Differential Effects on Cell Fusion Activity of Mutations in Herpes Simplex Virus 1 Glycoprotein B (gB) Dependent on Whether a gD Receptor or a gB Receptor Is Overexpressed

Abstract

Glycoprotein B (gB) of herpes simplex virus (HSV) is one of four glycoproteins essential for viral entry and cell fusion. Recently, paired immunoglobulin-like type 2 receptor (PILRα) was identified as a receptor for HSV type 1 (HSV-1) gB. Both PILRα and a gD receptor were shown to participate in HSV-1 entry into certain cell types. The purpose of this study was to determine whether insertional mutations in gB had differential effects on its function with PILRα and the gD receptor, nectin-1. Previously described gB mutants and additional newly characterized mutants were used in this study. We found that insertional mutations near the N terminus and C terminus of gB and especially in the central region of the ectodomain reduced cell fusion activity when PILRα was overexpressed much more than when nectin-1 was overexpressed. Most of the insertions reduced the binding of gB to PILRα, for at least some forms of gB, but this reduction did not necessarily correlate with the selective reduction in cell fusion activity with PILRα. These results suggest that the regions targeted by the relevant mutations are critical for functional activity with PILRα. They also suggest that, although both the binding of gB to a gB receptor and the binding of gD to a gD receptor may be required for HSV-induced cell fusion, the two receptor-binding activities may have unequal weights in triggering fusogenic activity, depending on the ratios of gB and gD receptors or other factors.

Manifestations of disease caused by herpes simplex virus (HSV) include recurrent mucocutaneous lesions in the mouth or on the face or genitalia and, more rarely, meningitis or encephalitis. The infection of host cells occurs by the fusion of the virion envelope with a cell membrane to deliver the nucleocapsid containing the viral genome into the host cell. This entry process and virus-induced cell fusion require glycoprotein B (gB), along with gD, gH, and gL. The membrane-fusing activity of HSV depends in part on the binding of gD to one of its receptors, herpesvirus entry mediator (HVEM), nectin-1, nectin-2, or 3-O-sulfated heparan sulfate (18). HVEM is a member of the tumor necrosis factor receptor family and is expressed by cells of the immune system, as well as many other cell types, such as epithelial, stromal, and endothelial cells (23). Nectin-1 and nectin-2 are cell adhesion molecules belonging to the immunoglobulin superfamily and are widely expressed by a variety of cell types, including epithelial cells and neurons (20). Specific sites in heparan sulfate generated by particular 3-O-sulfotransferases can serve as gD-binding entry receptors (17). This binding of gD to a receptor is associated with conformational changes in gD that are thought to enable gD to interact with gB and/or the heterodimer gH-gL to trigger fusogenic activity (8, 12). Both gB and gH have properties of fusogenic viral proteins (1, 7). Although evidence has been presented that gD and gH-gL are sufficient for hemifusion and that gB, in addition, is required for fusion pore formation (19), the specific roles each plays in HSV-induced membrane fusion have not been fully defined.

gB was recently discovered to bind to paired immunoglobulin-like type 2 receptor (PILRα) in an interaction that can mediate viral entry and cell fusion, provided that gD also binds to one of its receptors (14). For cells such as CD14⁺ monocytes, antibodies specific for either HVEM or PILRα were shown to block HSV entry. Also, entry requires the presence of both gD and gB in the virion. Although the overexpression of either a gD receptor or a gB receptor can enhance the susceptibility of cells to HSV entry and HSV-induced cell fusion, there are very few, if any, cell types that do not express at least low levels of endogenous receptors. Thus, the possibility exists that these endogenous receptors are cooperating with the introduced receptors to render the cells susceptible to HSV-induced membrane fusion.

PILRα belongs to the paired-receptor families, which consist of activating and inhibitory receptors (4, 11, 19). They are conserved among mammals (24). The inhibitory form PILRα has an immunoreceptor tyrosine-based inhibition motif in its cytoplasmic domain and transduces inhibitory signals (4). On the other hand, the activating form PILRβ associates with the immunoreceptor tyrosine-based activation motif-bearing DAP12 adaptor molecule and delivers activating signals (16). Both PILRα and PILRβ are expressed on cells of the immune system, especially monocytes, dendritic cells, and NK cells (4, 11, 19), and also in neurons (14). CD99 has been identified as a natural ligand for both PILRα and PILRβ (16). The binding of either PILRα or PILRβ to CD99 depends on the presence of sialyated O-linked glycans on CD99 (22).

