Truncated Forms of Glycoprotein D of Herpes Simplex Virus 1 Capable of Blocking Apoptosis and of Low-Efficiency Entry into Cells Form a Heterodimer Dependent on the Presence of a Cysteine Located in the Shared Transmembrane Domains

Abstract

Earlier studies have shown that herpes simplex virus 1 (HSV-1) virions of mutant lacking glycoprotein D (gD) and made in either complementing (gD^−/+ stocks) or noncomplementing cells (gD^−/− stocks) induce apoptosis. Subsequent studies have shown that apoptosis induced by gD^−/− mutant virus stocks can be blocked by in trans delivery of viral genes that encode either intact gD or a mixture of two genes encoding the glycoprotein ectodomain plus transmembrane domain (gD-B) and transmembrane domain plus the cytoplasmic carboxyl terminus of the protein (gD-D), respectively. Since the presence of the transmembrane domains was critical for precluding apoptosis in the bipartite system, the question arose whether the two components, gD-B and gD-D, form a heterodimer mediated by an unpaired cysteine located in the transmembrane domain. We report the following. (i) The substitution of the unpaired cysteine with serine in either gD-B or gD-D truncated forms of gD disabled the ability of gD-D and gD-B to block apoptosis. (ii) Immunoprecipitation of gD-D coprecipitated gD-B only from lysates of cells transduced with gD-D and gD-B containing the cysteine in the transmembrane domains. Replacement of cysteine with serine ablated coprecipitation of the components. (ii) The mixture of gD-D and gD-B complemented at a low level gD^−/+ virions. We conclude that the gD-B and gD-D can form a heterodimer dependent on the presence of cysteines in the transmembrane domain and the heterodimer can substitute for intact gD but at a much reduced efficiency.

Glycoprotein D (gD) is the most extensively studied herpes simplex virus 1 (HSV-1) glycoprotein. The extensive literature is based primarily on four functions expressed in the course of the replicative cycle of the virus: (i) gD interacts with at least three different receptors to enable productive entry of the virion to the cell surface (2, 5, 7, 13, 19), (ii) gD is a component of a quartet of viral glycoproteins (along with gB, gH, and gL) to enable the fusion of the viral envelope with the plasma membrane (1, 3, 6, 10, 16, 18), and (iii) gD is required for the spread of virus from cell to cell by functions that may be related to those required for the initial entry of the virus into cells (8, 14, 15, 17, 20). (iv) gD expressed at the cell surface mediates restriction to infection by sequestering its own receptor (4, 9).

In recent studies, this laboratory has reported that gD also block apoptosis induced by HSV-1 early in infection (22, 23, 24). These studies may be summarized as follows.

(i) gD minus viruses may be classified into two kinds of stocks depending on how they are made. Stocks made in complementing cells expressing gD from a transduced gene lack the gene but contain the glycoprotein in their envelopes. These stocks, designated gD^−/+, can infect cells, but the progeny accumulates in the cytoplasm and lacks both the gene and the glycoprotein in the envelope. These progeny stocks, designated gD^−/−, can attach to sulfated proteoglycans on cell surfaces but do not enter by fusion of the envelope with the plasma membrane. Electron microscopic studies have shown that gD^−/− virions are taken up into endosome-like vesicles and degraded; infection does not ensue.

(ii) Both gD^−/− and gD^−/+ stocks induce apoptosis but both requirements for induction of apoptosis differ slightly. gD^−/− stocks require ca. 100 PFU equivalents per cell to induce apoptosis whereas gD^−/+ stocks require ca. 10 PFU per cells to trigger the same response. In contrast wild-type parent does not induce apoptosis at any multiplicity of infection tested.

