Abstract
Herpes simplex virus type 1 glycoprotein M (gM) is a type III membrane protein conserved throughout the family Herpesviridae. However, despite this conservation, gM is classed as a non-essential protein in most alphaherpesviruses. Previous data have suggested that gM is involved in secondary envelopment, although how gM functions in this process is unknown. Using transfection-based assays, we have previously shown that gM is able to mediate the internalization and subcellular targeting of other viral envelope proteins, suggesting a possible role for gM in localizing herpesvirus envelope proteins to sites of secondary envelopment. To investigate the role of gM in infected cells, we have now analysed viral envelope protein localization and virion incorporation in cells infected with a gM-deletion virus or its revertant. In the absence of gM expression, we observed a substantial inhibition of glycoprotein H–L (gH–L) internalization from the surface of infected cells. Although deletion of gM does not affect expression of gH and gL, virions assembled in the absence of gM demonstrated significantly reduced levels of gH–L, correlating with defects of the gM-negative virus in entry and cell-to-cell spread. These data suggest an important role of gM in mediating the specific internalization and efficient targeting of gH–L to sites of secondary envelopment in infected cells.
Introduction
Herpes simplex virus type 1 (HSV-1) is a common human pathogen that causes most forms of non-genital herpes simplex infection. It is capable of infecting humans of all ages, becoming latent in the nervous system and reactivating periodically throughout life. As with all members of the family Herpesviridae, HSV-1 has a dsDNA genome contained in an icosahedral capsid, surrounded by a proteinaceous tegument and a cell-derived lipid envelope carrying many viral glycoproteins. In the widely accepted model of herpesvirus replication, the formation of mature virions involves the budding of tegumented capsids into the lumen of cytoplasmic membrane compartments in a process termed secondary envelopment (Mettenleiter et al., 2009). The precise nature of the cytoplasmic membranes where virus assembly occurs is unclear, but is thought to be derived either from the trans-Golgi network (TGN) or from endosomes, or from some combination of these compartments (Johnson & Baines, 2011).
At least 15 different viral transmembrane proteins have been identified in HSV-1 particles, all of which must be correctly localized to the appropriate cellular membranes to be incorporated into virions (Adams et al., 1998; Foster et al., 2001; Ghiasi et al., 1998; Loret et al., 2008). All membranous compartments that are functionally located between the TGN and the cell surface, including many classes of endosomes, are highly dynamic with a continual exchange of lipid and protein driven by vesicle transport (Seaman, 2008). The localization of membrane proteins within this dynamic system is usually mediated by discrete targeting motifs within their cytoplasmic domains that interact with vesicle-forming machinery. Many herpesvirus envelope proteins contain such potential targeting signals, for example tyrosine-based motifs (YXXΦ, where X = any amino acid and Φ = a bulky hydrophobic residue). YXXΦ motifs play important roles in endocytosis and targeting to intracellular compartments through interactions with clathrin adaptor proteins (Favoreel, 2006). Functional YXXΦ endocytosis motifs have been identified in the HSV envelope proteins glycoprotein B (gB) and glycoprotein E (gE) (Alconada et al., 1999; Beitia Ortiz de Zarate et al., 2004; Fan et al., 2002). YXXΦ motifs are also present in the cytoplasmic domains of the HSV envelope proteins glycoprotein M (gM), glycoprotein K (gK) and pUL20, and both gM and the gK–pUL20 complex localize to intracellular compartments showing overlap with TGN markers (Crump et al., 2004; Foster et al., 2004). However, there are no recognizable targeting motifs in glycoprotein D (gD) or the glycoprotein H–L complex (gH–L). These essential envelope proteins, together with gB, constitute the virus entry machinery and thus it is critical for the assembly of infectious viruses that both gD and gH–L are localized to the intracellular membranes where secondary envelopment occurs (Heldwein & Krummenacher, 2008). The mechanism for this process is currently unclear, but presumably relies on the activity of other viral proteins, because gD and gH–L are localized at the plasma membrane when expressed in the absence of HSV-1 infection. We have previously demonstrated that gM may play a role in this because both gD and gH–L are efficiently internalized and targeted to intracellular compartments when co-expressed with gM (Crump et al., 2004).
