Glycoprotein gIII of pseudorabies virus is multifunctional.

One of the major glycoproteins of pseudorabies virus, gIII, is nonessential for growth in cell culture. Mutants defective in gIII, however, consistently yield lower titers of infectious virus (3- to 20-fold) than does wild-type virus. The interactions of gIII- mutants with their host cells were compared with those of wild-type virus in an attempt to uncover the functions of gIII. We show that gIII plays a major role in the stable adsorption of the virus to its host cell; in the absence of gIII, the rate of adsorption is reduced and adsorption is easily reversed by washing. Thus, adsorption of pseudorabies virus can be said to occur in at least the following two ways: (i) a gIII-mediated rapid adsorption or (ii) a slower and more labile adsorption that is independent of gIII. After virions have been complexed with monoclonal antibodies against gIII (but not some monoclonal antibodies against other glycoproteins), both modes of adsorption were inhibited. Glycoprotein gIII affects virus stability and virus release, as well as adsorption. The effect on virus release is marked when the virus is defective in additional functions. Thus, although we found no obvious difference in the release of virus from gIII- or wild-type virus-infected rabbit kidney cells, release of a gIII-/gI- double mutant from the cells occurred less readily than did release of a gI- mutant. The gIII-/gI- and gIII- mutants, however, adsorbed to cells at a similar rate, indicating that the effects of gIII on adsorption and virus release constitute separate functions. The Bartha vaccine strain of pseudorabies virus has a defective gIII gene and is released poorly from rabbit kidney cells. After the resident Bartha gIII gene was replaced by the gIII gene of wild-type virus, virus release was enhanced considerably. Since inactivation of gIII in wild-type pseudorabies virus did not significantly affect virus release, the Bartha strain must be defective in another function which, in conjunction with gIII, significantly affects virus release. These results indicate again that gIII affects virus release in conjunction with other functions. Also, although the Bartha strain was functionally defective in virus release, it adsorbed to cells as well as wild-type virus did, showing that the effects of gIII on virus adsorption and release constitute separate functions. We conclude that gIII is a multifunctional glycoprotein.     

Role of glycoprotein gIII of pseudorabies virus in virulence.

Deletion mutants of pseudorabies virus unable to express glycoprotein gIII, gI, or gp63 or double and triple mutants defective in these glycoproteins were constructed, and their virulence for day-old chickens inoculated intracerebrally was determined. Mutants of wild-type pseudorabies virus defective in glycoprotein gIII, gI, or gp63 were only slightly less virulent (at most, fivefold) for chickens than was the wild-type virus. However, mutants defective in both gIII and gI or gIII and gp63 were avirulent for chickens, despite their ability to grow in cell culture in vitro to about the same extent as mutants defective in gIII alone (which were virulent). These results show that gIII plays a role in virulence and does so in conjunction with gI or gp63. The effect of gIII on virulence was also shown when the resident gIII gene of variants of the Bartha vaccine strain (which codes for gIIIB) was replaced with a gIII gene derived from a virulent wild-type strain (which codes for gIIIKa); gIIIKa significantly enhanced the virulence of a variant of the Bartha strain to which partial virulence had been previously restored by marker rescue. Our results show that viral functions that play a role in the virulence of the virus (as measured by intracerebral inoculation of chickens) may act synergistically to affect the expression of virulence and that the ability of the virus to grow in cell culture is not necessarily correlated with virulence.     

Herpes simplex virus glycoproteins E and I facilitate cell-to-cell spread in vivo and across junctions of cultured cells.