In addition to binding to PILRα, gB can bind to heparin and heparan sulfate and may contribute, along with gC, to the binding of HSV to cell surface heparan sulfate (17). Also, gB and gC can bind to DC-SIGN, which serves as a binding receptor for the infection of dendritic cells (2). An X-ray structure of the HSV-1 gB ectodomain reveals a homotrimeric conformation with structural homology to vesicular stomatitis virus (VSV) G glycoprotein, the single glycoprotein responsible for the entry of VSV. Both HSV-1 gB and VSV G glycoprotein have features of class 1 and class 2 viral fusion proteins and have been designated class 3 fusion proteins (7, 14, 15). The heparan sulfate-binding determinant of gB has been localized to a lysine-rich domain in the N terminus and shown to be dispensable for viral entry (9). It lies within a region that is probably disordered and was not included in the defined coordinates of the X-ray structure. The binding of DC-SIGN to gB probably depends on high-mannose N-glycans of gB (6).

In a previous study (10), 81 insertion mutants of HSV-1 gB were characterized to assess the effects of the insertions on protein processing and function in cell fusion with gD receptors, in relation to structural domains of gB identified in an X-ray structure (7). Only 27 mutants were found to be processed into mature glycosylated forms and transported to the cell surface. Only 11 of these retained fusion activity toward target cells expressing nectin-1 or HVEM. For the present study, we used 25 previously described gB insertion mutants shown to be expressed on cell surfaces and also identified an additional 10 such mutants.

The present study was designed to determine whether the effects of insertions in gB on cell fusion activity would be dependent on whether a gD receptor (nectin-1) or a gB receptor (PILRα) was overexpressed in target cells that also expressed unidentified weak endogenous receptors. In addition, we assessed the abilities of the gB mutants to bind to PILRα. Our results showed that some insertions inhibited cell fusion activity when PILRα was overexpressed significantly more than when nectin-1 was overexpressed, but without necessarily preventing the binding of PILRα to gB, at least to some stable oligomeric forms of gB. The results indicate that, although both a gB receptor and a gD receptor may be required for cell fusion activity, the two receptor-binding activities have unequal weights in triggering fusogenic activity, depending on the ratios of gB and gD receptors or other factors.

MATERIALS AND METHODS

Cells.

Chinese hamster ovary K1 (CHO-K1) cells (ATCC), 293 PEAK Rapid cells (Edge Biosystems), and a mutant CHO cell line selected for the absence of PILRα ligands (14) were used for this study. The CHO-K1 cell line and derivatives were grown in Ham's F-12 medium supplemented with 10% fetal bovine serum (FBS), and the 293 PEAK Rapid cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% FBS.

Plasmids.

Plasmids expressing HSV-1(KOS) gB (pPEP98), gD (pPEP99), gH (pPEP100), and gL (pPEP101) were described previously (11), as were plasmids expressing human nectin-1 (pBG38) (5) and a soluble human PILRα-immunoglobulin G (IgG) Fc hybrid protein (PILRα-Ig; pME18S-PILRα-Ig) (14). A plasmid expressing human PILRα (pQF003) was generated by subcloning the PILRα open reading frame from a commercial cDNA clone (Origene, SC108389) into pcDNA3 by using primers 5′-AATAAGCTTGCCGCCACCATGGGTCGGCCCCTGCTG-3′ and 5′-AATGGTACCGGGCTGTCCATTGGTTAGGC-3′ for PCR amplification and subjecting the products to digestion with HindIII and KpnI and ligation into the prepared vector.

gB mutants.

Previously described gB mutants (10) generated by the GPS-LS linker scanning system (New England Biolabs, Ipswich, MA), as well as 10 new mutants generated in the same way, were used. The new mutants were identified among a larger set that was first screened for cell surface expression of the protein. The plasmids encoding gB mutants capable of folding correctly enough for transport to the cell surface were then sequenced to identify the precise position of each insertion.

Preparation of PILRα-Ig supernatant.

Plates of 293 PEAK Rapid cells were transfected with pME18S-PILRα-Ig by using 293 Fectin (Invitrogen). Culture supernatants were collected at 48 and 72 h after transfection and pooled. The concentrations of PILRα-Ig in the culture supernatants were measured using a human IgG enzyme-linked immunosorbent assay (ELISA) kit (Immunology Consultants Laboratory, Inc.). The supernatants were used as described below.

