml of penicillin and streptomycin. The NBCS concentration in the medium was reduced to 2% during infection. MDBK (Madin-Darby bovine kidney) cells (ATCC CCL-22; China Center for Type Culture Collection) were grown in DMEM supplemented with 5% NBCS and 100 U/ml of penicillin and streptomycin. The NBCS level in the medium was reduced to 2% during infection. The WT PRV Ea strain was propagated in PK-15 cells, as described previously (67). PRV EGFP-VP1/2, which has an enhanced green fluorescent protein (EGFP) fused to the N terminus of VP1/2, was constructed by homologous recombination. PRV ΔUL21 is a full-length UL21 open reading frame-deleted virus, in which the UL21 gene is replaced by an EGFP expression cassette. PRV ΔUL21 R is a repaired strain, in which the EGFP expression cassette was replaced by the UL21 gene. PRV UL21 CΔ300 is a pUL21-truncated virus in which the carboxyl terminus (aa 301 to 533) was deleted. PRV UL21 CΔ300 R is a repaired strain in which the full-length pUL21 expression is restored. The full-length PRV Ea genome was cloned as bacterial artificial chromosome (BAC) (K. Zheng, X. R. Song, J. Liu, R. Z. Zhang, H. C. Chen, and Z. F. Liu, unpublished data). PRV UL21 CΔ300 and the repaired strain were constructed by a two-step Red-mediated recombination (68). The anti-α-tubulin monoclonal antibody was purchased from Sigma-Aldrich (catalog no. T5168). The anti-6×His monoclonal antibody was purchased from Roche (catalog no. 11922416001). Anti-EGFP (catalog no. 66002-1-Ig) and anti-hemagglutinin (catalog no. 66006-1-Ig) monoclonal antibodies were purchased from Proteintech (Wuhan, China). The anti-gB monoclonal antibody was purchased from Keqian Animal Biological Products Co. Ltd. (Wuhan, China). The anti-LC8-1 (catalog no. A14496) and anti-Tctex-1 (catalog no. A4150) polyclonal antibodies were purchased from ABclonal Biotechnology Co. Ltd. (Wuhan, China). Polyclonal antibodies against peptides PRV VP5 (CSQRHSYADRLYNGQYN), VP26 (FDPNNPRTITAQTLEGC), pUL21 (FDGIPGVRPPLSGETRC), VP13/14 (LQDVLVGEVRRPDFC), VP16 (CHLIPRDALNRMFEM), and Roadblock-1 (TEGIPIKSTMDNPTC) were raised in New Zealand rabbits (Experimental Animal Center, Huazhong Agricultural University, Wuhan, China).

Plasmids.

Plasmids pLumin, pFos-myc-LC151, and pJun-HA-KN151 are generous gifts from Zhihong Zhang (Britton Chance Center for Biomedical Photonics, Huazhong University of Science and Technology) (44). Plasmids pMD18-T (catalog no. D101A) and pcDNA3.1(+) (catalog no. V709-20) were purchased from TaKaRa (Dalian, China) and Invitrogen, respectively.

Immunoprecipitation assay and Western blot analysis.

For the immunoprecipitation (IP) assay, HEK293 cells were transfected with Lipofectamine 2000 (catalog no. 11668-019; Invitrogen) and lysed at 24 h posttransfection. PK-15 cells were infected with the corresponding PRV strains and lysed. The resulting cell lysates were incubated with 0.5 μg of the corresponding antibody for IP overnight at 4°C, followed by incubation with 40 μl of protein A+G agarose beads (catalog no. P2012; Beyotime, Shanghai, China) for 3 h at 4°C. The protein A+G agarose beads were washed for five times with lysis buffer and resuspended in 1× sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer. These samples were boiled for 10 min and then subjected to Western blot analysis. Western blot analysis was performed as described previously (69, 70).

Purification of extracellular virions and KCl treatment assay.