Despite being one of the few envelope proteins conserved throughout the family Herpesviridae, the function of gM is poorly understood. Encoded by the UL10 gene in HSV-1, gM is a type III membrane protein with eight predicted transmembrane domains (Baines & Roizman, 1993; MacLean et al., 1991, 1993). In several herpesviruses, gM has been shown to form a disulphide-linked complex with glycoprotein N (gN), although an interaction between HSV-1 gM and gN has yet to be demonstrated (Jöns et al., 1998; Koyano et al., 2003; Lake et al., 1998; Lipińska et al., 2006; Mach et al., 2000; Rudolph et al., 2002; Wu et al., 1998). It has been reported that gM is non-essential for the replication of many alphaherpesviruses, including HSV-1 (Baines & Roizman, 1991; MacLean et al., 1993), pseudorabies virus (PRV) (Dijkstra et al., 1996), bovine herpesvirus type 1 (König et al., 2002), varicella-zoster virus (Yamagishi et al., 2008), infectious laryngotracheitis virus (Fuchs & Mettenleiter, 1999), equine herpesvirus type 1 (EHV-1) (Osterrieder et al., 1996) and EHV type 4 (Ziegler et al., 2005). However, gM is essential for replication of the alphaherpesvirus Marek’s disease virus, of human cytomegalovirus (a betaherpesvirus) and of murid herpesvirus 4 (a gammaherpesvirus) (Hobom et al., 2000; May et al., 2005; Tischer et al., 2002). The phenotype of deleting gM in many alphaherpesviruses includes small reductions in virus-replication rates, reduced plaque sizes and, where studied, reduced virus-penetration rates (Baines & Roizman, 1991; Dijkstra et al., 1996; König et al., 2002; Leege et al., 2009; MacLean et al., 1991, 1993; Osterrieder et al., 1996; Yamagishi et al., 2008; Ziegler et al., 2005). A role for gM in secondary envelopment has been suggested in HSV-1, PRV and EHV-1 infection based on electron microscopy studies. However, the defects are generally mild unless gM is deleted simultaneously with either the glycoprotein E–I or pUL11 (Brack et al., 1999; Kopp et al., 2004; Leege et al., 2009; Seyboldt et al., 2000). Thus gM may have a partially redundant or synergistic role in the envelopment process.
In order to investigate the role of gM in the localization of other HSV-1 envelope proteins in infected cells, we have analysed a gM-deletion virus together with its revertant. Our data demonstrate that loss of gM expression causes a specific defect in the internalization of gH–L from the plasma membrane and reductions in virion incorporation of gH–L in infected cells. We suggest that gM has an important role in the assembly of infectious HSV-1 due to its ability to control gH–L localization.
Results
Construction of viruses and growth properties
HSV-1 lacking gM expression (ΔgM) was constructed by replacing codons 134–466 of UL10 with a β-galactosidase reporter gene, in the virus strain SC16, using homologous recombination (Fig. 1). Correct insertion of the lacZ reporter gene into the HSV-1 genome (SC16) was verified by β-galactosidase activity and Southern blot (data not shown). A revertant virus (gMR) was generated as a control by replacing the lacZ reporter gene with wild-type UL10.
Construction of HSV-1 ΔgM. A schematic of the HSV-1 genome with relevant restriction sites and ORFs highlighted. The region of UL10 between the indicated MscI and BsrGI restriction enzyme recognition sites in HSV-1 strain SC16 was replaced with a lacZ reporter gene, leaving 399 bp of UL10 remaining at the 5′ end and 22 bp at the 3′ end.