Assays for cell surface expression of gB and binding of PILRα.

For the initial characterization of the new mutants, CHO-K1 cells were transfected with plasmids expressing individual gB mutants or wild-type gB or with an empty vector, along with plasmids expressing gD, gH, gL, and T7 polymerase. The cells were washed once with phosphate-buffered saline (PBS) 16 h after transfection, and a cell-based ELISA was performed with the anti-gB serum R74 as described previously (10). Briefly, after incubation of the cells with the antibodies, the cells were washed, fixed, and incubated with biotinylated goat anti-rabbit IgG (Sigma), followed by streptavidin-horseradish peroxidase (streptavidin-HRP; GE Healthcare) and an HRP substrate (BioFX). Values obtained for wild-type gB (optical densities at 380 nm) ranged from 0.7 to 1.1. Negative control values (for samples in which an empty vector replaced the plasmid expressing wild-type gB) ranged from 0.17 to 0.37. Negative control values were subtracted from each of the experimental values before the results were normalized relative to those obtained with wild-type gB. To assess the expression of gB mutants on cell surfaces, along with their abilities to bind to PILRα, CHO-K1 cells seeded into six-well plates were transfected with 1.5 μg of an empty vector or a plasmid expressing wild-type gB or a gB mutant and 5 μl of Lipofectamine 2000. After 24 h of incubation, the cells were detached using a solution of 0.2 g of EDTA per liter of PBS, washed with PBS, incubated separately with anti-gB antiserum R74 at a 1:10,000 dilution or PILRα-Ig at 70 ng/ml for 1 h at room temperature, and then washed twice with PBS and lysed with 200 μl of lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 10 mM NaF, 1 mM Na₃VO₃, 1% Nonidet P-40) containing a protease inhibitor mixture (Roche Diagnostics, Indianapolis, IN). Protein G agarose (Thermo Scientific) was used to precipitate the proteins associated with the antibody or PILRα-Ig. Proteins eluted from the beads were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 4 to 20% acrylamide gels after being boiled for 5 min under nonreducing conditions. Western blot analyses were performed by using the rabbit anti-gB antiserum R74 at a 1:10,000 dilution for both the R74- and PILRα-Ig-precipitated samples, anti-rabbit secondary antibodies coupled to HRP, and enhanced chemiluminescence Western blotting detection reagents (GE Healthcare). The developed films were scanned, and ImageJ software (http://rsbweb.nih.gov/ij/) was used to quantify the density of each gB band. The values obtained for each mutant gB band (corresponding to a monomer or an oligomer in precipitates obtained with anti-gB or PILRα-Ig) were normalized with respect to the density of the wild-type gB control band precipitated by the anti-gB serum.

Cell fusion assay.

The cell fusion assay was done as described previously (11). Briefly, CHO-K1 and CHO PILRα ligand-negative cells were seeded into six-well plates 1 day before transfection. The CHO-K1 cells (effector cells) were transfected with 400 ng each of plasmids expressing T7 RNA polymerase, gD, gH, and gL; 800 ng of an empty vector or a plasmid expressing either wild-type or mutant gB; and 5 μl of Lipofectamine 2000. The CHO PILRα ligand-negative cells (target cells) were transfected with 400 ng of a plasmid carrying the firefly luciferase gene under the control of the T7 promoter, 1.5 μg of an empty vector (pcDNA3) or a plasmid expressing either human PILRα (pQF003) or human nectin-1 (pBG38), and 5 μl of Lipofectamine 2000. Six hours after transfection, the cells were detached with 0.2 g of EDTA per liter of PBS and suspended in 1.5 ml of Ham's F-12 medium supplemented with 10% FBS. Effector and target cells were mixed in a 1:1 ratio and replated into 96-well plates for 18 h. Luciferase activity was quantitated by a luciferase reporter assay system (Promega) using a Wallac-Victor luminometer (Perkin Elmer). Negative control values obtained with the empty vector (no gB) were subtracted from each of the experimental values before the normalization of the results relative to those obtained with wild-type gB.

RESULTS

Identification of new gB mutants.

