|Home | About | Journals | Submit | Contact Us | Français|
Of the four required herpes simplex virus (HSV) entry glycoproteins, the precise role of gH-gL in fusion remains the most elusive. The heterodimer gH-gL has been proposed to mediate hemifusion after the interaction of another required glycoprotein, gD, with a receptor. To identify functional domains of HSV-1 gH, we generated 22 randomized linker-insertion mutants. Analyses of 22 gH mutants revealed that gH is relatively tolerant of insertion mutations, as 15 of 22 mutants permitted normal processing and transport of gH-gL to the cell surface. gH mutants that were not expressed well at the cell surface did not function in fusion or viral entry. The screening of gH mutants for function revealed the following: (i) for wild-type gH and some gH mutants, fusion with nectin-1-expressing target cells occurred more rapidly than with herpesvirus entry mediator (HVEM)-expressing target cells; (ii) some gH mutants reduced the rate of cell fusion without abrogating fusion completely, indicating that gH may play a role in governing the kinetics of fusion and may be responsible for a rate-limiting first stage in HSV-1 fusion; and (iii) only one gH mutant, located within the short cytoplasmic tail, completely abrogated function, indicating that the gH cytoplasmic tail is crucial for cell fusion and viral infectivity.
Herpes simplex virus (HSV), an enveloped neurotropic virus, infects target cells via membrane fusion, a process executed by viral fusion proteins capable of inserting into target membranes. Unlike many enveloped viruses that induce fusion through the activity of a single viral fusion protein, HSV requires four glycoproteins, glycoprotein B (gB), glycoprotein D (gD), glycoprotein H (gH), and glycoprotein L (gL), to execute fusion (6, 40, 42). The focus of this study, gH, is expressed as a heterodimer with gL (gH-gL). HSV gH and gL rely on one another for proper folding, posttranslational processing, and transport to the cell and virion surface (5, 23, 35).
A sequential model of entry is the prevailing working hypothesis of HSV entry (1-3, 28, 32, 41). Viral attachment is mediated by the binding of glycoprotein C (gC) or gB to cell surface glycosaminoglycans such as heparan sulfate (38). The subsequent fusion between the virion envelope and host cell membrane is thought to result from a series of concerted events. First, gD binds to one of its host cell receptors. These receptors include herpesvirus entry mediator (HVEM), a member of the tumor necrosis factor (TNF) receptor family; nectin-1 and nectin-2, cell adhesion molecules of the Ig superfamily; and heparan sulfate modified by specific 3-O-sulfotransferases (39).
It was previously proposed that gD binding a receptor induces a conformational change that allows for interactions between gD, gB, and/or gH-gL (1, 2, 8, 10, 16, 25, 32). It is thought that while gD functions primarily in receptor binding, gB and gH-gL function as the core fusion machinery of HSV.
Based on its crystal structure, gB has structural features typical of viral fusion proteins in general and is structurally similar to vesicular stomatitis virus (VSV) glycoprotein G, the fusion protein of VSV (22, 34). In addition to its resemblance to other viral fusogens, gB also binds its own receptor, paired immunoglobulin-like receptor (PILRalpha) (36, 37). Importantly, HSV gB does not successfully execute fusion in the absence of gD or gH-gL (41). Compared to the other required HSV entry glycoproteins, relatively little is known about the specific roles of gH-gL during fusion. The structure of gH-gL is unknown, although in silico analyses and studies of synthetic gH peptides suggested that gH also has fusogenic properties (12, 13, 17-20).
gD, a gD receptor, and gH-gL have been shown to be sufficient for inducing hemifusion, the mixing of the proximal leaflets of the viral and host cell bilayers (41). Several lines of research suggest that the subsequent step in fusion is an interaction between gH-gL and gB, with the latter glycoprotein being required for a committed and expanding fusion pore (1-3, 16, 28, 41). However, it is still unclear whether the gB and gH-gL interaction requires that gD first bind a receptor (1, 3), indicating that another viable model of HSV entry may be nonsequential gD-gB-gH-gL complex formation.