Herpes simplex virus (HSV) glycoproteins E and I (gE and gI) can act as a receptor for the Fc domain of immunoglobulin G (IgG). To examine the role of HSV IgG Fc receptor in viral pathogenesis, rabbits and mice were infected by the corneal route with HSV gE- or gI- mutants. Wild-type HSV-1 produced large dendritic lesions in the corneal epithelium and subsequent stromal disease leading to viral encephalitis, whereas gE- and gI- mutant viruses produced microscopic punctate or small dendritic lesions in the epithelium and no corneal disease or encephalitis. These differences were not related to the ability of the gE-gI oligomer to bind IgG because the differences were observed before the appearance of anti-HSV IgG and in mice, in which IgG binds to the Fc receptor poorly or not at all. Mutant viruses produced small plaques on monolayers of normal human fibroblasts and epithelial cells. Replication of gE- and gI- mutant viruses in human fibroblasts were normal, and the rates of entry of mutant and wild-type viruses into fibroblasts were similar; however, spread of gE- and gI- mutant viruses from cell to cell was significantly slower than that of wild-type HSV-1. In experiments in which fibroblast monolayers were infected with low multiplicities of virus and multiple rounds of infection occurred, the presence of neutralizing antibodies in the culture medium caused the yields of mutant viruses to drop dramatically, whereas there was a lesser effect on the production of wild-type HSV. It appears that cell-to-cell transmission of wild-type HSV-1 occurs by at least two mechanisms: (i) release of virus from cells and entry of extracellular virus into a neighboring cell and (ii) transfer of virus across cell junctions in a manner resistant to neutralizing antibodies. Our results suggest that gE- and gI- mutants are defective in the latter mechanism of spread, suggesting the possibility that the gE-gI complex facilitates virus transfer across cell junctions, a mode of spread which may predominate in some tissues. It is ironic that the gE-gI complex, usually considered an IgG Fc receptor, may, through its ability to mediate cell-to-cell spread, actually protect HSV from IgG in a manner different than previously thought.     

Early interactions of pseudorabies virus with host cells: functions of glycoprotein gIII.

Adsorption of mutants of pseudorabies virus (PrV) lacking glycoprotein gIII is slower and less efficient than is that of wild-type virus (C. Schreurs, T. C. Mettenleiter, F. Zuckermann, N. Snugg, and T. Ben-Porat, J. Virol. 62:2251-2257, 1988). To ascertain the functions of gIII in the early interactions of PrV with its host cells, we compared the effect on wild-type virus and gIII- mutants of antibodies specific for various PrV proteins. Although adsorption of wild-type virus was inhibited by polyvalent antisera against PrV as well as by sera against gIII and gp50 (but not sera against gII), adsorption of the gIII- mutants was not inhibited by any of these antisera. These results suggest that, in contrast to adsorption of wild-type PrV, the initial interactions of the gIII- mutants with their host cells are not mediated by specific viral proteins. Furthermore, competition experiments showed that wild-type Prv and the gIII- mutants do not compete for attachment to the same cellular components. These findings show that the initial attachment of PrV to its host cells can occur by a least two different modes--one mediated by glycoprotein gIII and the other unspecific. gIII- mutants not only did not adsorb as readily to cells as did wild-type virus but also did not penetrate cells as rapidly as did wild-type virus after having adsorbed. Antibodies against gIII did not inhibit the penetration of adsorbed virus (wild type or gIII-), whereas antibodies against gII and gp50 did. It is unlikely, therefore, that gIII functions directly in virus penetration. Our results support the premises that efficient adsorption of PrV to host cell components is mediated either directly or indirectly by gIII (or a complex of viral proteins for which the presence of gIII is functionally essential) and that this pathway of adsorption promotes the interactions of other viral membrane proteins with the appropriate cellular proteins, leading to the rapid penetration of the virus into the cells. The slower penetration of the gIII- mutants than of wild-type PrV appears to be related to the slower and less efficient alternative mode of adsorption of PrV that occurs in the absence of glycoprotein gIII.     

Bovine herpesvirus 1 attachment to permissive cells is mediated by its major glycoproteins gI, gIII, and gIV.