For this study, we screened a previously uncharacterized set of gB insertion mutants in order to identify others capable of being expressed on the surfaces of CHO cells. Figure 1 shows cell surface expression levels of 10 new mutants normalized relative to that of wild-type gB. Only two of the mutants (those with insertions after S500 and K895) had expression levels indistinguishable from that of wild-type gB, but all those that exhibited levels of cell surface expression in excess of 20% of wild-type levels were included in further analyses. Other mutants failed to meet this criterion or were otherwise aberrant in terms of the insertion and were not studied further.

FIG. 1.

Cell surface expression of the newly characterized gB insertion mutants in comparison with that of wild-type (WT) gB. CHO-K1 cells in 96-well plates were transfected as if they were to be used as effector cells in cell fusion assays. However, instead of being mixed with target cells, these cells expressing wild-type gB or mutant forms with insertions after the indicated amino acids (along with HSV-1 gD, gH, and gL) were washed and incubated with an anti-gB antiserum. The cells were then washed thoroughly prior to fixation and incubation with a secondary antibody and an HRP detection system. The data are presented as the means and standard deviations of results obtained in seven independent experiments. Although all the gB mutants were reproducibly expressed at levels higher than 20% of wild-type gB levels, only the mutants with insertions after S500 and K895 were indistinguishable from wild-type gB by one-way analysis of variance.

The new mutants, with insertions after the amino acids listed in Fig. 1, along with previously described mutants capable of being expressed on the cell surface were selected for this study. Each mutant contained a 5-amino-acid insertion.

Relative abilities of endogenous receptors, PILRα, and nectin-1 to serve as fusion receptors for wild-type HSV-1 glycoproteins in CHO cells.

Normal CHO cells are resistant to HSV entry and are poor target cells in cell fusion assays unless transfected to express entry/fusion receptors. However, they do express low levels of, or inefficient, endogenous HSV entry/fusion receptors (15). To provide a basis for the comparison of the gB mutants with wild-type gB in mediating fusion with various fusion receptors, we first compared the activities of the fusion receptors with wild-type HSV-1 glycoproteins in the CHO cell system. CHO cells were transfected with plasmids expressing wild-type gB or an empty control vector, along with plasmids expressing gD, gH, gL, and T7 polymerase, to generate effector cells. The target cells were CHO cells selected for the absence of ligands of PILRα (14) and were transfected with an empty vector or plasmids expressing PILRα or nectin-1 and the luciferase plasmid. The data presented in Table 1 summarize the results of 26 independent experiments. The absolute values obtained for luciferase activity, as a measure of cell fusion, differed from experiment to experiment. Therefore, the results had to be normalized with respect to the activity of wild-type gB in each experiment. Reproducibly, cell fusion activity was significantly higher in the presence of gB than in its absence, even when the target cells were transfected with an empty vector and expressed only endogenous receptors. If there were no endogenous receptors, the background levels of cell fusion activity should have been the same whether or not the full complement of HSV-1 glycoproteins required for fusion was present. When overexpressed along with endogenous receptors, PILRα was not as active as nectin-1 in serving as a fusion receptor (it displayed 14.1% of the activity of nectin-1) but its ability to enhance cell fusion (by fivefold) over that observed with endogenous receptors was significant (P < 0.0001) (Table 1).

TABLE 1.

Cell fusion activity of wild-type HSV-1 gB with CHO cells expressing only endogenous receptors or transfected with PILRα or nectin-1

Receptor(s)a	Presence or absence of gBb	Luminometer reading (arbitrary units)c		Cell fusion activity (% of gB activity with nectin-1)d
		Minimum	Maximum	Mean	SD
Endogenous	Present	3,590	75,990	2.8	1.27
receptors	Absent	260	4,970
PILRα	Present	13,990	232,930	14.1	4.17
	Absent	170	5,150
Nectin-1	Present	136,130	1,908,470	NA	NA
	Absent	830	8,510

^aTarget cells were PILRα ligand-negative CHO cells transfected with an empty vector or with a plasmid expressing PILRα or nectin-1 (and also with a plasmid expressing luciferase under the control of the T7 promoter).

^bEffector cells were transfected with a plasmid expressing wild-type gB or an empty vector and also with plasmids expressing HSV-1 gD, gH, gL, and T7 polymerase.

^cMinimum and maximum raw values obtained for cell fusion activity in 26 experiments. Absolute levels of luciferase activity varied from experiment to experiment.

