Structure-Function Analysis of Herpes Simplex Virus Type 1 gD and gH-gL: Clues from gDgH Chimeras

Abstract

In alphaherpesviruses, glycoprotein B (gB), gD, gH, and gL are essential for virus entry. A replication-competent gL-null pseudorabies virus (PrV) (B. G. Klupp and T. C. Mettenleiter, J. Virol. 73:3014-3022, 1999) was shown to express a gDgH hybrid protein that could replace gD, gH, and gL in cell-cell fusion and null virus complementation assays. To study this phenomenon in herpes simplex virus type 1 (HSV-1), we constructed four gDgH chimeras, joining the first 308 gD amino acids to various gH N-terminal truncations. The chimeras were named for the first amino acid of gH at which each was truncated: 22, 259, 388, and 432. All chimeras were immunoprecipitated with both gD and gH antibodies to conformational epitopes. Normally, transport of gH to the cell surface requires gH-gL complex formation. Chimera 22 contains full-length gH fused to gD308. Unlike PrV gDgH, chimera 22 required gL for transport to the surface of transfected Vero cells. Interestingly, although chimera 259 failed to reach the cell surface, chimeras 388 and 432 exhibited gL-independent transport. To examine gD and gH domain function, each chimera was tested in cell-cell fusion and null virus complementation assays. Unlike PrV gDgH, none of the HSV-1 chimeras substituted for gL for fusion. Only chimera 22 was able to replace gH for fusion and could also replace either gH or gD in the complementation assay. Surprisingly, this chimera performed very poorly as a substitute for gD in the fusion assay despite its ability to complement gD-null virus and bind HSV entry receptors (HveA and nectin-1). Chimeras 388 and 432, which contain the same portion of gD as that in chimera 22, substituted for gD for fusion at 25 to 50% of wild-type levels. However, these chimeras functioned poorly in gD-null virus complementation assays. The results highlight the fact that these two functional assays are measuring two related but distinct processes.

For many alphaherpesviruses, four glycoproteins are required for virus entry into mammalian cells (32, 52, 53). In the case of herpes simplex virus (HSV) and bovine herpesvirus type 1, gB, gD, and the gH-gL heterodimer function independently or in concert to effect fusion of the virion envelope with the plasma membrane. All four glycoproteins are essential for the spread of these viruses from cell to cell and for cell fusion (2, 35, 46, 57). However, little is known about the mechanism of these processes. Interestingly, although the same set of four glycoproteins are required for pseudorabies virus (PrV) entry, gD and gL are not essential for cell-cell spread (24, 25). Thus, HSV and PrV glycoproteins have evolved to have somewhat different functions.

Klupp and Mettenleiter passaged a gL-null virus in cell culture and obtained a virus, PrV-ΔgLpass, that was able to enter cells, replicate, and spread from cell to cell (24). PrV-ΔgLpass had undergone gene rearrangements such that it lacked wild-type (WT) gH and instead contained a gDgH chimera resulting from a fusion between the first 271 amino acids of gD and the C-terminal 590 amino acids of gH. Moreover, the gDgH chimera was able to substitute for gH, gD, and gL in both virus entry and cell-cell fusion assays (24, 25). The authors stated, “It will be interesting to analyze a similar fusion protein in an HSV-1 background to test whether the different properties of the gH proteins influence the outcome of this experiment.”

We decided to investigate this matter in HSV. However, because both gD and gL are required for cell spread of HSV, we could not employ a similar selection method to obtain a gDgH fusion protein for HSV. Therefore, we used available structure-function information about the two proteins (5, 24, 44) and constructed several best-guess hybrids to mimic the PrV gDgH protein. In choosing where to truncate each protein, we were faced with the fact that the homology between the HSV and PrV homologues of gD and gH is poor. Klupp and Mettenleiter suggested that residue 271 (after the signal sequence) of PrV gD correlates with residue 295 of HSV type 1 (HSV-1). Therefore, we decided to truncate gD downstream of this residue. Our earlier studies suggested that gD306t is able to bind HSV receptors and to block virus infection (26, 39, 55, 59). Thus, our first constructs contained residues 1 to 308 of gD. For gH, we relied on the concept that the HSV structure was dependent upon disulfide bond arrangement and that cleaving between a disulfide pair would structurally alter the molecule (31). It was previously hypothesized that, within the first 432 residues, the disulfide bond arrangement of HSV gH consists of a linkage of cysteine 1 to cysteine 2 and cysteine 3 to cysteine 4 (44). Therefore, the constructs contained full-length gH (beginning at residue 22, after the signal peptide); gH beginning at residue 259 or 388, lacking cysteines 1 and 2; and gH beginning at residue 432, lacking cysteines 1 to 4 (Fig. 1).

FIG. 1.

Schematic diagram showing the four gDgH chimeras compared to WT HSV-1 gD and gH. Chimeras contain the first 308 amino acids of gD (black box), including the gD signal sequence (sig), fused to N-terminal truncations of gH (white box). Chimeras are abbreviated with the amino acid number of gH at which each is truncated. All chimeras retain the gH transmembrane domain (TM), represented by a gray box. Cysteines within gH thought to be involved in disulfide bond formation are denoted by “C.”

Of these four chimeras, only the one containing full-length gH was able to facilitate virus entry in the absence of authentic gD or gH. Although two other chimeras could substitute for gD in at least one functional assay, none were able to substitute for gH. In every case, gL was also required. Interestingly, certain mutants supported virus entry better than they did cell fusion (and others vice versa), implying that these two processes use somewhat different mechanisms.