A bovine herpesvirus 1 (BHV-1) gIII deletion mutant (gIII-) was produced by means of recombinant DNA that retained the ability to replicate in cell culture. However, the gIII- mutant was functionally defective, showing impaired attachment to permissive cells, a delay in virus replication, and reduced extracellular virus production. The attachment defect exhibited by the gIII- mutant is an indication of the role played by gIII in the normal infection process. This was shown by dramatically decreased binding of radiolabelled gIII- virus to permissive cells and a slower adsorption rate, as measured by plaque formation, than the wild-type (wt) virus. Furthermore, treatment of the gIII- virus with neomycin increased virus adsorption and plaque formation by severalfold, whereas neomycin treatment had no effect on the wt virus. This observation showed that the gIII- mutant was strictly defective in adsorption but fully competent to produce productive infections once induced to attach. The gIII- mutant showed greater sensitivities than did the wt virus to anti-gI and anti-gIV antibody-mediated neutralization. Analyses with panels of monoclonal antibodies to gI and gIV revealed that the epitopes gI-IV and gIV-III were the main targets for enhanced neutralization. This provided evidence that gI and gIV may also participate in virus attachment. Finally, when affinity-purified gI, gIII, and gIV were tested for their ability to inhibit virus adsorption, gIII had the most pronounced inhibitory effect, followed by gI and then gIV. gIII was able to completely inhibit wt virus adsorption, and at a high concentration, it also partially inhibited the gIII- mutant. gI and gIV inhibited wt and gIII- mutant adsorption to a comparable extent. Our results collectively indicate that gIII plays a predominant role in virus attachment, but gI and gIV also contribute to this process. In addition, a potential cooperative mechanism for virus attachment with these three proteins is presented.     

Interaction of glycoprotein gIII with a cellular heparinlike substance mediates adsorption of pseudorabies virus.

Glycoprotein gIII is one of the major envelope glycoproteins of pseudorabies virus (PrV) (Suid herpesvirus 1). Although it is dispensable for viral growth, it has been shown to play a prominent role in the attachment of the virus to target cells, since gIII- deletion mutants are severely impaired in adsorption (C. Schreurs, T. C. Mettenleiter, F. Zuckermann, N. Sugg, and T. Ben-Porat, J. Virol. 62:2251-2257, 1988). We show here that during the process of adsorption of PrV, the viral glycoprotein gIII interacts with a cellular heparinlike receptor. This conclusion is based on the following findings. (i) Heparin inhibits plaque formation of PrV by preventing the adsorption of wild-type virions to target cells. However, heparin does not interfere with the plaque formation of PrV mutants that lack glycoprotein gIII. (ii) Wild-type virions readily adsorb to matrix-bound heparin, whereas gIII- mutants do not. (iii) Pretreatment of cells with heparinase reduces considerably the ability of wild-type PrV to adsorb to these cells and to form plaques but does not negatively affect gIII- mutants. (iv) Glycoprotein gIII binds to heparin and appears to do so in conjunction with glycoprotein gII. Although heparin significantly reduces the adsorption of wild-type virus to all cell types tested, quantitative differences in the degree of inhibition of virus adsorption by heparin to different cell types were observed. Different cell types also retain their abilities to adsorb wild-type PrV to a different extent after treatment with heparinase and differ somewhat in their relative abilities to adsorb gIII- mutants. Our results show that while the primary pathway of adsorption of wild-type PrV to cells occurs via the interaction of viral glycoprotein gIII with a cellular heparinlike receptor, an alternative mode of adsorption, which is not dependent on either component, exists. Furthermore, the relative abilities of different cell types to adsorb PrV by the gIII-dependent or the alternative mode vary to some extent.     

Complex between glycoproteins gI and gp63 of pseudorabies virus: its effect on virus replication.

To ascertain the biological functions of different glycoproteins that are nonessential for pseudorabies virus growth in vitro, we have constructed mutants defective in one (or a combination) of these glycoproteins and have examined various aspects of their role in the infective process. We made the following two observations. (i) Glycoproteins gI and gp63 are noncovalently complexed to each other. They are coprecipitated by antisera against either one of these glycoproteins but do not share antigenic determinants: monoclonal antibodies against gp63 do not immunoprecipitate gI from extracts of gp63- mutant-infected cells, and monoclonal antibodies against gI do not immunoprecipitate gp63 from extracts of gI- mutant-infected cells. (ii) Mutants unable to synthesize either gI or gp63 have some common biological characteristics; they have a growth advantage in primary chicken embryo fibroblasts. Furthermore, we have shown previously that in conjunction with glycoprotein gIII, gI and gp63 are necessary for the expression of virulence (T. C. Mettenleiter, C. Schreurs, F. Zuckermann, T. Ben-Porat, and A. S. Kaplan, J. Virol. 62, 2712-2717, 1988). These results show that the functional entity affecting virus replication in chicken embryo fibroblasts, as well as affecting virulence, is the complex between gI and gp63. The gI-gp63 complex of pseudorabies virus does not appear to have Fc receptor activity as does its homolog, the gI-gE complex of herpes simplex virus.     