(iii) Apoptosis induced by gD^−/− and gD^−/+ is blocked by either gJ or gD expressed by genes delivered in trans. In studies designed to test what component of gD is essential for blocking apoptosis, we found that the ectodomain of gD (construct gD-A, Fig. 1) effectively blocked gD^−/+ but not the gD^−/− mutant virus stocks. The components of gD required for blocking apoptosis induced by gD^−/− stocks were either the intact gD (gD-WT, Fig. 1) or a mixture consisting of the ectodomain + transmembrane domain (gD-B, Fig. 1) and transmembrane domain plus cytoplasmic domain (gD-D, Fig. 1). Mixtures containing the ectodomain (gD-A) and construct gD-D were not effective. These results supported two hypotheses. First, the key events that lead to induction and blockage of apoptosis by the two virus stocks take place in different cellular compartments. Second, structural requirements of the binary gD components suggested that they form a heterodimer through the transmembrane domain.

FIG. 1.

Schematic diagrams of the structure of gD and of truncated forms of gD employed in the present study. TM, transmembrane domain; C, cytoplasmic domain. The filled circle indicates that Cys7 was replaced by serine. The filled square indicates that an HA tag was inserted in frame before the stop codon.

(iv) Evidence in support of the first hypothesis rests on two series of experiments. The first showed that chloroquine blocks apoptosis induced by gD^−/− but not gD^−/+ stocks. The data infer that apoptosis induced by gD^−/− stock is triggered concurrent with of subsequent to destruction of gD in the endosomal compartment and by extension, that the events leading to apoptosis induced by gD^−/+ virus take place in a different compartment. The second series of experiments stemmed from reports that gD interacts with mannose-6-phosphate receptor (M6PR) and that virions and gD colocalize with M6PR in late endosomal compartment. In our studies, overexpression of M6PR blocked apoptosis induced by either gD^−/− or gD^−/+ virus. The significance of these results stems in turn on the well established role of M6PR as a regulator of lysosomal enzymes. The results of these studies indicated that while apoptosis induced by gD^−/− stocks is the consequence of lysosomal degradation of large amounts of gD^−/− virus taken up by endocytosis early in infection, the induction of apoptosis by gD^−/+ virus is an event associated with exocytosis of newly made viral exocytic vesicles.

The objective of this report is to test the second hypothesis, i.e., that the gD-B and gD-D block apoptosis by the formation of an obligatory formation of a heterodimer. The experimental design of these studies rests on the report that the seventh cysteine (Cys7) located in the transmembrane domain of HSV-1 gD is unpaired and that mutagenesis of this cysteine blocks the formation of gD homeodimers (11, 21). We report that the requirement of gD-D and gD-B components of gD to block apoptosis reflects the formation of a heterodimer with attributes similar in some respects to those of a monomeric intact gD molecule.

MATERIALS AND METHODS

Cells and Viruses.

SK-N-SH cells were obtained from American Type Culture Collection (Rockville, Md.) and maintained in Dulbecco modification of Eagle minimal essential medium (DMEM) containing 10% fetal bovine serum. The insect cell line Sf9 (Spodoptera frugiperda) was obtained from PharMingen (San Diego, Calif.). Unless indicated, cultures were seeded less than 20 h prior to infection and assayed at 60 to 70% confluence. gD^−/− and gD^−/+ mutant viruses were produced as described in detail elsewhere (22).

Antibodies.

Monoclonal antibodies against gD (clone H170), monoclonal antibodies to ICP0 (clone H1083), and HA-specific polyclonal antiserum were purchased from the Goodwin Institute (Plantation, Fla.) and were described elsewhere (22).

Construction of baculovirus recombinants expressing mutant gD-WT(Cys7Ser), gD-B(Cys7Ser), and gD-D(Cys7Ser).

The baculovirus transfer vector pAc-CMV containing human cytomegalovirus immediate-early promoter-enhancer sequences in the XhoI-EcoRI site of pAc-SG2 was described elsewhere (1). To construct mutant gD-WT, gD-B, and gD-D, gD, truncated gD-B, and gD-D (23) were mutagenized by using the QuikChange site-directed mutagenesis kit (Stratagene, Cedar Creek, Tex.) with a mutagenic primer (CCTGGTCATTTCCGGAATTGTGTAC) to change the seventh Cys codon into Ser (Fig. 1).