The ΔgM, gMR and wild-type (SC16) viruses were analysed by single- and multi-step growth assays in HT29 (human colon epithelial) and Vero (primate kidney epithelial) cells. SC16 and gMR had virtually identical growth properties in both cell lines (Fig. 2a–d). In HT29 cells treated with a high m.o.i. (10), ΔgM demonstrated a slight delay in the onset of virus production and a small replication deficit with approximately sixfold reduction in final cell-associated titre (Fig. 2a). A similar early delay and reduction in final infectious titre (approximately 12-fold) was also observed for ΔgM virus released into the culture medium. Greater inhibition of replication of ΔgM was observed in HT29 cells at low m.o.i. (0.01) with up to 37-fold reduction in infectious titre at 36–48 h post-infection (p.i.), suggesting additional defects in cell-to-cell spread (Fig. 2c). In Vero cells there was little difference in the growth properties of ΔgM compared with the control viruses at high m.o.i., with two- to threefold reduction in final cell-associated infectious titre (Fig. 2b), comparable to previous data on an independently generated SC16-based gM-deletion virus in Vero cells (Browne et al., 2004). There was also a delay and a similar reduction in final infectious titre (approx. 12-fold) for ΔgM virus released into the medium from infected Vero cells. At low m.o.i., ΔgM also showed a delay in replication kinetics in Vero cells, with four- to eightfold-lower titres at 36–48 h p.i., again suggesting defects in spread between cells (Fig. 2d). In addition, the plaque sizes of SC16 and gMR were virtually identical, whereas the ΔgM plaques were significantly smaller on both HT29 and Vero monolayers, the plaque diameter being reduced by approximately 60 % in both cell types (Fig. 3). Taken together, these data suggest that the absence of gM causes a minor defect in the assembly of infectious virions, a small delay in the release of infectious HSV-1 into the culture medium, and less efficient cell-to-cell spread. Given that plaque size and growth kinetics were virtually identical between wild-type SC16 and gMR, further experiments were performed using ΔgM and gMR only.
Virus-growth analysis. Single-step (a, b) and multi-step (c, d) growth analysis at an m.o.i. of 10 and 0.01, respectively, were performed on HT29 and Vero cells. For single-step growth analyses (a, b) infectious titres from both culture media and cells are shown. For multi-step growth analyses (c, d) total infectious titres from cells and media combined are shown. Data are shown as mean titres and sem (n = 3).
Plaque-size quantification. Representative images of plaques on HT29 and Vero monolayers (a). The relative plaque sizes of ΔgM and gMR calculated as the percentage of SC16 (set as 100 %) (b). Data are shown as means and sem (n = 50). ***, Significantly different at P<0.001 (two-tailed Student’s t-test).
HSV-1 protein expression levels in infected cells
The expression of several viral proteins in infected HT29 and Vero cells was analysed by Western blotting. No detectable gM expression could be observed in either cell line infected by ΔgM, as expected (Fig. 4a, b). The expression levels of all viral proteins tested, including the major capsid protein VP5, the major tegument proteins VP16 and VP22, and the envelope proteins gB, gC, gD and gH, were highly similar between ΔgM- and gMR-infected HT29 or Vero cells. In Vero cells, a greater proportion of higher-molecular-mass species was observed for gB, gC and gD in cells infected with ΔgM. Importantly, despite the proximity of the UL11 gene to the engineered deletion within the gM gene (UL10), only a slight reduction in the expression of pUL11 could be observed in ΔgM-infected cells. The expression of pUL9 could not be assessed due to lack of specific antisera although, given the essential nature of pUL9 as the origin-binding protein for genome replication, it seems unlikely that expression of this gene is affected by the engineered deletion in UL10.
Western blot analysis of viral protein expression levels in infected cells. HT29 cells (a) or Vero cells (b) were infected with ΔgM or gMR at an m.o.i. of 10 and harvested at 24 h p.i. Twofold dilutions of cell lysates were loaded and samples were probed with antibodies against HSV-1 proteins. The tubulin signals were used as loading controls.
The absence of gM leads to defects in the incorporation of gH–L into virions
To analyse whether the deletion of gM had any effect on the incorporation of envelope proteins into virions, ΔgM and gMR viruses were purified from the medium of infected cells and the viral protein content was examined by SDS-PAGE and Western blotting. Coomassie blue-stained SDS-PAGE analysis of purified ΔgM and gMR virions demonstrated equivalent purity of virus preparations and the major viral structural proteins were clearly visible (Fig. 5a, b, upper left panels). Interestingly, an equivalent amount of p.f.u. of ΔgM virions demonstrated a greater overall protein level than gMR virions. Furthermore, analysis of a range of p.f.u. by Western blotting demonstrated that ΔgM virions contained approximately twice the amount of the capsid and tegument proteins than the equivalent p.f.u. of gMR (Fig. 5a, b, upper right panels). This suggests that the particle : p.f.u. ratio is higher for ΔgM and that the loss of gM causes virions to be less infectious. When comparing similar levels of capsid (VP5) loading (compare 16×106 p.f.u. gMR with 8×106 p.f.u. ΔgM), the two viruses harvested from HT29 cells had no observable difference in the levels of tegument proteins VP16 and VP22 or envelope proteins gB and gD. However, a small decrease in gC and a large decrease in gH levels were observed in ΔgM virions (Fig. 5a). Quantification of signals and normalization of the data to VP5 levels demonstrated that ΔgM virions contained approximately sixfold less gH and threefold less gC than gMR virions (Fig. 5a, lower panel). Similarly, viruses purified from the supernatant of infected Vero cells had no detectable difference in VP16 and VP22, but approximately threefold less gH was found in ΔgM virions (Fig. 5b). However, unlike HT29 virions, viruses prepared from Vero cells demonstrated little difference in the incorporation of gC and, surprisingly, ΔgM virions contained two to three times more gB and gD (Fig. 5b). We also observed that gL levels were reduced in ΔgM virions to the same extent as gH in virions from both cell types (data not shown). Overall, these data show that gM plays an important role in the efficient incorporation of gH–L into HSV-1 particles.