Several domains important for fusion within HSV gH have been discerned. The only function associated with the N-terminal domain of HSV gH, to date, is gL binding. Residues 377 to 397 within a predicted alpha-helix in the gH ectodomain are required for cell-cell fusion and complementation of a gH-null virus (18). The mutation of a predicted heptad repeat region spanning residues 443 to 471 abrogated cell-cell fusion (17). Insertion mutations within what has been termed the pretransmembrane region of gH have also been shown to abrogate fusion and viral entry (11). The glycine residue at position 812 within the predicted gH transmembrane domain was shown previously to be important for fusion (21). Finally, although the deletion of the final six residues of gH (residues 832 to 838), which are within its short cytoplasmic tail, has no effect on fusion, further deletions were shown to decrease polykaryocyte formation by a syncytial HSV strain (4, 43).
We used a transposon-based comprehensive random linker-insertion mutagenesis strategy to generate a library of mutants spanning the entire length of HSV-1 gH, an 838-amino-acid type I membrane protein. A panel of 22 insertion mutants was generated, 15 of which were expressed at near-normal levels on the cell surface. Interestingly, some insertions reduced the rate of cell fusion rather than abrogating cell fusion activity altogether, suggesting that gH may have a role in governing the kinetics of fusion and may be responsible for a rate-limiting first stage in HSV-1 fusion. Additionally, one insertion mutation that completely abrogated cell fusion and viral infectivity is located within the gH cytoplasmic tail, indicating that the short C-terminal tail of gH is critical for cell fusion and entry mediated by HSV-1.
Cell lines used were CHO cells, CHO cells stably expressing human HVEM (29) (CHO-HVEM) or nectin-1 (15) (CHO-nectin-1), Vero cells, and Vero-VgHC4 cells carrying the HSV-1 gH gene, which were used for the propagation and titration of the gH-negative mutant HSV-1 (KOS) virus named KOS gH87 (30). The CHO lines were grown in Ham's F12 medium supplemented with 10% fetal bovine serum (FBS). The Vero lines were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS.
The GPS-LS linker scanning system (New England Biolabs, Ipswich, MA) was used as recommended by the manufacturer for the insert excised from pPEP100 (33) to generate random linker-insertion mutations of HSV-1 gH. After the religation of the library of inserts into the pCAGGS vector and transformation of competent bacteria, 87 unique gH linker-insertion mutants were isolated. After the removal of the transposon, each mutant plasmid was fully sequenced to verify the position of each insert. Plasmids containing inserts outside of the gH open reading frame (ORF), stop codon insertional mutations, and duplicates of identical inserts were dropped from further study.
CHO cells seeded into 96-well plates grown to approximately 80% confluence were transfected with 20 ng of empty vector or a plasmid expressing wild-type (WT) gH or mutant gH, 20 ng each of WT gD and gL, and 40 ng of gB with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) diluted in Opti-MEM (Gibco, Carlsbad, CA). After a 2-h transfection, the cells were washed once with Opti-MEM and then incubated with serum-containing medium for 10 h. Cell-based enzyme-linked immunosorbent assays (CELISAs) were performed, as described elsewhere previously (14), by using anti-gH monoclonal antibody (MAb) 52S or gH-gL monoclonal antibody 53S at a 1:10,000 dilution. For time course experiments, CHO cells seeded into 6-well plates grown to approximately 80% confluence were transfected with 400 ng of empty vector or a plasmid expressing WT gH or mutant gH by using Lipofectamine 2000. After a 4-h transfection, cells were washed once and then incubated with serum-containing medium for the indicated times. In addition, CHO cells were transfected with all of the plasmids used to prepare effector cells for the cell fusion assay described below, and those effector cell populations were seeded for both fusion and CELISAs so that gH expression could be assessed in the same cell populations used for each fusion assay.