The extracellular domain of herpes simplex virus gE is indispensable for efficient cell-to-cell spread: evidence for gE/gI receptors

Herpes simplex virus (HSV) spreads rapidly and efficiently within epithelial and neuronal tissues. The HSV glycoprotein heterodimer gE/gI plays a critical role in promoting cell-to-cell spread but does not obviously function during entry of extracellular virus into cells. Thus, gE/gI is an important molecular handle on the poorly understood process of cell-to-cell spread. There was previous evidence that the large extracellular (ET) domains of gE/gI might be important in cell-to-cell spread. First, gE/gI extensively accumulates at cell junctions, consistent with being tethered there. Second, expression of gE/gI in trans interfered with HSV spread between epithelial cells. To directly test whether the gE ET domain was necessary for gE/gI to promote virus spread, a panel of gE mutants with small insertions in the ET domain was constructed. Cell-to-cell spread was reduced when insertions were made within either of two regions, residues 256 to 291 or 348 to 380. There was a strong correlation between loss of cell-to-cell spread function and binding of immunoglobulin. gE ET domain mutants 277, 291, and 348 bound gI, produced mature forms of gE that reached the cell surface, and were incorporated into virions yet produced plaques similar to gE null mutants. Moreover, all three mutants were highly restricted in spread within the corneal epithelium, in the case of mutant 277 to only 4 to 6% of the number of cells compared with wild-type HSV. Therefore, the ET domain of gE is indispensable for efficient cell-to-cell spread. These observations are consistent with our working hypothesis that gE/gI can bind extracellular ligands, so-called gE/gI receptors that are concentrated at epithelial cell junctions. This fits with similarities in structure and function of gE/gI and gD, which is a receptor binding protein.     

Characterization and mapping of a nonessential pseudorabies virus glycoprotein.

Antigenic variants of pseudorabies virus (PRV) containing mutations in a viral glycoprotein with a molecular weight of 82,000 (gIII) were isolated by selecting for resistance to a complement-dependent neutralizing monoclonal antibody (MCA82-2) directed against gIII. These mutants were completely resistant to neutralization with MCA82-2 in the presence of complement. Two mutants selected for further studies either did not express gIII or expressed an improperly processed form of the glycoprotein. The mutations were also associated with an altered plaque morphology (syncytium formation). The gIII gene was mapped by marker rescue of a gIII- mutant with cloned restriction enzyme fragments to the long unique region of the PRV genome between 0.376 and 0.383 map units. This corresponds to the map location of a glycoprotein described by Robbins et al. (J. Mol. Appl. Gen. 2:485-496, 1984). Since gIII is nonessential for viral replication in cell culture and has several other characteristics in common with the herpes simplex virus glycoprotein gC, gIII may represent the PRV equivalent to herpes simplex virus gC.     

Glycoproteins E and I facilitate neuron-to-neuron spread of herpes simplex virus.