Western blot analysis of purified virions. ΔgM or gMR were purified from the culture supernatant of HT29 (a) and Vero cells (b). Purified viruses (8×106 p.f.u.) were separated by SDS-PAGE and stained with Coomassie blue (upper left panels). Molecular mass markers and the position of major viral proteins as expected from a previous publication (Dargan et al., 1995) are shown. Twofold dilution series of each virus (2–16×106 p.f.u. as indicated) were analysed by Western blotting with antibodies against HSV-1 proteins (upper right panels). The integrated intensity (I.I.) values for viral protein signals from Western blots were normalized to VP5 and the values of log2 (ΔgM : gMR) after the normalization are shown (lower panels).
Delayed entry of ΔgM virus correlates with reduced levels of gH–L in the virion
As we observed a consistent decrease in levels of the membrane-fusion mediator gH–L in ΔgM virions, the entry kinetics of purified ΔgM and gMR viruses in Vero cells were investigated. A significant delay in the rate of ΔgM virus entry compared with gMR was observed for viruses purified from both HT29 and Vero cells; the time to reach 50 % entry for ΔgM was approximately double that of gMR (Fig. 6). ΔgM purified from HT29 cells demonstrated a more severe attenuation with a significantly delayed initial entry rate (Fig. 6b). These observations correlate with the reduction of gH–L in ΔgM virions prepared from the different cell lines, HT29-purified ΔgM virions having the greatest reductions in gH–L and the most dramatic delay in entry. The reduced entry rate of ΔgM virions could at least partly account for the observed defects in cell-to-cell spread of this virus.
HSV-1 penetration rate assay. The relative penetrations of purified ΔgM and gMR from Vero (a) and HT29 (b) cells were determined. The means and sem from triplicate experiments are shown.
gH–L localization in infected cells
To investigate whether gM plays a role in targeting viral glycoproteins to cytoplasmic membranes during infection, the subcellular localization of gH–L and gD was examined by immunofluorescence microscopy. When cells were fixed and permeabilized prior to antibody staining, those infected with gMR showed gH–L predominantly localized to intracellular compartments that showed some co-localization with TGN marker protein TGN46 (Fig. 7a). In cells infected with ΔgM, while some gH–L localized to intracellular compartments, a much greater level of gH–L was observed in a diffuse pattern reminiscent of plasma-membrane localization (Fig. 7a). No apparent difference in the distribution of gD was observed, with localization predominantly to intracellular compartments that showed some overlap with TGN46 (Fig. 7a). To specifically label cell-surface glycoproteins, infected cells were incubated with gH–L- or gD-specific antibodies on ice prior to fixation, thereby avoiding any signal from glycoproteins in intracellular membranes (Fig. 7b). Only faint signals for gH–L could be detected on the surface of cells infected with gMR, whereas considerably higher levels of gH–L could be observed on cells infected with ΔgM. There was no obvious difference in the level of cell-surface gD, with strong signals for both ΔgM- and gMR-infected cells. In addition, the cell-surface levels of gH–L and gD in populations of infected cells were analysed by flow cytometry. Virtually identical surface levels of gD were observed in cells infected with either virus. However, an increase in cell-surface gH–L was observed in cells infected with ΔgM (Fig. 7c). To quantify the relative gH–L levels on the surface of HSV-1-infected cells, the anti-gH–L fluorescence signals of ΔgM-infected cells were normalized to those of gMR-infected cells for 14 independent sets of samples. These data demonstrated a 2.4-fold increase in mean fluorescence (P = 0.00017; Student’s t-test) and a 2.7-fold increase in median fluorescence (P = 0.00023; Student’s t-test) of cell-surface gH–L in ΔgM-infected samples. Taken together, these data suggest that gM plays an important role in controlling gH–L localization and that, in the absence of gM function, greater levels of gH–L are present at the plasma membrane.