HA-tagged gD-D and mutant gD-D with HA tag.

To construct the hemagglutinin (HA)-tagged gD-D into pAc-CMV, gD-D with the HA tag was amplified by a forward primer (GGAATTCATGGGCCTGATCGCCGGCGC) and a reverse primer in which HA was inserted right before the stop codon (GAAGATCTCTAAGCATAATCTGGCACATCATAGTAAAACAAGGGCTGGTGCG). The PCR fragment was inserted into EcoRI and BglII site of pAc-CMV. To construct mutant gD-D with the HA tag, HA-tagged gD-D was mutagenized with the primer described above for mutagenesis. The procedures for the generation of recombinant baculovirus and infection of mammalian cells were as described elsewhere (22).

Immunofluorescence.

A total of 5 × 10⁴ of SK-N-SH were seeded onto four-well glass slides and infected with recombinant baculoviruses for 24 h. Cells were fixed in ice-cold methanol for 20 min at −20°C and then blocked in phosphate-buffered saline (PBS) containing 1% bovine serum albumin at room temperature, rinsed three times with PBS, and reacted for 24 h at 4°C with either a 1:1,000 dilution of gD monoclonal antibody, a 1:2,000 dilution of mouse monoclonal antibody against ICP0, or a 1:500 dilution of HA polyclonal antibody in PBS. The cells were rinsed five times in PBS, reacted for 1 h with 1:64 dilution of a goat anti-mouse immunoglobulin G (IgG) or 1:160 dilution of an anti-rabbit IgG conjugated to fluorescein isothiocyanate (Sigma, St. Louis, Mo.) in PBS, and then rinsed five times with PBS and mounted in 90% glycerol. Slides were analyzed in a Zeiss confocal fluorescence microscope.

Double infection.

Subconfluent cultures of SK-N-SH cells in 25-cm² flasks were first exposed at 37°C for 2 h to 10 PFU of recombinant baculovirus per cell and then to either 100 PFU equivalent of gD^−/− as indicated. The cultures were then maintained for an additional 18 h at 37°C in medium containing 2.5 mM sodium butyrate.

Immunoprecipitation assay.

Subconfluent cultures of SK-N-SH cells in 25-cm² flasks were infected either with 10 PFU of recombinant baculovirus gD-B plus HA-gD-D, mutant gD-B plus HA-gD-D, gD-B plus mutant HA-gD-D, or mutant gD-B plus mutant HA-gD-D as indicated. The cells were harvested 24 h postinfection and washed with 5 ml of PBS twice. The pellets were resuspended with 200 μl of lysis buffer (20 mM Tris[pH 8.0], 1 mM EDTA, 1% NP-40, 400 mM NaCl, 2 mM dithiothreitol [DTT], 0.1 mM NaVO₄, 10 mM NaF, 1× protease inhibitor cocktail [Sigma, St. Louis, Mo.]) and chilled on ice for 40 min; this was followed by centrifugation at 1,000 rpm for 2 min. A total of 150 μl of cell lysate was diluted by 150 μl of low-salt lysis buffer (20 mM Tris[pH 8.0], 1 mM EDTA, 1% NP-40, 16 mM NaCl, 2 mM DTT) and incubated with 5% rabbit preimmune serum at 4°C for 1 h, followed by incubation with 50 μl of protein A-Sepharose for 1 h at 4°C and centrifugation at 3,000 rpm for 3 min to remove nonspecifically bound proteins. The supernatant was incubated with HA-specific polyclonal antiserum at 4°C for 16 h, and then incubated with 20 μl of protein A-Sepharose at 4°C for 1 h. Antigen-antibody complexes were rinsed three times with rinse buffer (50 mM Tris[pH 7.4], 10 mM MgCl₂, 5 mM DTT) and pelleted by centrifugation at 3,000 rpm for 3 min. The antigen-antibody complexes were disrupted by boiling the pellet in sample buffer for 5 min. Supernatant fluids containing precipitated interaction proteins were recovered after centrifugation for 1 min and analyzed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blots were probed with monoclonal antibodies to gD to identify associated proteins as described as elsewhere (22).

gD-B and gD-D functional complementation assay.