^dNormalized levels of cell fusion activity for endogenous receptors and PILRα as a percentage of wild-type gB activity with nectin-1. Within each experiment, all three receptors were tested and background values (for activity in the absence of gB) were subtracted from experimental values before the normalization of results according to those obtained with nectin-1. Student's t test (paired, two-tailed) revealed that the increase in cell fusion activity conferred by PILRα expression (compared to activity with only endogenous receptors) was significant (P < 0.0001; n = 26). NA, not applicable.

Effects of the insertional mutations on fusion activity with target cells expressing endogenous receptors, PILRα, or nectin-1.

Experiments were performed as described in the preceding section except that the ability of each of the gB mutants to induce fusion with the three types of target cells was compared with that of wild-type gB (Fig. 2 and 3A). Note that many of the insertions in gB abolished fusion activity with cells expressing endogenous receptors, as well as with cells expressing PILRα or nectin-1. This finding is additional evidence that the cell fusion activity seen with wild-type gB by using CHO cells expressing only endogenous receptors as target cells (Table 1) is real and depends on gB function.

FIG. 2.

Cell fusion activity of the gB mutants with target cells expressing endogenous receptors, PILRα, or nectin-1. PILRα ligand-negative CHO cells were transfected with pcDNA3 (empty vector), a PILRα-expressing plasmid (pQF003), or a nectin-1-expressing plasmid (pBG38), along with a reporter plasmid expressing luciferase under the control of the T7 promoter. They were replated with CHO-K1 cells previously transfected with plasmids expressing T7 polymerase, gD, gH, gL, and wild-type (WT) gB or one of the mutant gB forms with insertions after the indicated amino acids. After 18 h, the cells were lysed and mixed with a luciferase substrate and the luciferase activity was quantified as a measure of cell fusion. Each bar shows the means of three to six independent determinations, with the results expressed as the percentage of wild-type gB activity observed with target cells expressing each of the receptors indicated. Data for the endogenous receptors are the same in the upper and lower panels.

FIG. 3.

Cell fusion activity of the gB mutants with target cells expressing PILRα or nectin-1 (A) and precipitation of cell surface forms of the gB mutants by an anti-gB antiserum or PILRα-Ig (B). (A) The cell fusion results from Fig. 2 comparing PILRα and nectin-1 as receptors are repeated here. The stars correspond to mutants whose activity with PILRα was significantly lower than that with nectin-1 (Student's t test; P < 0.05 [n = 4 to 5]). (B) CHO-K1 cells were transfected with plasmids expressing wild-type (WT) gB or each of the mutants with insertions after the indicated amino acids. Intact cells were detached and incubated either with anti-gB or with PILRα-Ig. After being washed, the cells were lysed for the isolation of complexes of anti-gB or PILRα-Ig with cell surface gB. The complexes were prepared for SDS-PAGE using conditions that usually fully dissociate gB oligomers but failed to fully dissociate oligomers formed by some of the mutants. The anti-gB antiserum was used to probe Western blots for the gB precipitated by anti-gB or PILRα-Ig (Fig. 4). The quantitation of the gB bands on the blots demonstrated the relative abilities of anti-gB and PILRα-Ig to bind to cell surface gB and indicated whether the forms bound were readily dissociated by preparation for SDS-PAGE. The results presented are the means and standard deviations of three determinations.

Figure 2 (top) shows that the cell fusion activities of the mutants, relative to that of wild-type gB, were similar whether endogenous receptors alone or PILRα in addition was expressed by the target cells. However, some of the mutants retained activity with nectin-1 (plus endogenous receptors) to a greater extent than with endogenous receptors alone (Fig. 2, bottom) or with PILRα (Fig. 3A). Seven mutants, namely, those with insertions after K105, D408, R470, I495, T497, F732, and N819, showed significantly higher activities with nectin-1 than with PILRα. Mutants with insertions after R470, I495, and T497 were notable for exhibiting cell fusion activities with nectin-1 that were comparable to that of wild-type gB while exhibiting significantly reduced or no activity with PILRα. Thus, for these mutants, the effects of the insertions on cell fusion activity were highly dependent on the fusion receptors expressed by the target cells.

Effects of insertional mutations on cell surface expression of gB and binding to PILRα.