For the original cell fusion assay performed (Table (Table1),1), CHO cells (effectors) were seeded into 96-well plates 18 h before transfection. Effectors were transfected with 30 ng of plasmid expressing gB and 20 ng each of plasmids expressing the T7 RNA polymerase, gD, gL, and either empty vector, WT gH, or mutant gH by using Lipofectamine 2000. In parallel, target cells were seeded into 6-well plates and transfected with 400 ng of plasmid carrying the firefly luciferase gene under the control of the T7 promoter and 1.8 μg empty vector with Lipofectamine 2000. After 2 h of transfection, cells were washed, and target cells were detached and overlaid onto effectors and coincubated for 10 h before lysis and quantification of luciferase signals as described elsewhere previously (26). For the fusion time course assay, effector and target cells were seeded into 6-well plates 18 h prior to transfection. Effector cells were transfected with 600 ng of plasmid containing gB and 400 ng each of plasmids containing T7 polymerase, gD, gL, and either empty vector, WT gH, or mutant gH. Target cells were transfected with 400 ng of plasmid containing luciferase and 1.8 μg of empty vector. After 4 h of transfection, cells were washed, detached, and overlaid as described elsewhere previously (26). After coincubation for the indicated times, cells were washed and lysed, and luciferase activity was measured as described above.
The viral complementation assay was done as described previously for the complementation of a gD-negative virus by WT gD and mutants (27) except that Vero cells were transfected with 1.0 μg of plasmid expressing WT or mutant gH and later infected with gH-negative mutant HSV-1 (30). Virus stocks were prepared, and titrations were performed on Vero-VgHC4 cells. A blue plaque assay was performed after 2 days of incubation at 37°C. Cells were fixed (2% paraformaldehyde [PFA] and 0.2% glutaraldehyde in phosphate-buffered saline [PBS]) and then permeabilized (2 mM MgCl2, 0.01% deoxycholate, 0.02% NP-40) for 5 min and washed twice. Ferricyanide (FeCN) buffer containing 0.5 mg X-gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) per ml was added to cells and incubated about 1 h until blue plaques developed. Plaques were counted, using a dissection microscope.
For analysis of selected mutants, CHO cells seeded into 25-cm2 flasks were transfected with 5 μg each of plasmid expressing WT gL and either empty vector or a plasmid expressing WT gH or a gH mutant and 20 μl Lipofectamine 2000. After 4 h of transfection, cells were washed and incubated with serum-containing medium for 18 h. Cells were then washed with PBS, detached with versene, and counted. Cells were resuspended in 1 ml of lysis buffer (50 mM Tris [pH 8], 150 mM NaCl, 1% Nonidet P-40) containing protease inhibitor per 10 × 106 cells, mixed with sample buffer for SDS-PAGE, and boiled for 5 min. Samples were separated by electrophoresis on 4 to 20% gels, and immunoblots were performed by using rabbit anti-gH-gL antiserum R137 at a 1:10,000 dilution (31), which was kindly provided by Gary H. Cohen and Roselyn J. Eisenberg. For analysis of complemented virus, the supernatant was centrifuged at 48,000 × g for 1 h at 4°C over a 10% sucrose cushion. The pellet was dissolved in sample buffer and separated by SDS-PAGE on a 4 to 20% gel under reducing conditions. Western blot analysis was performed by using R137 at a 1:10,000 dilution and VP5 MAb (EastCoast Bio) at a 1:5,000 dilution.