MATERIALS AND METHODS

Plasmid DNAs.

The US6 gene encoding WT HSV-1 gD (KOS) was PCR amplified from infected Vero cells with primers 5′-CGCGGAAGCTTCCGGTATGGGGGGGGCT and 5′-CGCCCTCGAGTATCTAGTAAAACAAGGGCT. By the use of the underlined restriction digest sites, the gD fragment was digested and ligated with pcDNA3.1 vector to generate the plasmid pSC390. Construction of the plasmids pCMV3gL-1, encoding HSV-1 gL, and pCMV3gH1 (pHC138), encoding HSV-1 gH, was described previously (10). The HSV-1 gDgH chimeras were created by PCR amplification of pSC390 and pHC138. gD was amplified with primers encoding the restriction sites (underlined) for KpnI (5′-GGGGTACCATGGGGGGGACTGCTGCCAGGTTG) and BamHI (5′-GCAGGATCCGGATGATACGGCGTCGCGGCGTC for gD308t and 5′-ATTTGCGGCGGATCCGTCCCCACGGGGTC for gD285t). The KpnI-BamHI-digested gD-containing PCR product was then cloned into vector pcDNA3.1 to generate either plasmid pDT332 (gD308t) or pTC477 (gD285t). Primers containing the BglII restriction site (chimera 22, 5′-GCCAGATCTCACTGAGCAGACAGATCCCTGGTT; chimera 259, 5′-GCTAGATCTTGCCCTGGTGCGAGCACGCTAC; chimera 388, 5′-GGAAGATCTAGGATTCGCATTCGTGAACGCCGCACACGCAAA; chimera 432, 5′-GGCGAGATCTTGTGTTCTTCAATGTGTCAGTCTT) and the XhoI restriction site (5′-CCCCTCGAGTTATTCTCGTCTCCAAAAAAA) were used to amplify the gH gene. These gH-containing BglII-XhoI-digested PCR fragments were then ligated with BglII-XhoI-digested pDT332 to generate plasmids pTC430 (chimera 22), pDT349 (chimera 259), pTC431 (chimera 388), and pTC432 (chimera 432). All of these plasmids contain sequences encoding the first 308 amino acids of gD fused to the gH sequences. The plasmid pTC477 was digested and ligated similarly to the gH-containing fragments to generate plasmids pTC478 (chimera t22) and pTC479 (chimera t388), which contain sequences encoding the first 285 amino acids of gD linked to gH sequences. All plasmids were sequenced to ensure the absence of PCR errors. The SacI-XhoI fragments from pTC430, pDT349, pTC431, pTC432, pTC478, and pTC479 were ligated into vector pCAGGS/MCS (41) to generate plasmids for use in the fusion assay: pTC441 (chimera 22), pTC442 (chimera 259), pTC443 (chimera 388), pTC444 (chimera 432), pTC496 (chimera t22), and pTC497 (chimera t388). Plasmids pT7EMCLuc (encoding the firefly luciferase gene) and pCAGT7 (encoding T7 polymerase), as well as pPEP98, pPEP99, pPEP100, and pPEP101 (encoding the genes for HSV-1 gB, gD, gH, and gL, respectively) were gifts of P. G. Spear (41, 46).

Cells and viruses.

293T cells were grown in Dulbecco modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS). African green monkey kidney (Vero) cells were grown in 5% FBS-DMEM. CHO-K1 cells were grown in Ham's F-12 medium containing 10% FBS. The CHO cell lines CHO-HVEM12, expressing the HSV receptor HveA (56), and CHO-R3A, expressing nectin-1 (15), were grown in 10% FBS-F-12 containing 250 μg of G418/ml. CHO-K1 and CHO-HVEM12 cells were kindly provided by P. Spear. Propagation of the gD-null virus FgDβ on VD60 cells (gifts of D. C. Johnson) and the gH-null virus SCgHZ on F6 cells (gifts of A. Minson) was as previously described (12, 29).

Antibodies.

Polyclonal antibodies (PAbs) used in this study were as follows: rabbit (R) serum R7 was raised against HSV-2 gD and cross-reacts with HSV-1 gD (22), R137 was prepared against purified gHt-gL (44), and rabbit antibodies anti-UL1-1 and anti-UL1-2 (gifts of D. C. Johnson) were prepared against peptide sequences of gL (21). gD monoclonal antibodies (MAbs) used that recognize linear epitopes were as follows: 1D3 (within gD residues 11 to 19) (6, 7) and DL6 (within gD residues 272 to 279) (22). gD MAbs used that recognize discontinuous (conformational) epitopes were as follows: type 1-specific antibodies 41S and 45S (obtained from the American Type Culture Collection [ATCC]) (50); type common antibodies HD1 (36, 45), DL11 (8, 36), I-99 (42), 11S (ATCC) (50), and AP7 (6, 34). The anti-gH MAb H12 recognizes a linear epitope within amino acids 475 to 648 (44). MAbs that recognize discontinuous epitopes of gH or gH-gL were as follows: 52S and 53S (ATCC) (50), 46S (58), and LP11 (gift of A. Minson) (3).

Immunoprecipitation assays.