Two herpes simplex virus (HSV) glycoproteins E and I (gE and gI) form a heterooligomer which acts as an Fc receptor and also facilitates cell-to-cell spread of virus in epithelial tissues and between certain cultured cells. By contrast, gE-gI is not required for infection of cells by extracellular virus. HSV glycoproteins gD and gJ are encoded by neighboring genes, and gD is required for both virus entry into cells and cell-to-cell spread, whereas gJ has not been shown to influence these processes. Since HSV infects neurons and apparently spreads across synaptic junctions, it was of interest to determine whether gD, gE, gI and gJ are also important for interneuronal transfer of virus. We tested the roles of these glycoproteins in neuron-to-neuron transmission of HSV type 1 (HSV-1) by injecting mutant viruses unable to express these glycoproteins into the vitreous body of the rat eye. The spread of virus infection was measured in neuron-rich layers of the retina and in the major retinorecipient areas of the brain. Wild-type HSV-1 and a gJ- mutant spread rapidly between synaptically linked retinal neurons and efficiently infected major retinorecipient areas of the brain. gD mutants, derived from complementing cells, infected only a few neurons and did not spread in the retina or brain. Mutants unable to express gE or gI were markedly restricted in their ability to spread within the retina, produced 10-fold-less virus in the retina, and spread inefficiently to the brain. Furthermore, when compared with wild-type HSV-1, gE- and gI- mutants spread inefficiently from cell to cell in cultures of neurons derived from rat trigeminal ganglia. Together, our results suggest that the gE-gI heterooligomer is required for efficient neuron-to-neuron transmission through synaptically linked neuronal pathways.     

Glycoprotein D-negative pseudorabies virus can spread transneuronally via direct neuron-to-neuron transmission in its natural host, the pig, but not after additional inactivation of gE or gI.

Envelope glycoprotein D (gD) is essential for entry of pseudorabies virus (PRV) into cells but is not required for the subsequent steps in virus replication. Phenotypically complemented gD mutants can infect cells and can spread, both in vitro and in mice, by direct cell-to-cell transmission. Progeny virions released by infected cells are noninfectious because they lack gD. The aim of this study was to determine the role of gD in the neuropathogenicity of PRV in its natural host, the pig. We investigated whether gD-negative PRV can spread transneuronally via synaptically linked neurons of the olfactory and trigeminal routes. High doses of a phenotypically complemented gD mutant and gD mutants that are unable to express either gI or gI plus gE were inoculated intranasally in 3- to 5-week-old pigs. Compared with the wild-type virus, the virulence of the gD mutant was reduced. However, pigs inoculated with the gD mutant still developed fever and respiratory signs. Additional inactivation of either gI or gI plus gE further decreased virulence for pigs. Immunohistochemical examination of infected pigs showed that a PRV gD mutant could replicate and spread transneuronally into the central nervous system (CNS). Compared with the wild-type virus, the gD mutant had infected fewer neurons of the CNS on day 2. Nevertheless, on day 3, the gD-negative PRV had infected more neurons and viral antigens were present in second- and third-order neurons in the olfactory bulb, brain stem, and medulla oblongata. In contrast, gD mutants which are unable to express either gI or gI plus gE infected a limited number of first-order neurons in the olfactory epithelium and in the trigeminal ganglion and did not spread transneuronally or infect the CNS. Thus, transsynaptic spread of PRV in pigs can occur independently of gD. Possible mechanisms of transsynaptic transport of PRV are discussed.     

Herpes simplex virus gE/gI expressed in epithelial cells interferes with cell-to-cell spread

The herpes simplex virus (HSV) glycoprotein heterodimer gE/gI plays an important role in virus cell-to-cell spread in epithelial and neuronal tissues. In an analogous fashion, gE/gI promotes virus spread between certain cell types in culture, e.g., keratinocytes and epithelial cells, cells that are polarized or that form extensive cell junctions. One mechanism by which gE/gI facilitates cell-to-cell spread involves selective sorting of nascent virions to cell junctions, a process that requires the cytoplasmic domain of gE. However, the large extracellular domains of gE/gI also appear to be involved in cell-to-cell spread. Here, we show that coexpression of a truncated form of gE and gI in a human keratinocyte line, HaCaT cells, decreased the spread of HSV between cells. This truncated gE/gI was found extensively at cell junctions. Expression of wild-type gE/gI that accumulates at intracellular sites, in the trans-Golgi network, did not reduce cell-to-cell spread. There was no obvious reduction in production of infectious HSV in cells expressing gE/gI, and virus particles accumulated at cell junctions, not at intracellular sites. Expression of HSV gD, which is known to bind virus receptors, also blocked cell-to-cell spread. Therefore, like gD, gE/gI appears to be able to interact with cellular components of cell junctions, gE/gI receptors which can promote HSV cell-to-cell spread.     

The gIII glycoprotein of pseudorabies virus is involved in two distinct steps of virus attachment.