The 22 linker-insertion gH mutants are listed in Table Table1,1, along with summary data for each mutant. The number and name of the amino acid 5′ of the 5-amino-acid insertion denote the position of each insertion. To assess expression levels of gH mutants on the cell surface, plasmids encoding each mutant were transfected into CHO cells along with plasmids encoding wild-type (WT) HSV-1 gD, gB, and gL. The transfected cells were incubated with two monoclonal antibodies against HSV-1 gH for the detection of gH by cell-based enzyme-linked immunosorbent assay (CELISA). One antibody (52S) recognizes gH regardless of the presence of gL, while the other (53S) is a conformational antibody that needs both gH and gL in order to bind. Figure Figure1A1A shows the locations of the mutants along a linear representation of gH and its proposed or predicted functional domains and levels of cell surface gH detected for each mutant as a percentage of WT gH expression.
Of the 22 mutants screened, 15 were detected on cell surfaces at ≥60% of WT gH levels. The 277S, 386L, 437N, 497I, 666P, 731G, and 740A mutants were detected at lower levels, <40% of WT gH levels (Fig. (Fig.1A).1A). Interestingly, the 386L mutant showed higher levels of expression detected by anti-gH-gL antibody than detected by the anti-gH antibody, suggesting that the 386L insertion site may alter the anti-gH (52S) antibody epitope. For all other mutants, cell surface levels detected by the two antibodies were not significantly different, indicating that complexes with gL were formed even if overall levels of the cell surface complexes were reduced.
Lysates of cells transfected with selected mutants were also probed with anti-gH-gL polyclonal antibody by Western blotting to determine if gH was translated despite low levels of cell surface expression. Figure Figure1B1B shows that all mutants with low expression levels at the cell surface were translated, indicating that their insertions disrupted the posttranslational processing required for transport to and/or expression at the cell surface.
To assess the function of the gH mutants, we first used a quantitative luciferase-based cell-cell fusion assay. All 22 gH mutants were coexpressed with WT HSV-1 gD, gB, and gL in CHO cells (effectors) and mixed with nectin-1-expressing or HVEM-expressing CHO cells (targets). Initially, mutants were screened by the direct transfection of effectors seeded into 96-well plates. This served as a primary screen for the function of mutants and revealed that a number of mutants had cell fusion activity greater than 80% of that of WT gH; some mutants had decreased fusion activity despite near-normal levels of cell surface expression, and all mutants that were expressed on the cell surface at low levels (<40% of WT gH levels) lacked significant cell fusion activity (Table (Table1).1). These mutants with reduced cell surface expression and no fusion activity were not investigated further.
For all other mutants, our secondary screen for function was the same luciferase-based fusion assay performed over a time course. Effector and target cells were transfected as described above, but transfections were scaled to a 6-well plate. After transfection, effector cells were washed, detached, and seeded into 96-well plates in triplicate. Transfected target cells were overlaid onto effector cells at a 1:1 ratio. Coincubation mixtures of effectors and targets were harvested at 5, 8, and 18 h after cell mixing, and the luciferase activity was determined as a measure of cell fusion that had occurred up to each time point.
The first significant observation with this assay was that cells expressing WT HSV-1 glycoproteins induced fusion with nectin-1-expressing target cells at a much higher rate than that observed with HVEM-expressing target cells. Maximal or near-maximal cell fusion was observed with nectin-1-expressing target cells by around 8 h after cell mixing but, with HVEM-expressing target cells, not until some time after 18 h. The decline in luciferase activity at 18 h with nectin-1-expressing targets may be due to the decay of luciferase mRNA or protein after cell fusion had gone to completion. It is not clear whether this receptor-dependent difference in rates relates to the nature of each receptor or to the effective quantity of each receptor expressed on the cell surface of the target cells that we used. However, in a direct comparison of nectin-1 and HVEM, nectin-1 was previously shown to be more efficient at promoting entry than HVEM (24).