293T cells were transfected with the desired plasmids according to the GenePORTER protocol (Gene Therapy Systems, Inc.). At 48 h posttransfection, cells were lysed in 10 mM Tris (pH 8)-150 mM NaCl-10 mM EDTA-1% NP-40-0.5% deoxycholic acid-1 mM phenylmethylsulfonyl fluoride. Typically, 5% of the total cell extract (from a 35-mm-diameter plate) was incubated with the appropriate antibody for 18 h at 4°C. Proteins were precipitated with protein A agarose beads (Gibco BRL) for 2 h at 4°C, separated by electrophoresis on a sodium dodecyl sulfate (SDS)-12% polyacrylamide gel, and detected by Western blotting with the desired counterantibody.

CELISA.

To detect gDgH cell surface expression, we used the cellular enzyme-linked immunosorbent assay (CELISA) (9, 14, 33). Cells growing on 35-mm-diameter plates were transfected with an appropriate plasmid via the GenePORTER protocol and then reseeded 24 h later onto 96-well plates (pretreated with 0.2% gelatin-phosphate-buffered saline [PBS]). At 48 h posttransfection, cells were fixed with 3% paraformaldehyde for 20 min, quenched with 50 mM NH₄Cl for 10 min, and then incubated for 2 h with various concentrations of anti-gD PAb (R7) diluted in 5% FBS-DMEM, all at room temperature. Secondary antibody goat anti-rabbit coupled to horseradish peroxidase was added, and bound antibody was detected with ABTS [2,2′-azinobis(3-ethylbenzthiazolinesulfonic acid)] peroxidase substrate (Moss, Inc.).

ELISA.

293T cells growing on 35-mm-diameter plates were transfected by the GenePORTER method and then lysed at 48 h posttransfection. Proteins from cell lysates were separated by electrophoresis on an SDS-12% polyacrylamide gel, probed with the anti-gD MAb 1D3, and quantified by laser densitometry. Equivalent amounts of protein from the cellular extracts were added to wells of a 96-well ELISA plate pretreated with either 4 μg of HveA(200t)/ml or 10 μg of HveC(346t)/ml purified protein (26, 59) and blocked with 5% milk-PBS containing 0.2% Tween 20. Baculovirus-produced and purified gD306t (51) was suspended in lysis buffer and used to test for any effects of detergent on gD-receptor binding. Proteins bound to receptor were detected with ABTS substrate through PAb R7 and goat anti-rabbit-horseradish peroxidase binding.

Fusion assays.

The luciferase reporter gene activation assay was previously described by Okuma et al. (41) and Pertel et al. (46). Briefly, CHO-K1, CHO-HVEM12, and CHO-R3A cells growing in six-well plates were transfected by the Lipofectamine protocol (Invitrogen). Cells that were transfected with plasmids encoding HSV glycoproteins and the T7 RNA polymerase (0.4 μg of each/well) were designated as effector cells. CHO-HVEM12 and CHO-R3A cells that were transfected with the plasmid encoding luciferase under the control of the T7 promoter (0.4 μg/well) were designated as target cells. At 8 h posttransfection, cells were refed with 10% FBS-F-12 medium. Four hours later, target and effector cells were cocultivated into 24-well plates at a 1:1 ratio (targets to effectors). At 20 h postcocultivation, cells were washed once with PBS and then lysed in 100 μl of 1× reporter lysis buffer (Luciferase Assay System; Promega) per well. Cell lysates were incubated at −20°C for at least 18 h. Finally, 20 μl of lysate from each sample was mixed with 100 μl of luciferase substrate (Promega) and immediately assayed for light output by luminometry.

To test for syncytium formation, 24-well plates containing mouse melanoma cells expressing nectin-1 (designated C10) were transfected with 1 μg of total DNA of the appropriate pCAGGS/MCS plasmids (0.25 μg of each of four plasmids total). C10 cells were used in place of CHO cells for this assay because they tended to form more extensive syncytia under WT conditions and therefore were easier to observe and photograph. At 26 h posttransfection, cells were fixed in methanol, stained with Giemsa stain (Gibco BRL) for 10 min, and scored for syncytium formation (a qualitative comparison of the positive-control well to the experimental sample wells). Transfected CHO and C10 cells were tested for cell surface protein expression by CELISA.

Complementation assay.

The complementation protocol was adapted from those of Muggeridge et al. (37) and Chiang et al. (6), with some modifications. A 12-well plate containing Vero cells was transfected by the GenePORTER protocol overnight and then refed the next morning with 5% FBS-DMEM and allowed to recover for several hours. Next, cells in each well were infected with 10⁶ PFU of FgDβ or SCgHZ in 200 μl of serum-free DMEM at 37°C. After 1 h, 1 ml of 5% FBS-DMEM was added and incubation was continued for an additional hour, after which the virus was inactivated by incubation with 40 mM sodium citrate (pH 3.0)-10 mM KCl-135 mM NaCl for 1 min. The acid was removed, fresh medium was added, and cells were incubated overnight. At 24 h postinfection, total virus was collected by freeze-thawing the cells on the plate three times and then scraping the cell debris and medium into a 15-ml conical tube. Samples were centrifuged at 2,500 rpm in a Beckman GPR centrifuge for 10 min to remove cell debris. For gH-null complementation, virus-containing supernatants were serially diluted and their titers were determined on F6 cells. For gD-null complementation, the virus titer was determined on VD60 cells. One hour postinfection, F6 and VD60 cells were overlaid with DMEM containing 1% carboxymethyl cellulose and 5% FBS and incubated for an additional 3 days. Plaques were scored after cells were fixed by using 5% formaldehyde-PBS and stained with crystal violet.