The entry of herpesviruses into cells involves two distinct stages: attachment or adsorption to the cell surface followed by internalization. The virus envelope glycoproteins have been implicated in both stages. Pseudorabies virus attaches to cells by an early interaction that involves the viral glycoprotein gIII and a cellular heparinlike substance. We examined the role of gIII in the attachment process by analysis of a set of viruses carrying defined gIII mutations. The initial attachment of gIII mutants with an internal deletion of 134 amino acids (PrV2) to MDBK cells was indistinguishable from that of wild-type virus. The adsorption of these mutants was, however, much more sensitive than that of wild-type virus to competing heparin. Furthermore, while attachment of wild-type virus to MDBK cells led to a rapid loss of sensitivity to heparin, this was not the case with PrV2, which could be displaced from the cell surface by heparin after it had attached to the cells. We conclude that glycoprotein gIII is involved in two distinct steps of virus attachment and that the second of these steps but not the first is defective in PrV2.     

Role of a structural glycoprotein of pseudorabies in virus virulence.

The virulence of deletion mutants of pseudorabies virus defective in the expression of glycoprotein gI, gp63, or both was tested in 1-day-old chickens and young pigs. In the absence of expression of gI, the virulence of a fully virulent laboratory strain, PrV(Ka), for 1-day-old chickens was reduced approximately fourfold. Inactivation of glycoprotein gp63 appeared also to affect the virulence of PrV(Ka) only slightly, as did inactivation of both gI and gp63. The level of reduction in virulence, however, was considerably more marked in Bartha 43/25aB4, a less virulent virus strain. Inactivation of the expression of gI in Bartha 43/25aB4 reduced virulence for chickens at least 100-fold. The results obtained when the virulence of the mutants for pigs was determined were compatible with those obtained for chickens. These results indicate that gI plays a role in virulence, but that it does so in conjunction with at least one other viral function (a function that is defective in Bartha 43/25aB4).     

Herpes simplex virus gE/gI must accumulate in the trans-Golgi network at early times and then redistribute to cell junctions to promote cell-cell spread

Herpes simplex virus (HSV) glycoprotein heterodimer gE/gI is necessary for virus spread in epithelial and neuronal tissues. Deletion of the relatively large gE cytoplasmic (CT) domain abrogates the ability of gE/gI to mediate HSV spread. The gE CT domain is required for the sorting of gE/gI to the trans-Golgi network (TGN) in early stages of virus infection, and there are several recognizable TGN sorting motifs grouped near the center of this domain. Late in HSV infection, gE/gI, other viral glycoproteins, and enveloped virions redistribute from the TGN to epithelial cell junctions, and the gE CT domain is also required for this process. Without the gE CT domain, newly enveloped virions are directed to apical surfaces instead of to cell junctions. We hypothesized that the gE CT domain promotes virus envelopment into TGN subdomains from which nascent enveloped virions are sorted to cell junctions, a process that enhances cell-to-cell spread. To characterize elements of the gE CT domain involved in intracellular trafficking and cell-to-cell spread, we constructed a panel of truncation mutants. Specifically, these mutants were used to address whether sorting to the TGN and redistribution to cell junctions are necessary, and sufficient, for gE/gI to promote cell-to-cell spread. gE-519, lacking 32 C-terminal residues, localized normally to the TGN early in infection and then trafficked to cell junctions at late times and mediated virus spread. By contrast, mutants gE-495 (lacking 56 C-terminal residues) and gE-470 (lacking 81 residues) accumulated in the TGN but did not traffic to cell junctions and did not mediate cell-to-cell spread. A fourth mutant, gE-448 (lacking most of the CT domain), did not localize to cell junctions and did not mediate virus spread. Therefore, the capacity of gE/gI to promote cell-cell spread requires early localization to the TGN, but this is not sufficient for virus spread. Additionally, gE CT sequences between residues 495 and 519, which contain no obvious cell sorting motifs, are required to promote gE/gI traffic to cell junctions and cell-to-cell spread.     

Synthesis, cellular location, and immunogenicity of bovine herpesvirus 1 glycoproteins gI and gIII expressed by recombinant vaccinia virus.