Purification of extracellular virions was performed as previously described with minor modifications (7, 71). PK-15 cells were grown in 150-cm² flasks until 90% confluence. Cells were then infected with WT PRV or PRV EGFP-VP1/2 at an MOI of 10. At 6 hpi, the cells were washed twice with serum-free DMEM and then incubated with serum-free DMEM. The extracellular medium was collected at 24 hpi and clarified by centrifugation at 600 × g for 10 min at 4°C. The supernatant was treated with 50 μg/ml DNase I (catalog no. Z358A; Promega) for 30 min at 4°C and then filtered through a 0.45-μm filter. Extracellular virions were pelleted by centrifugation at 20,000 × g for 45 min at 4°C. The viral pellet was resuspended in MNT buffer (30 mM morpholineethanesulfonic acid [MES], 10 mM NaCl, 20 mM Tris-HCl [pH 7.4]), layered over a 10% Ficoll 400 cushion (catalog no. F4375; Sigma-Aldrich), and centrifuged at 110,000 × g for 2 h at 4°C in an SW41 Ti rotor (Beckman). The resulting pellet was washed by resuspension in MNT buffer and concentrated by centrifugation at 110,000 × g for 30 min at 4°C. Purified virions were resuspended in MNT buffer and stored at −80°C. KCl treatment of extracellular virions was performed as previously described, with minor modifications (12). A total of 5 × 10⁸ PFU of extracellular virions of WT PRV or PRV EGFP-VP1/2 were left untreated or incubated with 0.5 mg/ml trypsin (catalog no. 27250018; Gibco) at 37°C for 1 h. The samples were then incubated with 100 mM phenylmethylsulfonyl fluoride (PMSF, catalog no. 0754; Amresco) and a protease inhibitor cocktail (catalog no. P8340; Sigma-Aldrich) for 5 min on ice. Virions were lysed for 30 min on ice by adding an equal volume of 2× lysis buffer (2% Triton X-100, 20 mM MES, 20 mM dithiothreitol [DTT], 30 mM Tris [pH 7.4], protease inhibitor cocktail, 320 μg/ml PMSF) containing KCl to achieve final concentrations of 0.1 M, 0.5 M, or 1 M. The samples were then layered on top of a 20% (wt/vol) sucrose cushion (30 mM MES, 10 mM DTT, 20 mM Tris [pH 7.4], protease inhibitor cocktail) with corresponding concentrations of KCl and centrifuged at 110,000 × g for 15 min at 4°C. The resulting pellet was resuspended in BRB80 [80 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES; pH 6.8), 12 mM MgCl₂, 1 mM EGTA] with 10 mM DTT, protease inhibitor cocktail, 0.1 U/μl DNase I, and 100 μg/ml RNase A (catalog no. A7973; Promega). Following incubations at 37°C for 30 min and 4°C overnight, the capsids were centrifuged at 110,000 × g for 8 min at 4°C and resuspended in BRB80. The capsids were then incubated at 4°C overnight for dissolving and subsequently stored at −80°C. The normalized relative protein amounts were quantified by densitometry scanning (ImageJ2x).

Indirect immunofluorescence assay.

Cells were transfected or infected and fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 20 min and permeabilized with 0.5% Triton X-100 in PBS for 10 min. Then, cells were blocked with 5% bovine serum albumin (BSA) in PBS and incubated with primary antibody at 37°C for 2 h. After incubation with a Texas Red/fluorescein isothiocyanate (FITC)-conjugated secondary antibody at 37°C for 1 h, cells were incubated with 4′,6-diamidino-phenylindole (DAPI; Biosharp, Anhui, China) in PBS for 10 min at room temperature. Sample cells were observed under an inverted fluorescence microscope (Nikon eclipse Ti-U microscope/Olympus FluoView FV1000 confocal microscope).

Growth curve determination and plaque morphology comparison.

One-step growth curve determination and plaque morphology comparison were performed as described previously (69, 70). Multistep growth curve determination was performed as previously described, with minor modifications (72–74). Confluent PK-15 cells were infected with PRV at an MOI of 0.01. After 1.5 h of attachment at 4°C, residual input virus was removed. The cultures were washed three times with phosphate-buffered saline (PBS), and then fresh medium was added to each flask before further incubation at 37°C. At 0, 12, 24, 48, and 72 hpi, medium and cells were collected and frozen separately. The virus yields were determined by titration on MDBK cells. Each data point represents the average of the results from three samples obtained from separate infections.

Primary neuronal culture and the microfluidic chamber system.

Microfluidic chambers (catalog no. SND450; Xona) were assembled with coverslips that had been pretreated with poly-l-lysine (catalog no. P1399; Sigma-Aldrich). Dorsal root ganglia (DRG) neurons were prepared as described previously (75, 76). Neurons were cultured in the microfluidic chambers for 10 days before infection. PRV was inoculated into either the axonal compartment or somal compartment. The entire contents of the axonal and somal compartments were harvested separately. Virus yields were determined by titration on MDBK cells. Four chambers were used for each virus strain.

Animal experiment.

Twenty-four 8-week-old male BALB/c mice (Experimental Animal Center, Huazhong Agricultural University) were infected intraocularly with 1 × 10⁵ PFU of WT PRV, PRV ΔUL21, or the repaired strain (eight mice per group). Eighteen 8-week-old male BALB/c mice were infected intranasally with 1 × 10⁵ PFU of WT PRV, PRV ΔUL21, or the repaired strain (six mice per group). The survival time of each mouse was recorded. SCG were isolated immediately after death. The isolated SCG were homogenized, and virus loads were determined by titration on MDBK cells. Animal experiments on mice infected with WT PRV, PRV UL21 CΔ300, and the repaired strain were conducted with similar procedures.

Ethics statement.

Our study was approved by the Institutional Animal Care and Use Committee (IACUC) of Huazhong Agricultural University (license no. HZAUMO-2016-013). Animal-related experimental procedures were conducted according to instructions by the Regulations for the Administration of Affairs Concerning Experimental Animals of the People’s Republic of China.

Statistical analysis.

Statistical significance was determined by unpaired two-tailed Student’s t test for comparisons (Microsoft Office Excel and GraphPad Prism 7).