RESULTS

gDgH protein expression and epitope recognition.

Work done with PrV demonstrated that the requirement for gL during virus entry could be superseded by a gDgH chimeric protein (24, 25, 49). We questioned whether a similar protein could be constructed for HSV and if the minimal protein domains needed for entry would mimic those of PrV. Several different chimeric proteins were synthesized, each containing progressively shorter regions of gH through N-terminal truncations (Fig. 1). All chimeras contained the gD signal sequence, the first 308 amino acids of gD, and the gH transmembrane domain and cytoplasmic tail.

To examine the antigenic profile of each chimera, 293T cells were transfected with the various plasmids. These cells were chosen because of the high level of protein expression seen upon transfection. Cytoplasmic extracts of cells expressing each chimera were immunoprecipitated by MAbs that recognized discontinuous epitopes to either gD or the C-terminal half of gH. The eight gD MAbs used in these experiments represent six different antigenic group classifications. Representative blots from these experiments are shown in Fig. 2, and an experimental summary is found in Table 1. Each chimera was immunoprecipitated by all gD MAbs, as well as by anti-gH MAbs 46S, 52S, and H12, and all migrated at the expected position predicted by their molecular size. The data strongly suggest that the gD and gH portions of each chimera are properly folded.

FIG. 2.

Protein profiles from transfected cells expressing the gDgH chimeras or WT HSV proteins. Representative blots depict results obtained when cellular lysates from transfected 293T cells were immunoprecipitated with either the anti-gH MAb 52S (A) or the anti-gD MAb DL11 (B). Upon transfer to nitrocellulose, blots were subsequently probed with either an anti-gH or an anti-gD PAb, respectively. For comparison, panel C shows a Western blot of both 293T and Vero cell lysates probed with anti-gH-gL PAb R137. Numbers to the left of panels A and B and to the left and right of panel C are molecular masses in kilodaltons.

TABLE 1.

Immunoprecipitation of gDgH chimeras with and without gL

Immunoprecipitation antibody	Result for gDgH chimera:
	22	22-gL	259	259-gL	388	388-gL	432	432-gL
gH-gL
LP11	−	+	−	−	−	−	−	−
53S	−	+	−	−	−	−	−	−
gH (cfa)
46	+	+	+	+	+	+	+	+
52S	+	+	+	+	+	+	+	+
gH
H12	+	+	+	+	+	+	+	+
gL
L4	NDb	+	ND	+	ND	−	ND	−
UL1-1	−	−	−	+	−	−	−	−
UL1-2	−	+	−	+	−	−	−	−
gD (cfa)
AP7	+	+	+	+	+	+	+	+
DL11	+	+	+	+	+	+	+	+
HD1	+	+	+	+	+	+	+	+
I-99	+	+	+	+	+	+	+	+
11S	+	+	+	+	+	+	+	+
41S	+	+	+	+	+	+	+	+
45S	+	+	+	+	+	+	+	+
gD
DL6	+	+	+	+	ND	ND	ND	ND
1D3	+	+	+	+	+	+	+	+

^acf, antibody to a conformational epitope.

^bND, not determined.

Association of gL with gDgH chimeras.

Certain antibodies to conformational epitopes of the gH N terminus (LP11 and 53S) require the presence of gL (19, 21). That said, only chimera 22, containing the complete gH N terminus, was immunoprecipitated with the MAbs LP11 and 53S, and that occurred only when chimera 22 was coexpressed, indicating that chimera 22 binds gL (Table 1). Coexpression of the chimeras with gL did not affect protein expression, with one exception. Upon coexpression of chimera 22 and gL, there was a second prominent protein band that migrated just below the full-length chimera on an SDS gel (Fig. 2A). The identity of this second major protein band is currently unknown, but it most likely contains very little or no gD, since it could not be visualized with antibodies to gD (Fig. 2B). More of this protein was expressed in 293T cells than in Vero cells (Fig. 2C). This is an important point because the complementation assays are carried out in Vero-derived F6 and VD60 cells.

To test specifically for gL binding (independent of gH-gL conformation), each chimera was immunoprecipitated with UL1-2, a gL PAb directed against gL C-terminal sequences (Fig. 3A, top panel), and then probed via Western blotting with H12, an anti-gH antibody. Chimeras 22 and 259 were both immunoprecipitated with UL1-2, suggesting an interaction between the chimera and gL. The two shortest chimeras (388 and 432) were not captured by the antibody. Taken together, these data suggest that a major site of gH-gL association lies between gH amino acids 259 and 388.

FIG. 3.

Association of gL with gDgH chimeras. (A) Immunoprecipitation of proteins with anti-gL PAb UL1-2 (top) or UL1-1 (bottom), followed by Western blotting with anti-gH MAb H12. (B) Working model of WT HSV-1 gH-gL interaction (44). gH is shown in gray, and gL is shown in black, with the C termini labeled COO⁻. The complex is anchored to the membrane through the gH transmembrane region (dotted box). The hatched boxes on either end of gL represent the peptides used to generate the UL1-1 and UL1-2 PAbs. The position of gH residue 259 is indicated.