A soluble hybrid form of PILRα, PILRα-Ig, can bind to gB on cell surfaces (14) and to its natural ligand CD99 (22). To assess the levels of wild-type and mutant gBs on CHO cell surfaces, in conjunction with the quantification of PILRα-Ig binding to the cell surface gBs, experiments were done as follows. CHO cells were transfected with plasmids expressing wild-type gB or selected gB mutants. The cells were then detached (intact), incubated with anti-gB antibodies or with PILRα-Ig, washed thoroughly, and lysed for the collection of Fc-containing complexes on protein G beads. The collected complexes were prepared for SDS-PAGE, and Western blotting was performed with an anti-gB serum to quantitate the amounts of gB precipitated by anti-gB and by PILRα-Ig. The quantitative results are shown in Fig. 3B, and a representative set of blots is shown in Fig. 4.

FIG. 4.

Western blots showing the relative amounts of wild-type (WT) gB and mutants, with insertions after the indicated amino acids, accessible on CHO cell surfaces for the binding of anti-gB antibodies or PILRα-Ig and subsequent precipitation. The anti-gB antiserum used for precipitation for the upper set of blots was also used for the probing of all the blots. In each panel, the lower band is monomeric gB. For some of the mutants, oligomeric gB was not fully dissociated by SDS or heat and oligomeric bands can be seen. This set of blots is representative of three determinations, with the experiments done as described in the legend to Fig. 3 and in the text. IP, immunoprecipitation. Numbers to the left of each panel indicate the molecular weights of marker proteins.

There appear to be at least two populations of gB precipitated from cell surfaces, one that could be fully dissociated into monomers by heat and one that could not (Fig. 3B and 4). Only certain mutants displayed the nondissociable oligomeric forms, and these results were highly reproducible. The amounts of fully dissociable forms of most mutant gBs precipitated by anti-gB ranged from about 50 to 150% of wild-type gB levels. (In some cases, the results presented in Fig. 3B for anti-gB precipitation are not entirely consistent with those shown in Fig. 1 for cell-based ELISA using the same anti-gB antibodies. This discrepancy may be accounted for in part by the different methodologies and by the use of monolayer versus detached cells.) For wild-type gB, none of the nondissociable form was detected. The amounts of nondissociable mutant gBs precipitated by anti-gB were invariably smaller than those of the dissociable forms (Fig. 3B).

Comparisons of the data in Fig. 3A and B show that the two mutants expressed at the highest levels on the cell surface (those with insertions after F732 and M742) exhibited reduced levels of cell fusion activity with all receptors. Conversely, the two mutants exhibiting the highest levels of cell fusion activity with both nectin-1 and PILRα (those with insertions after T868 and S869) were detected at reduced levels on cell surfaces by anti-gB. The latter two mutants have their insertions in the cytoplasmic tail of gB. Previous studies have shown that deletions or amino acid substitutions within the cytoplasmic tail of gB can either significantly enhance or reduce cell fusion activity and that the levels of activity do not necessarily correlate with levels of cell surface expression (3, 13).

The amount of wild-type gB precipitated by PILRα-Ig was comparable to that precipitated by anti-gB (Fig. 3B). Only one mutant (that with an insertion after R470) was indistinguishable from wild-type gB in this respect. All other mutant gBs exhibited reduced levels of precipitation of the dissociable form by PILRα-Ig, compared with the amounts available on the cell surface as assessed by anti-gB precipitation (Fig. 3B, top). Similar results were obtained using anti-gB and PILRα-Ig to precipitate gB from lysates of cells that had been transfected to express wild-type gB or the mutants and then biotinylated to selectively label cell surface proteins (data not shown). Unexpectedly, the V014 mutant had reduced ability to bind to PILRα-Ig and also reduced activity with PILRα, although the difference between activities with nectin-1 and PILRα did not quite reach statistical significance. The insertion point for this mutant is in the signal peptide, but the insertion is predicted not to alter the position of the signal peptidase cleavage site. Possibly, some aspect of the processing of this mutant is altered by its having a signal peptide of 35 amino acids instead of 30 amino acids.

As indicated in the bottom panel of Fig. 3B, PILRα-Ig was more efficient at precipitating the nondissociable forms of mutant gBs than was anti-gB. This finding suggests that the conformations of cell surface gB associated with resistance to dissociation into monomers had enhanced abilities to bind to PILRα.