gH–L and gD localization in infected cells. (a) Vero cells infected with ΔgM or gMR were fixed, permeabilized and then stained with anti-gH–L or anti-gD together with anti-TGN46. (b, d) Vero cells infected with ΔgM or gMR were incubated with gH–L- or gD-specific antibodies on ice for 1 h (b) or at 37 °C for 1 h (d) and then fixed, permeabilized and labelled with anti-TGN46 followed by fluorescent secondary antibodies. (c) Vero cells infected with ΔgM or gMR were labelled with gH–L- or gD-specific antibodies and analysed by FACS. Histograms show gH–L or gD cell-surface signals for uninfected controls (grey filled), ΔgM-infected cells and gMR-infected cells. (a, b, d) Green represents gH–L or gD and red represents TGN46.
An increase in gH–L levels at the cell surface could be due to either enhanced rates of gH–L transport through the secretory pathway to the plasma membrane, or reduced rates of internalization of gH–L. To investigate this, antibody-feeding assays were conducted where cells were incubated with antibodies specific to gH–L or gD at 37 °C for 1 h prior to fixation. We observed that, whereas gH–L-specific antibodies were efficiently internalized to cytoplasmic organelles in gMR-infected cells, in cells infected with ΔgM, gH–L-specific antibodies remained on the plasma membrane (Fig. 7d). No obvious differences were observed in the internalization of a gD-specific antibody between cells infected with either ΔgM or gMR (Fig. 7d).
Overall, the data presented here provide strong evidence for a specific role of gM in mediating the internalization of gH–L from the cell surface and targeting to intracellular compartments during HSV-1 infection. Without this functionality of gM, gH–L is incorporated less efficiently into assembling virions, leading to viruses that are attenuated in entering and spreading between cells.
Discussion
Five envelope proteins, gB, gH, gL, gM and gN, have been found to be conserved in all herpesviruses, indicating their importance to this diverse family of viruses. The essential function of gB and the gH–L complex in herpesvirus entry is now well-established and crystal structures of HSV-1 gB and HSV type 2 gH–L have recently been solved (Chowdary et al., 2010; Heldwein et al., 2006). However, the role of gM and gN in the herpesvirus life cycle remains unclear. To investigate the function of gM in HSV-1, we deleted its gene (UL10) from the low-passage strain SC16 and analysed the effects of this deletion in infected cells. Consistent with previously published work, our data suggest that HSV-1 gM is relatively dispensable for virus assembly, but is important for efficient cell-to-cell spread. However, a considerable reduction in the incorporation of gH–L into ΔgM virions was observed, correlating with reduced penetration rates. The absence of gM also appeared to marginally inhibit the release of infectious virus. It therefore seems likely that both the reduced penetration and the delayed release of ΔgM could contribute to the observed inhibition of cell-to-cell spread. Importantly, we observed an increase in cell-surface gH–L levels and a concomitant inhibition of gH–L internalization from the plasma membrane in ΔgM-infected cells. These data demonstrate a functional role of gM in the targeting of gH–L to sites of secondary envelopment for efficient virion incorporation. We believe that this is the first observation of a specific role of another HSV-1 membrane protein on gH–L incorporation into virions. Previous data on HSV-1 lacking gB, gC or gD demonstrated no effect on gH–L levels in purified viruses (Rodger et al., 2001).