In the proposed model for gH-gL structure (Fig. 3B), the interaction between the two proteins occurs at the N termini of both. Unlike UL1-2, the peptide antibody UL1-1 (raised against gL residues 26 to 44) was unable to immunoprecipitate the WT gH-gL complex (Fig. 3A, bottom panel) (47). Therefore, gH appears to obscure the gL N terminus upon binding. Chimera 22 behaved like WT gH and was not immunoprecipitated by UL1-1. In contrast, UL1-1 immunoprecipitated chimera 259 in the presence of gL. We interpret these observations to mean that amino acids 1 to 258 of gH mask the gL N terminus from antibody.

gDgH cell surface protein expression.

Since maturation and transport of gH depend on coexpression of gL (21, 47), it is not clear whether HSV gL functions merely as a gH chaperone or whether it has a function of its own in virus entry and fusion. In the case of PrV, it appears that gL normally functions only as a gH chaperone, since the gDgH chimera functions independently of gL. Based on the PrV data, we asked whether an HSV gDgH protein could reach the cell surface in the absence of gL. Vero cells were transfected with plasmids encoding WT or chimeric genes, transferred to a 96-well plate, incubated with fivefold dilutions of anti-gD antibody, and then evaluated by CELISA (Fig. 4). Chimeras 388 and 432 were detected on the surface of Vero cells in the absence of gL (Fig. 4A). When chimera 22 was expressed alone, low levels of it were found on the cell surface (Fig. 4B). However, much greater surface expression was detected when gL was also present. Chimera 259 was not detected on the surface regardless of whether gL was present.

FIG. 4.

gDgH cell surface expression by CELISA. Transfected Vero cells were fixed with 3% paraformaldehyde and then incubated with fivefold serial dilutions of the anti-gD PAb R7 (x axis). Results for chimeras 388 and 432 (expressed with and without gL) are shown in panel A, while chimeras 22 and 259 are shown in panel B. OD₄₀₅, optical density at 405 nm.

Functional analysis of the gH domain of the chimeras.

With the basic characteristics of each chimera established, each chimera was then tested for its ability to substitute for gH in both virus entry and cell-cell fusion. Chimera 259 was negative in all assays of gD and gH function (data not shown), most likely because it was not presented efficiently on the cell surface (Fig. 4B). Therefore, this chimera will not be mentioned further.

Virus entry was examined by a gH-null virus complementation assay. Vero cells were transfected with plasmids expressing either WT gH or one of the chimeras and then infected 18 h later with SCgHZ (gH-null) virus. Total virus was collected by freeze-thaw and assayed for plaque formation on F6 cells (which provide gH in trans upon infection with HSV). As seen in Fig. 5A, only chimera 22 was able to complement gH-null virus and allow entry into F6 cells at near-WT virus levels. This finding shows that full-length gH is functional for virus entry within the context of a gDgH chimera. In contrast, viruses complemented with chimera 388 or 432 were unable to form plaques on F6 cells.

FIG. 5.

Examination of gH functions. (A) gH-null virus complementation was performed as described in Materials and Methods. Each sample was assayed in three separate experiments, and the average value (with standard deviation as represented by error bars) was plotted on the graph. The percentage of WT activity was calculated as follows: (sample titer/WT titer) × 100. (B) Cell-cell fusion in the absence of gH was performed as described in Materials and Methods. Target CHO cells (expressing the luciferase protein and either HveA or nectin-1) were cocultivated with effector CHO cells (expressing T7 polymerase, gB, gD, and gL, plus either gH, gDgH, or empty vector DNA) and tested for light production 20 h later. The percentage of WT activity was calculated as for panel A. Shown is one assay, done in triplicate for each sample. (C) Observation of syncytium formation. C10 cells were transfected with plasmids expressing gB, gD, and gL plus either WT (gH+), vector (gH−), or chimera 22. Cells were then fixed with methanol and stained with Giemsa stain. Prominent syncytia are denoted by white arrows.

We next asked whether any of the chimeras could function in a cell-cell fusion assay. For HSV-1, expression of gB, gD, gH, and gL and a gD receptor on the cell surface is sufficient to support cell-cell fusion (2, 35, 46, 57). The fusion assay has been compared to the virus-cell fusion event during entry and to the cell-cell fusion seen with syncytial mutants (53). Also, the fusion assay allows us to determine if each mutant has a receptor-specific fusion phenotype, since the cells used here express either HveA or nectin-1 as their sole HSV receptor. Using a luciferase reporter assay system (41, 46), we tested the HSV chimeras for their fusogenic properties on receptor-expressing target cells, consisting of CHO-HVEM12 cells (HveA) or CHO-R3A cells (nectin-1) that were also transfected with a plasmid encoding the luciferase protein under the control of the T7 promoter. Effector cells were prepared by cotransfection of CHO cells with plasmids containing the genes for T7 polymerase, gB, gD, gL, and one of the chimeras (in place of gH).

Target and effector cells were mixed, incubated for 20 h, lysed, and then assayed for light production (i.e., fusion). In the absence of WT gH, only chimera 22 enabled cell-cell fusion when expressed together with gB, gD, and gL (Fig. 5B). Effector cells expressing proteins from this set of plasmids produced a signal on target cells expressing HveA or nectin-1 at levels that were 10 and 18% of the WT level, respectively. These signals, although diminished compared to WT gH, were significantly above background levels.