Two of the major glycoproteins of bovine herpesvirus 1 (BHV-1) are gI, a polypeptide complex with apparent molecular weights of 130,000, 74,000, and 55,000, and gIII (a 91,000-molecular-weight [91K] glycoprotein), which also exists as a 180K dimer. Vaccinia virus (VAC) recombinants were constructed which carry full-length gI (VAC-I) or gIII (VAC-III) genes. The genes for gI and gIII were each placed under the control of the early VAC 7.5K gene promoter and inserted within the VAC gene for thymidine kinase. The recombinant viruses VAC-I and VAC-III retained infectivity and expressed both precursor and mature forms of glycoproteins gI and gIII. The polypeptide backbones, partially glycosylated precursors, and mature gI and gIII glycoproteins were indistinguishable from those produced in BHV-1-infected cells. Consequently, they were apparently cleaved, glycosylated, and transported in a manner similar to that seen during authentic BHV-1 infection, although the processing efficiencies of both gI and gIII were generally higher in recombinant-infected cells than in BHV-1-infected cells. Immunofluorescence studies further demonstrated that the mature gI and gIII glycoproteins were transported to and expressed on the surface of cells infected with the respective recombinants. Immunization of cattle with recombinant viruses VAC-I and VAC-III resulted in the induction of neutralizing antibodies to BHV-1, which were reactive with authentic gI and gIII. These data demonstrate the immunogenicity of VAC-expressed gI and gIII and indicate the potential of these recombinant glycoproteins as a vaccine against BHV-1.     

Glycoproteins E and I of Marek's Disease Virus Serotype 1 Are Essential for Virus Growth in Cultured Cells

The role of glycoprotein E (gE) and gI of Marek's disease virus serotype 1 (MDV-1) for growth in cultured cells was investigated. MDV-1 mutants lacking either gE (20DeltagE), gI (20DeltagI), or both gE and gI (20DeltagEI) were constructed by recE/T-mediated mutagenesis of a recently established infectious bacterial artificial chromosome (BAC) clone of MDV-1 (D. Schumacher, B. K. Tischer, W. Fuchs, and N. Osterrieder, J. Virol. 74:11088-11098, 2000). Deletion of either gE or gI, which form a complex in MDV-1-infected cells, resulted in the production of virus progeny that were unable to spread from cell to cell in either chicken embryo fibroblasts or quail muscle cells. This was reflected by the absence of virus plaques and the detection of only single infected cells after transfection, even after coseeding of transfected cells with uninfected cells. In contrast, growth of rescuant viruses, in which the deleted glycoprotein genes were reinserted by homologous recombination, was indistinguishable from that of parental BAC20 virus. In addition, the 20DeltagE mutant virus was able to spread from cell to cell when cotransfected into chicken embryo fibroblasts with an expression plasmid encoding MDV-1 gE, and the 20DeltagI mutant virus exhibited cell-to-cell spread capability after cotransfection with a gI expression plasmid. The 20DeltagEI mutant virus, however, was not able to spread in the presence of either a gE or gI expression plasmid, and only single infected cells were detected by indirect immunofluorescence. The results reported here demonstrate for the first time that both gE and gI are absolutely essential for cell-to-cell spread of a member of the Alphaherpesvirinae.     

Mutagenesis of varicella-zoster virus glycoprotein I (gI) identifies a cysteine residue critical for gE/gI heterodimer formation, gI structure, and virulence in skin cells