Despite the lower levels of gH–L in virions in the absence of gM (three- to sixfold), there was not a complete block of gH–L incorporation, otherwise gM-deletion strains of HSV-1 would be completely attenuated. The incorporation of some gH–L into ΔgM virions suggests that gH–L can access the virus-assembly compartments prior to reaching the plasma membrane. This agrees with the fact that we could still observe low levels of gH–L in intracellular compartments in ΔgM-infected cells when the localization of total gH–L was analysed. The most likely source of such populations of gH–L would be newly synthesized protein being transported through the secretory pathway. There is current uncertainty about the nature of the compartment where herpesvirus secondary envelopment occurs, with suggestions of Golgi, TGN and/or endosomal origins for viral membranes (Johnson & Baines, 2011). Our data suggest that gM-mediated retrieval of gH–L from the plasma membrane is important to maintain optimal levels of this glycoprotein complex at intracellular sites of virus assembly. However, we cannot rule out an additional role of gM in transporting gH–L directly to virus-assembly compartments from other intracellular sites and that, in the absence of gM, more gH–L escapes to the plasma membrane.
Previous reports have demonstrated synergistic and/or redundant functions between gM and pUL11 in the cytoplasmic assembly of both HSV-1 and PRV (Kopp et al., 2004; Leege et al., 2009). Interestingly, cells infected by our ΔgM virus showed slightly lower pUL11 expression levels than cells infected with the revertant virus. Whilst we would not expect such minor differences in pUL11 expression levels to affect HSV-1 assembly, we cannot rule out the possibility that the reduced expression of pUL11 may also contribute to the observed phenotypes. It would be interesting to investigate the virion incorporation and subcellular localization of envelope proteins in pUL11-null virus-infected cells compared with ΔgM to assess the relative importance of these two proteins in gH–L targeting.
We have previously shown that gM and/or the gM–N complex can cause the internalization of both gD and gH–L from the plasma membrane in transfection assays. It was therefore surprising that we observed no decrease in the incorporation of gD into ΔgM virions, and the localization of gD in infected cells was unaffected by the absence of gM. While gM has the ability to relocalize gD in transfection assays, in the context of HSV-1 infection other viral proteins must also control gD localization. We currently do not know what performs this function, although possible candidates are gK–pUL20 or pUL43, other type III membrane proteins expressed by HSV-1. Similarly to gM, gK–pUL20 and pUL43 may also remove viral glycoproteins from the cell surface as they have been shown to inhibit cell–cell fusion induced by the virus entry proteins of HSV-1 or PRV in transfection assays (Avitabile et al., 2004; Klupp et al., 2005). The role of other viral membrane proteins on gD localization in HSV-1 infected cells is currently being investigated.
It is interesting that the absence of gM produced a more severe phenotype when analysed in the human cell line HT29 than the primate cell line Vero. These observations suggest that the localization of gH–L to assembly compartments is more dependent on gM in HT29 cells. Why such effects are not apparent in Vero cells is currently unclear, but could reflect some adaptation to replication in Vero cells in the absence of gM because the viruses were propagated in this cell type. One of the surprising observations made during these studies that could partly explain this is the increased gB and gD incorporation into ΔgM virions that was observed in Vero cells. Analysis of viral protein expression indicated a greater proportion of higher-molecular-mass gB, gD and gC species in ΔgM-infected Vero cells, suggesting more rapid maturation of these glycoproteins in the absence of gM (Fig. 4b). Interestingly, gM has been shown to localize to the endoplasmic reticulum (ER) in infected cells (Baines et al., 2007; Zhang et al., 2009). It is conceivable that such a population of gM could retard the exit of glycoproteins from the ER, thereby delaying their processing in the Golgi. This could increase levels of mature glycoproteins in post-Golgi membranes in the absence of gM, leading to enhanced incorporation into virions. However, this appears to be a phenotype specific to Vero cells; there was no observable difference in viral glycoprotein expression patterns between ΔgM and gMR in HT29 cells or other human cell lines (Fig. 4 and data not shown).
Our observation that, in HT29 cells, less gC is incorporated into virions in the absence of gM may suggest a functional role of gM in targeting gC to virus-assembly compartments. There is currently no evidence to suggest that gC is involved directly in HSV-1 assembly, although it is conceivable that the combined reduction of gH–L and gC in virus-assembly compartments could contribute to the greater inhibition of HSV-1 replication in HT29 cells. The cytoplasmic domain of HSV-1 gC is 11 aa, with no consensus trafficking motifs. Therefore, gC would be predicted to be localized to the plasma membrane in the absence of other viral proteins, and targeting gC to intracellular membranes may be important for virion incorporation. However, no obvious difference in gC incorporation was observed in ΔgM viruses from Vero cells, although this could be complicated by the increased proportion of mature gC in ΔgM-infected Vero cells. The effect of gM on the localization and trafficking of gC is being investigated.