We wanted to visually confirm that the low level of luminescence really reflected cell-cell fusion by examining syncytium formation. Since syncytium formation on CHO cells was difficult to discern, mouse melanoma cells expressing the HSV receptor nectin-1 (B78H1-C10 cells) were used instead. C10 cells were transfected with plasmids expressing three of the four essential glycoproteins plus either WT gH, the desired chimera, or vector DNA. The cells were then stained with Giemsa stain, and the monolayer was scanned for the presence of syncytia. Plates containing syncytia with five or more nuclei were scored as positive for fusion. Cells transfected with chimera 388 or 432 formed no detectable syncytia (data not shown). However, syncytia were detected when C10 cells were transfected with plasmids expressing chimera 22, gB, gD, and gL (Fig. 5C). These syncytia were noticeably smaller (containing fewer nuclei) and were fewer than those formed when cells were transfected with the plasmid for WT gH along with gB, gD, and gL. These observations show that chimera 22 can substitute for gH in cell-cell fusion. Moreover, the syncytial data account for the low signal seen in the luciferase assay (Fig. 5B).

We next tested whether the chimeras could substitute for gL in the fusion assay by mixing HveA or nectin-1 target CHO cells with CHO effectors expressing gB, gD, gH, and one of the chimeras. None of the chimeras produced detectable light in the luciferase assay (Fig. 6), nor were any syncytia seen when transfected C10 cells were stained with Giemsa stain (data not shown). These results provide a clear distinction between the HSV gDgH chimeras and the PrV gDgH protein, as the latter protein replaced gL in several functional assays (11, 24, 25, 49). It was disappointing that chimera 22 was not adequately presented on the cell surface without gL, because we are precluded from assigning a clear function to HSV-1 gH or gL. For chimeras 388 and 432, their inability to function in the absence of gL cannot be explained by a protein transport problem, since they could reach the cell surface without gL. It is possible that these chimeras lack the gH sequences essential for fusion. Alternatively, gL may be required for fusion in its own right (40), or fusion might require both proteins.

FIG. 6.

Cell-cell fusion in the absence of gL was as described for Fig. 5, except that effector CHO cells were transfected to express T7 polymerase, gB, gD, and gH, plus either gL, gDgH, or empty vector DNA.

Functional analysis of the gD domain of the chimeras.

Each of the gDgH chimeras was tested as a substitute for gD in three assays of gD function, namely, gD-null virus complementation, receptor binding, and cell-cell fusion. Because each chimera displayed a remarkably different phenotype, we will discuss the results obtained for each chimera separately.

Analysis of chimera 22.

We first tested virus entry by the gD-null virus complementation assay. The assay was performed essentially the same way as for gH-null virus complementation, except that the gD-null virus FgDβ and VD60 complementing cells were used. Chimera 22 was able to substitute for gD in the complementation assay at 90% of WT gD levels (Fig. 7A). This observation demonstrates that, even when gD lacks the cytoplasmic tail, transmembrane domain, and a portion of the ectodomain (amino acids 309 to 320), it is functional for virus entry.

FIG. 7.

Functional analysis of the gD domain of chimera 22. (A) The gD-null virus complementation assay was as described in Materials and Methods and the Fig. 5 legend. (B) Receptor binding properties as determined by ELISA. Extracts from transfected 293T cells were normalized by Western blotting with anti-gD MAb 1D3 and subsequent densitometry. Serial twofold dilutions of extract (x axis) were incubated with purified HveAt or HveCt (nectin-1) bound to wells of a 96-well plate. Bound protein was detected with PAb R7. Values for mock-transfected cells, denoting the assay background, were subtracted from each data point. OD₄₀₅, optical density at 405 nm. (C) Cell-cell fusion in the absence of gD followed the same basic protocol as described for Fig. 5, except that effector CHO cells were transfected to express T7 polymerase, gB, gH, and gL, plus either gD, gDgH, or empty vector DNA.

To examine the binding of chimera 22 to receptor, extracts of transfected 293T cells were incubated with truncated forms of purified HveA or nectin-1 protein immobilized on wells of a 96-well plate. Binding of the chimeric proteins to receptor was detected by ELISA with a PAb to gD. Chimera 22 and WT gD bound equally well to HveA (Fig. 7B, top panel). However, unlike WT gD, chimera 22 bound poorly to nectin-1 (Fig. 7B, bottom panel), perhaps because a portion of full-length gH interfered with the nectin-1 binding site on gD. The addition of gL did not alter the binding of chimera 22 to either receptor (data not shown).

In contrast to complementation, chimera 22 was a poor substitute for WT gD in cell-cell fusion, yielding only 8% of the luciferase activity seen for WT gD on CHO cells expressing HveA (Fig. 7C, top panel). This result was surprising considering the fact that chimera 22-gL was efficiently expressed on the cell surface (Fig. 4B) and was able to bind HveA at WT levels (Fig. 7B). Fusion of effector cells bearing gB, gH, gL, and chimera 22 with target cells expressing nectin-1 (Fig. 7C, bottom panel) was also very poor (4% of WT level) but is understandable in light of the poor binding of chimera 22 to nectin-1 (Fig. 7B). The HveA data suggest that a gD function other than receptor binding is impaired when it is fused to gH.

Analysis of chimeras 388 and 432.