Varicella-zoster virus (VZV) is the alphaherpesvirus that causes chicken pox (varicella) and shingles (zoster). The two VZV glycoproteins gE and gI form a heterodimer that mediates efficient cell-to-cell spread. Deletion of gI yields a small-plaque-phenotype virus, ΔgI virus, which is avirulent in human skin using the xenograft model of VZV pathogenesis. In the present study, 10 mutant viruses were generated to determine which residues were required for the typical function of gI. Three phosphorylation sites in the cytoplasmic domain of gI were not required for VZV virulence in vivo. Two deletion mutants mapped a gE binding region in gI to residues 105 to 125. A glycosylation site, N116, in this region did not affect virulence. Substitution of four cysteine residues highly conserved in the Alphaherpesvirinae established that C95 is required for gE/gI heterodimer formation. The C95A and Δ105-125 (with residues 105 to 125 deleted) viruses had small-plaque phenotypes with reduced replication kinetics in vitro similar to those of the ΔgI virus. The Δ105-125 virus was avirulent for human skin in vivo. In contrast, the C95A mutant replicated in vivo but with significantly reduced kinetics compared to those of the wild-type virus. In addition to abolished gE/gI heterodimer formation, gI from the C95A or the Δ105-125 mutant was not recognized by monoclonal antibodies that detect the canonical conformation of gI, demonstrating structural disruption of gI in these viruses. This alteration prevented gI incorporation into virus particles. Thus, residues C95 and 105 to 125 are critical for gI structure required for gE/gI heterodimer formation, virion incorporation, and ultimately, effective viral spread in human skin.     

Role of pseudorabies virus glycoprotein gI in virus release from infected cells.

The Bartha vaccine strain of pseudorabies virus has a deletion in the short unique (Us) region of its genome which includes the genes that code for glycoproteins gI and gp63 (E. Petrovskis, J. G. Timmins, T. M. Gierman, and L. E. Post, J. Virol. 60:1166-1169, 1986). Restoration of an intact Us to the Bartha strain enhances its ability to be released from infected rabbit kidney cells and increases the size of the plaques formed on these cells (T. Ben-Porat, J. M. DeMarchi, J. Pendrys, R. A. Veach, and A. S. Kaplan, J. Virol. 57:191-196, 1986). To determine which gene function plays a role in virus release from rabbit kidney cells, deletions were introduced into the genomes of both wild-type virus and the "rescued" Bartha strain (Bartha strain to which an intact Us had been restored) that abolish the expression of either the gI gene alone or both gI and gp63 genes. The effect of these deletions on the phenotype of the viruses was studied. Deletion mutants of wild-type virus defective in either gI or gI and gp63 behave like wild-type virus with respect to virus release and plaque size on rabbit kidney cells. Deletion of gI from the rescued Bartha strain, however, strongly affects virus release and causes a decrease in plaque size. We conclude that gI affects virus release but that at least one other viral function also affects this process. This function is defective in the Bartha strain but not in wild-type virus; in its absence gI is essential to efficient release of the virus from rabbit kidney cells.     

Role of the pseudorabies virus gI cytoplasmic domain in neuroinvasion, virulence, and posttranslational N-linked glycosylation

The glycoproteins I and E of pseudorabies virus are important mediators of cell-to-cell spread and virulence in all animal models tested. Although these two proteins form a complex with one another, ascribing any function to the individual proteins has been difficult. We have shown previously, using nonsense mutations, that the N-terminal ectodomain of the gE protein is sufficient for gE-mediated transsynaptic spread whereas the cytoplasmic domain of the protein is required for full expression of virulence. These same studies demonstrated that the cytoplasmic domain of gE is also required for endocytosis of the protein. In this report, we describe the construction of viruses with nonsense mutations in gI that allowed us to determine the contributions of the gI cytoplasmic domain to protein expression as well as virus neuroinvasion and virulence after infection of the rat eye. We also constructed double mutants with nonsense mutations in both gE and gI so that the contributions of both the gE and gI cytoplasmic domains could be determined. We observed that the gI cytoplasmic domain is required for efficient posttranslational modification of the gI protein. The gE cytoplasmic domain has no effect on gE posttranslational glycosylation. In addition, we found that infection of all gE-gI-dependent anterograde circuits projecting from the rat retina requires both ectodomains and at least one of the cytoplasmic domains of the proteins. The gI cytoplasmic domain promotes transsynaptic spread of virus better than the gE cytoplasmic domain. Interestingly, both gE and gI cytoplasmic tails are required for virulence; lack of either one or both results in an attenuated infection. These data suggest that gE and gI play differential roles in mediating directional neuroinvasion of the rat; however, the gE and gI cytoplasmic domains most likely function together to promote virulence.