Many of the insertional mutations reduced the binding of PILRα-Ig to the dissociable forms of gB (but not to the nondissociable forms), whereas only a few of the mutants exhibited significantly reduced cell fusion activity with PILRα compared with nectin-1. Conversely, the R470 insertion mutant bound to PILRα-Ig as efficiently as wild-type gB but had significantly reduced cell fusion activity with PILRα compared with nectin-1. Thus, it seems unlikely that these two mutant phenotypes (selective loss of cell fusion activity with PILRα and reduced binding to PILRα-Ig) are causally related.

DISCUSSION

Some of the insertion mutations described herein alter the structure of gB so that cell fusion activity with PILRα is inhibited or significantly reduced, with less or no effect on activity with the gD receptor nectin-1. Insertions into either the N-terminal or the C-terminal region of gB (after K105, F732, and N819) or, especially, into the middle region (after D408, R470, I495, and T497) selectively inhibited activity with PILRα without necessarily preventing PILRα binding (although most of these insertions did reduce PILRα-Ig binding, at least to the dissociable form of gB). Thus, some of the features of gB required for cell fusion when HSV-1 glycoproteins and PILRα are overexpressed (and perhaps cooperate with endogenous receptors) are not required for cell fusion under conditions in which HSV-1 glycoproteins and the gD receptor nectin-1 are overexpressed (and perhaps cooperate with endogenous receptors). The conditional nature of the phenotypes of these mutants, at least with respect to cell fusion, indicates that the structural requirements for gB function in cell fusion depend on the receptors expressed and probably other factors. An implication of this finding is the possibility that there are multiple ways to trigger the fusogenic activity of gB and perhaps different conformational transitions that accompany each.

As noted previously (10) and confirmed here, insertions into large portions of the gB ectodomain eliminate cell fusion activity regardless of the fusion receptor expressed (Fig. 2 and 3). However, insertions into the flexible N-terminal domain, the region of the ectodomain proximal to the membrane, the C-terminal cytoplasmic tail, or a disordered region near the middle of the ectodomain are compatible with partial or complete retention of cell fusion activity, at least with nectin-1. Interestingly, the latter regions are all excluded from the resolved structure of the gB ectodomain (7) (Fig. 5).

FIG. 5.

Mapping of the insertion points in the gB mutants exhibiting greater impairment in fusion with PILRα-expressing cells than in that with nectin-1-expressing cells and of the potential positions for the addition of O-glycans. The triangles show the positions, on a linear representation of gB, of the insertions that selectively inhibit cell fusion activity with PILRα. The circles show the positions of Thr or Ser residues that may potentially be sites for the addition of O-glycans (identified by using the NetOGlyc 3.1 prediction server hosted by the Technical University of Denmark). The colored bars represent the structural domains of a crystallized portion of the gB ectodomain (7). White or hatched regions were not included in the X-ray structure. TM, transmembrane domain.

It is also noteworthy and perhaps relevant to gB function that the binding of PILRα to its natural ligand CD99 requires two sialylated O-glycans attached 5 amino acids apart on CD99 (22). In a previous study, sialic acid moieties on HSV-1 virion glycoproteins were shown to be required for viral infectivity (21), although the specific glycoprotein that carries the critical sialic acid was not identified. All potential sites for the addition of O-glycans to gB are located in the N-terminal region and the disordered region near the middle of the ectodomain (Fig. 5). The possibility exists that the binding of PILRα to gB, as to CD99, requires two sialylated O-glycans, perhaps located at widely separated positions on the gB polypeptide chain but in close proximity in certain conformations of the gB trimer. Many of the insertion mutants described herein may have alterations in conformation that are responsible for reduced or altered PILRα binding, either because of effects on glycosylation itself or because of effects on the proximity of the O-glycans. If there is any merit to these speculations, PILRα-Ig may prove to be a useful tool for identifying specific conformations of gB. The results presented in Fig. 3B support this suggestion, in that the nondissociable forms of several mutant gBs bound to PILRα-Ig much more efficiently than the dissociable forms. Moreover, the findings of another study (J. Wang, Q. Fan, T. Satoh, J. Arii, L. L. Lanier, P. G. Spear, Y. Kawaguchi, and H. Arase, unpublished results) show that sialylated O-glycans at amino acids T53 and T480 are required for the binding of PILRα to gB and for gB function in cell fusion with PILRα but not with nectin-1.