Next, chimeras 388 and 432 were tested in each of the gD assays. These chimeras complemented gD-null virus poorly, at a level that was just 7% of the level seen for WT gD (Fig. 8A). Since both chimeras bound HveA and nectin-1 similarly to WT gD (Fig. 8B), the poor complementation was not due to a binding defect. However, both chimeras were able to support fusion when tested as a surrogate for gD, at levels of 35 to 50% of WT activity on HveA-expressing CHO cells and 20 to 25% of WT activity on nectin-1-expressing CHO cells (Fig. 8C). This finding was intriguing, as the opposite situation was seen with chimera 22, which complemented gD-null virus but was poor as a replacement for gD for fusion.

FIG. 8.

Functional analysis of the gD domain of chimeras 388 and 432. Complementation of gD-null virus (A), binding of the chimeras to cellular receptors (B), and cell-cell fusion in the absence of gD (C) were as described for Fig. 7. OD₄₀₅, optical density at 405 nm.

gD285 constructs.

As seen in Fig. 7, chimera 22 did not bind nectin-1 as efficiently as did WT gD. It is known that when gD is truncated to amino acid 285 its affinity for both HveA and nectin-1 is dramatically increased (27, 48, 61). We wondered whether a chimera consisting of gD285 fused to full-length gH (t22, Fig. 9A) would show enhanced binding to nectin-1 or enhanced fusion activity or whether it would complement gD-null virus. Since chimera 388 bound nectin-1 with WT efficiency, we prepared a gD285/gH388 chimera to serve as a control (t388, Fig. 9A).

FIG. 9.

The gD285 chimeras. (A) Schematic diagram showing the gD285-gH proteins compared to original chimeras 22 and 388. gD and gH portions of the chimeric proteins are indicated at the top of the figure. gD285 chimeras contain the first 285 residues of gD (in addition to the gD signal sequence) fused to N-terminal gH truncations at amino acid 22 (t22) or 388 (t388). (B to D) Binding of the chimeras to cellular receptors (B), the cell-cell fusion assay in the absence of gD (C), and complementation of gD-null virus (D) were as described for Fig. 6. OD₄₀₅, optical density at 405 nm.

Chimera t388 had receptor binding properties similar to those of chimera 388 (Fig. 9B). However, chimera t22 showed enhanced binding to nectin-1 compared to the original chimera 22 (Fig. 9B). In the fusion assay, chimera t388 supported the fusion of CHO target cells expressing either HveA or nectin-1 at 20 to 35% of WT gD levels (Fig. 9C). However, t388 performed slightly better than chimera 388 in gD-null virus complementation with titers approximately 15% of WT gD (Fig. 9D) compared to 7% of WT gD for chimera 388 (Fig. 8A). Despite its enhanced receptor binding, chimera t22 performed very poorly as a gD substitute in the cell-cell fusion assay (Fig. 9C). Nevertheless, chimera t22 complemented gD-null virus at levels approaching that of WT gD (Fig. 9D). The complementation data imply that the gD285 fragment contains all the information needed to supply gD function in entry. Overall, the data from the gD285 chimeras suggest that there is a critical function of gD (other than receptor binding) that is lost when gD is linked to gH.

DISCUSSION

A fundamental conclusion of our study is that gD (in at least some version) and gH as well as gB and gL are required for HSV entry (as assessed by null virus complementation) and cell fusion. Although some chimeras could substitute for WT gH (Table 2) or gD (Table 3) for fusion, none eliminated the need for both, and in every case, gL was also essential. Thus, the HSV gDgH chimeras that we prepared do not functionally mimic the PrV gDgH chimera (24). Moreover, our data highlight the differences between gD, gH, and gL of the two viruses. It was previously noted that, unlike HSV gH, PrV gH could be incorporated into virions in the absence of gL (23). However, virus infectivity was significantly enhanced by the presence of gL and formation of a gH-gL complex in the virion. It is well established that, for HSV, gH transport, cell surface expression, and incorporation into the virion envelope require complex formation with gL (10, 21, 43, 44). In light of these differences it is perhaps not surprising that the requirement for gL was not eliminated by the presence of any of the gDgH chimeras in either entry or cell fusion assays.

TABLE 2.

Summary of data pertaining to gH function

WT or chimera	gL bindinga	Avg % of WT activityb
		Fusion assay		Complementation
		HveA	Nectin-1
gH	+	100	100	100
22	+	10	18	90
259	+	0	0	0
388	−	0	1	0
432	−	0	0	0

^agL binding was determined qualitatively based on immunoprecipitation results seen in Fig. 3.

^bAverage from three separate experiments.

TABLE 3.

Summary of data pertaining to gD function

WT or chimera	Avg % of WT activitya
	Receptor binding		Fusion assay		Complementation
	HveA	Nectin-1	HveA	Nectin-1
gD	100	100	100	100	100
gD308
22	90	25	8	4	90
259	90	25	0	0	0
388	100	100	50	35	7
432	100	100	25	20	7
gD285
t22	110	110	10	8	60
t388	110	110	35	20	15

^aAll results are averages from three separate experiments.

The chimeras provide insights into structural and functional aspects of gD, gH, and gL.

First, full-length gH was able to function at WT levels in a gH-null virus complementation assay with gD appended on its N terminus (in the case of chimera 22). Furthermore, chimera 22 bound gL and reacted with the MAb LP11, the antibody that is used as the “gold standard” for assessing gH-gL conformation (17, 19, 44). In addition, this chimera substituted for gH in the fusion assay, albeit poorly. Thus, gD did not place any structural constraints on the gH entry function but had an effect on cell-cell fusion.