gH mutants consistently clustered into three categories based on their fusion activity (Fig. 2A and B). Category 1 mutants were defined as having >50% of the cell fusion activity of WT gH at the 5-h time point with either receptor. These mutants were nearly indistinguishable from WT gH with regard to the kinetics of cell fusion with either receptor (Fig. (Fig.2A).2A). Also, the levels of cell fusion activity at each time point were similar although not identical to those of WT gH (either higher or lower but not lower than 70%). Category 2 mutants were defined as having <50% of the cell fusion activity of WT gH at the 5-h time point. The kinetics of cell fusion for category 2 mutants were distinct from those of WT gH (Fig. 2A and B). These mutants achieved the highest level of fusion with cells expressing either receptor at the 18-h time point (with the exception of the 447L mutant for nectin-1) and the highest level of cell fusion, as a percentage of WT gH cell fusion, at the 18-h time point (Fig. (Fig.2B).2B). The sole category 3 mutant (824L) was incapable of supporting cell fusion with either receptor at any of the time points.
In order to ensure that the decreased rate of fusion for category 2 mutants was not due to a lower rate of cell surface expression, we performed an expression time course. CHO cells were transfected as they were for cell fusion, washed, and allowed to incubate for 5, 8, and 18 h before expression levels were read by CELISA. As shown in Fig. Fig.3,3, only one gH mutant, the 775P mutant, which belonged to category 1 and resembled WT gH in fusion kinetics, exhibited a moderate stepwise increase in expression levels relative to those of WT gH over time. For all category 2 mutants, reduced fusion rates were not due to reduced rates of expression at the cell surface. Similarly, the expression levels of the 824L fusion-defective mutant were similar to those of WT gH at all time points, indicating that the abrogation of functional activity caused by the 824L insertion is not due to decreased expression levels.
In order to explore the effect of the insertions on gH function in viral entry and plaque formation, we used a viral complementation assay in which cells were transfected with WT or mutant gH expression constructs and subsequently infected with gH-null HSV-1. Complemented virus was harvested from the cell lysate (cell-associated virus) and supernatant (extracellular virus). Complemented viral titers were determined by plaque assay, and the level of gH incorporation into virions was determined by Western blotting.
Of particular interest were the mutants that were defective in fusion or that exhibited a reduced rate of cell fusion. As shown in Fig. Fig.4A4A for the category 2 mutants, a decreased rate of fusion correlated with decreased, but not absent, viral infectivity. Furthermore, the reduced viral infectivity does not seem to be due to a decreased incorporation of gH into the virions (Fig. (Fig.4B).4B). The nonfunctional 824L mutant was as defective for viral infectivity as the gH mutants that were not properly processed (Table (Table1)1) despite normal levels of gH incorporation into virions. Collectively, these data suggest that the regions of gH identified as described above to be important for rates of cell-cell fusion, or for detectable cell fusion activity, are also important for viral infectivity.
We generated a panel of HSV-1 gH insertion mutants in order to further identify gH functional domains. Fifteen gH mutants were expressed at the cell surface at levels greater than or equal to 60% of WT gH levels. We screened for the function of our gH mutants with a luciferase-based cell-cell fusion assay standard to the field but modified the assay by monitoring the extent of cell fusion as a function of time after mixing the effector and target cells. The first interesting finding revealed by our time course was that fusion with nectin-1-expressing target cells occurred more rapidly than that with HVEM-expressing target cells. Nectin-1-triggered fusion is so rapid that the standard effector and target cell coincubation time of 18 h or more is likely too long to accurately reflect the function of mutations with intermediate phenotypes. It is not clear whether the receptor-dependent difference in the rate of fusion that we observed relates to the nature of each receptor or to the effective quantity of each receptor expressed on our target cells.
Our study also revealed that a number of insertion mutations lowered the rate of cell-cell fusion without abrogating fusion altogether. Importantly, the decreased rate of cell-cell fusion was not due to a decreased rate of cell surface expression. The decreased rate of fusion observed for these mutants suggests that gH may have a role in governing the kinetics of fusion and may be responsible for a rate-limiting first stage in HSV-1 fusion. However, we acknowledge that our experiments measure fusion events within a population of cells and not the rate of individual fusion events. If we embrace the current working model of HSV fusion in which gH is responsible for hemifusion, our data suggest that the hemifusion process may be more easily slowed than shut down altogether. It is also possible that the mutations that affect rate of fusion, all of which are located in the gH ectodomain, interrupt but do not totally abolish the gH-gD and/or gH-gB interactions necessary for fusion.