Overall, our data demonstrate an important role for HSV-1 gM in the internalization of gH–L from the plasma membrane of infected cells and incorporation into mature virions. While many questions still remain over the mechanistic details, this work is the first demonstration of a functional role for gM in viral glycoprotein trafficking in infected cells. Given that gM appears to be an essential gene in beta- and gammaherpesviruses, it will be interesting to investigate whether the targeting of gH–L or other viral envelope proteins has an even greater reliance on gM function in these herpesvirus subfamilies.
Methods
Cells and viruses.
Vero cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10 % FCS, 2 mM l-glutamine, 100 U ml−1 penicillin and 100 µg ml−1 streptomycin. HT29 cells were grown in RPMI medium supplemented as above. All media and reagents listed above were from PAA. HSV-1 strain SC16 (Hill et al., 1975) was the strain used in this study. Viruses were propagated and titres were determined using Vero cells.
Construction of gM-deletion and -revertant HSV-1.
The plasmid pc73.1 contains the BamHI (21653)–BglII (25153) region of the HSV-1 genome that covers UL10 (23204–24625; numbering relates to GenBank accession no. NC_001806). The plasmid pc78.1 is pc73.1 with a lacZ gene inserted into an NruI site (24343) of UL10 (MacLean et al., 1991). Plasmid pc78.1 was digested with PpuMI (New England Biolabs) and religated to generate pc78.1ΔPpuMI, where the majority of UL10 (23765–24592) and the lacZ cassette were removed. A cytomegalovirus promoter-driven lacZ flanked by BsrGI and SmaI sites was ligated into MscI/BsrGI-cut pc78.1ΔPpuMI to create pc78.1ΔMscI/BsrGI-LacZ. Baby hamster kidney (BHK) cells were transfected with pc78.1ΔMscI/BsrGI-LacZ together with SC16-infected cell DNA as described previously (Balan et al., 1994). Resulting plaques were analysed by X-Gal (Melford) staining and limiting dilution was used to isolate single blue plaques. To generate an equivalently treated wild-type SC16 strain, BHK cells were transfected with high-molecular-mass DNA prepared from SC16-infected cells. A revertant virus (gMR) was made by transfecting BHK cells with pc73.1 together with high-molecular-mass DNA prepared from ΔgM-infected cells. Resulting plaques were identified by staining with X-Gal and single white plaques were isolated.
Virus growth analysis.
For single-step virus growth curves, cells were infected at 10 p.f.u. per cell at 37 °C for 1 h. Unabsorbed virus was inactivated with acid wash (40 mM citric acid, 135 mM NaCl, 10 mM KCl, pH 3.0). At various times p.i., cells were harvested by scraping into medium and infectious virus was liberated by sonication and freezing/thawing. Virus titres were determined by plaque assay. In multi-step growth curves, cells were infected at 0.01 p.f.u. per cell and treated as described above.
Plaque assay and plaque-size analysis.
Cell monolayers were infected at 37 °C for 1 h and then overlaid with CMC (0.6 % carboxymethyl cellulose in DMEM+2 % FCS; Sigma). After incubation at 37 °C for 3 days, cells were fixed with 3.7 % formaldehyde and stained with 0.1 % toluidine blue (Sigma). Plaques were scanned at 600 days p.i., and diameters were determined by using Adobe Photoshop.
Virus purification.
Cells were infected at 0.01 p.f.u. per cell and incubated for 3 days. The culture medium was harvested and centrifuged at 2000 r.p.m. (Beckman GH3.8) for 20 min to remove cell debris. Supernatant virions were pelleted at 18 000 r.p.m. (Beckman type 19 rotor) for 2 h, resuspended in 2 ml of 1 % FCS/PBS and centrifuged through a 30 ml 5–15 % continuous Ficoll gradient at 12 000 r.p.m. (Beckman SW 32Ti) for 1.5 h. The clear virus band in the middle of each gradient was harvested, and the virions were pelleted at 20 000 r.p.m. (Beckman SW 32Ti) for 2 h. All centrifugations were performed at 4 °C. The pellets were resuspended in PBS and stored at −70 °C.
Virus-penetration assay.