Second, the properties of chimeras 259 and 388 suggest that amino acids 259 to 387 play an important role in the transport of gH-gL to the cell surface. The presence of an internal retention signal in this region seems unlikely, as signals are usually located at or near a protein's cytoplasmic domain (16). One possible explanation is that the N terminus of gH (in this case amino acids 1 to 387) requires proper gL binding for protein folding. Perhaps, when gL is absent or misaligned (as may be the case with chimera 259), these amino acids are improperly folded, causing the molecule to be retained in the endoplasmic reticulum. Unfortunately, there are no conformation-dependent MAbs to the gH N terminus that do not also require the presence of gL.

Yet gL binding alone was not enough to restore protein transport, as chimera 259 bound gL but was not transported to the cell surface. Moreover, the chimera 259-gL complex had different structural properties from WT gH-gL. The evidence for this is as follows. A PAb to a peptide which mimics the N terminus of gL can bind free gL but not gL bound to WT gH (44, 47). This was interpreted to mean that the gH N terminus binds to and obscures the gL terminus, interfering with antibody binding (44). However, this antibody immunoprecipitated gL bound to chimera 259, suggesting that the gL N terminus was no longer obscured by gH. When the properties of the chimeras are taken together with our previous data from studies of C-terminal gH truncations (44), it appears that a major site for gL binding lies between amino acids 259 and 323 of gH. This region corresponds to a similar region suggested by Westra et al. (58).

Third, chimeras 388 and 432 had very interesting properties. Neither protein formed a complex with gL, yet both were transported to the cell surface. In spite of being transported, neither chimera successfully substituted for gH or gL in functional assays (Table 2). Our interpretation is that chimeras 388 and 432 fail because they lack key elements of gH that are needed to bind gL (44, 58). It is interesting that, although the C terminus of gH has been implicated as a fusion domain (1, 13, 20, 62), these sequences alone were not able to function in entry or fusion when fused to gD. Both chimeras were able to substitute for gD in cell-cell fusion provided that gH and gL were also present. A puzzling observation is that neither protein could substitute well for gD in complementation, raising the possibility that the chimeras are not incorporated into the virion or that masking of gD in the virion could prevent chimera function.

Further insights into gD function.

Previously we studied the ability of gD truncated at amino acid 306 (gD306t) to bind its receptors HveA, nectin-1, and 3-OST. gD306t had micromolar affinity for each of these receptors. gD truncated even further to amino acid 285 had a 100-fold-higher affinity for HveA and nectin-1 (27, 48, 61). Our data suggested that all of the information needed for gD to bind to either receptor was contained within amino acids 1 to 285. The three-dimensional structure of gD bound to HveA revealed that all of the HveA contact residues are at the N terminus of gD (5). The site of nectin-1 binding is not as well defined but probably overlaps that for HveA (38). The question here was how much of the gD protein is needed to function in fusion and complementation. It was recently reported that a chimera consisting of the first 315 amino acids of gD attached to a heterologous transmembrane region and cytoplasmic tail can complement a gD-null virus (60). The properties of chimera 22 and t22 indicate that all of the information needed for gD to function in entry is contained within the first 285 amino acids. An interesting observation is that chimeras 22 and t22 (containing full-length gH) functioned better for gD complementation than did chimeras 432, 388, and t388 (which contained roughly 400 fewer amino acids). This observation suggests that gD may be presented differently when it is linked to the shorter forms of gH.

Virus entry and cell-cell fusion.

The results of the complementation and cell-cell fusion assays agreed when gH function was tested (Table 2). Chimeras 388 and 432 were negative in both assays. Chimera 22 was able to substitute for gH in both assays although fusion levels were only 10 to 18% of that seen for WT. Chimeras 22 and t22 were poor substitutes for gD in the fusion assay even though both replaced gD at WT levels in complementation (Table 3). This finding was puzzling, but it is interesting that the shorter chimeras (388 and 432) exhibited the opposite trend when replacing gD (35% of WT level in cell-cell fusion versus 7% of WT level in complementation). Apparently, one cannot accurately predict the outcome of one of these assays on the basis of the other.

Our findings highlight the fact that the two assays do not always agree. Others have noted that cell-cell fusion and virus-cell fusion are inherently different (35, 40, 62) and may require a different set of protein-lipid machinery for each task. Alternatively, the interactions between the same set of proteins may be different depending on whether the protein is presented from the viral membrane or from the cellular membrane (18). A third possibility may be a difference in context. With the fusion assay, the context is only the four viral glycoproteins essential for fusion and an HSV receptor. In the null virus complementation assay, the chimera expressed on the viral envelope is in the context of all the other virion proteins. It is unknown whether coexpression of other viral proteins would alter the HSV-1 gDgH chimera phenotype in the fusion assay.

Future studies will be directed at understanding more of the structure of the gDgH chimeras and their potential as vaccine candidates. A goal will be to crystallize the gH-gL complex. Should this be problematic, chimera 22 or t22 might be a viable candidate. It is appealing that the structure of gD285 is already known (5), possibly making solution of the hybrid structure an easier task. In addition, both gD and gH-gL have been shown previously to protect animals from virus challenge (4, 28, 30, 43). HSV-2 gD has shown promise as a subunit vaccine in women (54). Perhaps a gDgH hybrid will offer a higher degree of protection than would either protein alone.