It is interesting that of the quartet of required entry glycoproteins, gD and gH are more tolerant of insertion mutations than the presumed fusogen gB. In the linker-insertion mutagenesis analysis reported here, 68% of mutants were expressed at the cell surface. A previous study also reported a panel of gH insertional mutants in which between 57 and 76% of mutants were expressed normally, depending on the antibody used to detect expression (11). In similar studies done with gD (7) and gB (26), cell surface expression was achieved for 86% and 33% of mutants, respectively. This difference between the panels of gH and gB insertion mutants generated in an identical fashion may indicate that gH, unlike gB, has a more flexible conformation and performs functions that are not so dependent on intrachain interfaces or dramatic conformational changes.
One of our insertion mutations, 824L, was incapable of performing cell fusion or complementing a gH-null virus despite normal cell surface expression and viral incorporation. The 824L gH mutant clearly indicates a crucial role for the short cytoplasmic tail of gH in both cell fusion and viral entry. Two previous studies demonstrated the importance of 831V, also located within the gH cytoplasmic tail, downstream of our 824L insertion mutant, to the fusion activity of a syncytial strain of HSV-1 (4, 43). An alanine substitution at 831V decreased the viral entry efficiency of a syncytial strain of complemented virus but did not abrogate viral fusion altogether, as our 824L mutant did. It is possible that our 824L insertion mutant and the previously identified 831A mutant are disrupting the same required functional domain of gH, although they may also be distinct functional domains. Regardless, the precise role of the short cytoplasmic tail of gH in fusion and viral entry remains to be defined. Since soluble gD is sufficient for triggering fusion, we can conclude that interactions between gH and gD via their cytoplasmic domains are not required for fusion (8). Similarly, it seems unlikely that the cytoplasmic domain of gH is interacting with gB, although perhaps an interaction between gH and gB via their cytoplasmic tails is required for fusion. It seems more likely that the gH cytoplasmic tail is mediating the multimerization of gH or the gH association with soluble proteins or is required for gH conformation and/or stability. Based on the normal cell surface expression and viral incorporation of the 824L mutant, it seems unlikely that the functional defect of 824L is due entirely to decreased protein stability.
It may be significant that the 824L mutation abrogated membrane fusion entirely, whereas the 831A mutation did not. Perhaps the most membrane-proximal region of the cytoplasmic tail, which is disrupted in the 824L mutant but not in the 831A mutant, is interacting with viral envelope or plasma membrane lipids and plays a direct role in the perturbation of the inner leaflet of the bilayer that is required for fusion to proceed. As others noted previously (43), gH proteins of the alphaherpesviruses HSV-1 and -2, equine herpesviruses 1 and 4, bovine rhinotracheitis virus type 1, varicella-zoster virus, and pseudorabies virus all contain two neutral, nonaromatic amino acids (ML or VL) in the most-membrane-proximal region of the cytoplasmic tail. Based on sequence prediction, the 824L mutation may extend the transmembrane of gH by two amino acids (9), thus perhaps pushing the conserved membrane-proximal VL amino acids downstream. Further studies are needed to define the functional regions of the gH cytoplasmic tail and their specific functions in membrane fusion.
We thank Gary H. Cohen and Roselyn J. Eisenberg for the generous gift of the R137 antibody and Nanette Susmarski for providing cells and her cell line expertise. We thank the members of the Longnecker, Spear, and Jardetzky laboratories for help and support.
This research was supported by grants CA117794 (R.L.) and CA021776 (P.G.S. and R.L.) from the National Cancer Institute.
Published ahead of print on 9 December 2009.