Vero cell monolayers were chilled at 4 °C followed by incubation with 300 p.f.u. purified viruses in ice-cold HEPES-buffered medium per well for 1 h. The virus-containing medium was removed and cells were washed in ice-cold PBS and then incubated at 37 °C. At various time points, unabsorbed virus was inactivated by acid wash. For the 0 min time point, the acid wash was performed immediately after the 4 °C virus-binding step. Samples without acid wash served as controls for input virus levels. The cells were overlaid with CMC and incubated at 37 °C before fixing and staining as described above. Plaque numbers were counted for each time point and the rate of penetration was calculated as the percentage of plaques produced at different time points compared with the final time point.
Western blot analysis.
Infected cell samples were harvested at 24 h p.i. and cells were resuspended in lysis buffer (50 mM Tris, pH7.9, 150 mM NaCl, 1 % Triton X-100, 1 % sodium deoxycholate) supplemented with protease inhibitor cocktail (Roche) and incubated on ice for 20 min. Cell debris was removed by centrifugation at 13 000 r.p.m. for 15 min at 4 °C. Lysates or purified virus samples were separated by SDS-PAGE followed by electrophoretic transfer to nitrocellulose and the presence of various proteins was determined using the primary antibodies described below. The LI-COR Odyssey Infrared Imager was used for signal scanning and the Odyssey v3.0 software was used to calculate the integrated intensity for quantification.
Primary antibodies used were: anti-VP5 (Abcam, ab6508), anti-VP16 (Abcam, ab110226), anti-VP22 (AGV30 from G. Elliott), anti-gM (anti-gM-B from H. Browne), anti-gD (LP14; Minson et al., 1986), anti-gH (Abcam, ab110227), anti-gB (R69 from G. Cohen), anti-gC (Abcam, ab6509), anti-UL11 (from T. Mettenleiter), anti-tubulin alpha (AbD Serotec, MCA77G).
Immunofluorescence microscopy.
Vero cells grown on glass coverslips were infected at 3 p.f.u. per cell and fixed in 3 % formaldehyde at 6 h p.i. Samples were incubated with perm/quench solution (50 mM NH4Cl, 0.2 % saponin in PBS; Sigma), and blocked with PGAS (0.2 % gelatin, 0.02 % saponin, 0.02 % NaN3 in PBS; Sigma) containing human IgG (0.5 mg ml−1) before being labelled with primary antibodies against gH–L (LP11) or gD (LP2) (Minson et al., 1986) together with TGN46 (AbD Serotec). Alternatively samples were incubated with LP11 or LP2 on ice for 1 h prior to fixation to specifically label cell surface-localized gH–L or gD, or incubated with LP11 or LP2 for 1 h at 37 °C prior to fixation to investigate gH–L or gD internalization. All samples were incubated with secondary antibodies (donkey anti-mouse–Alexa Fluor 488 and donkey anti-sheep–Alexa Fluor 568; Invitrogen), embedded in ProLong Gold (Invitrogen) and examined on an inverted confocal miscroscope (Leica TCS SP5 II). All channels were recorded separately and acquired images were processed with Adobe Photoshop.
FACS.
Vero cells were infected at 10 p.f.u. per cell, harvested at 6 h p.i. and incubated with LP11 or LP2 antibodies on ice for 1 h. After two ice-cold PBS washes, the cells were incubated with donkey anti-mouse–Alexa Fluor 488 on ice for 1 h. Cells were washed and resuspended in cold PBS and analysed using a BD FACSort cell sorter. Data from 10 000 cells were collected for each sample and the fluorescent signals were analysed using Summit 4.3 software.
Acknowledgements
We thank T. Minson and H. Browne (University of Cambridge, UK), G. Cohen (University of Pennsylvania, PA, USA), T. Mettenleiter (Friedrich-Loeffler-Institut, Germany) and G. Elliott (Imperial College London, UK) for antisera, and S. Efstathiou, M. Gill and D. Glauser for helpful discussions. This work was supported by the Royal Society (UF090010) and the Medical Research Council (G0700129). PhD studentship funding for Y. R. was received jointly from Cambridge Overseas Trust, Robinson College, the Board of Graduate Studies, and the Department of Pathology at the University of Cambridge. PhD studentship funding for S.-Y. K. L. was from the Biotechnology and Biological Sciences Research Council.