Replicate subconfluent cultures of SK-N-SH cells in 25-cm² flasks were either mock infected or exposed to recombinant baculoviruses (15 PFU/cell) that expressed gD, gD-B plus gD-D, mgD-B plus gD-D, gD-B plus mgD-D, or mgD-B plus mgD-D. After 2 h the cells were exposed to 10 PFU of gD^−/+ stock. per cell. The cells were the thoroughly rinsed with 1 M citrate acid (pH 3.0) for 1 min, followed by three rinses with PBS to remove unattached virus. After the rinses, the cells were incubated in DMEM containing 10% fetal bovine serum and harvested 18 h after infection. The frozen-thawed, sonicated cell lysates were used to infected SK-N-SH cells that were seeded on four-well glass slides. After 12 h of incubation the slide cultures were fixed and reacted with rabbit polyclonal antibody against ICP0 and then examined with the aid of a Zeiss confocal microscope.

RESULTS

Expression of Cys7Ser mutant gD, gD-B, gD-D, HA-tagged gD-D, and mutant gD-D.

As indicated in the introduction, the objective of the studies described here was to test the hypothesis that gD can form a heterodimer consisting of the ectodomain plus transmembrane domain (gD-B) and transmembrane domain plus cytoplasmic domain (gD-D) and linked through the seventh cysteines located in the transmembrane domain. To test this hypothesis, we constructed two sets of mutated versions of gD-WT, gD-B, and gD-D structures. In the first set designated mgD, mgD-B, and mgD-D, the seventh cysteine was replaced by serine as described in Materials and Methods. In the second set, we inserted an HA tag immediately before the stop codon in mgD-D and gD-D. These constructs were designated HA-mgD-D and HA-gD-D. All five constructs were sequenced and found to be exactly as predicted and cloned into baculoviruses as described in Materials and Methods.

As shown in Fig. 2, all constructs were expressed in SK-N-SH cells and reacted with either monoclonal antibody against gD (panels A to C) or the anti-HA tag (panels D and E). In each instance, intact gD or the truncated forms of gD accumulated in the cytoplasm.

FIG. 2.

Digitized images of cells infected with recombinant baculovirus as indicated. The cells were exposed to recombinant baculoviruses encoding intact gD or truncated forms as indicated. After 24 h of incubation, the cells were fixed and reacted with either anti-gD monolconal antibody (A to C) or anti-HA antibody (D and E) and then reacted with anti-IgG antibody conjugated to fluorescein isothiocyanate. The images were collected with a Zeiss confocal microscope as described in Materials and Methods.

Effects of the substitution of Cys7 with serine on the ability of intact and truncated forms of gD to block apoptosis induced by gD^−/− and gD^−/+ stocks.

The key questions addressed in this report is whether the seventh cysteine located in the transmembrane domain is essential for blocking apoptosis induced by gD^−/− or gD^−/+ viruses. Replicate cultures of SK-N-SH cells were exposed to 10 PFU of one or two recombinant baculoviruses per cell for 2 h and then infected with 100 PFU equivalents of gD^−/− stock per cell. The cells were harvested 24 h after gD^−/− virus infection and examined for the presence of fragmented DNA. As shown in Fig. 3A, baculoviruses expressing intact gD (gD-WT, lane 2) or Cys7Ser mutant (mgD, lane 3) blocked DNA fragmentation. Consistent with the earlier report, apoptosis induced by gD^−/− stock was blocked by exposure of cells to gD-B and gD-D (Fig. 3A, lane 5). However, exposure of cells to a mixture of Cys7 mutant gD-B (mgD-B) and wild-type gD-D (gD-D), to wild-type gD-B (gD-B) and Cys7 mutant gD-D (mgD-D), or to mutant gD-B (mgD-B) and mutant gD-D (mgD-D) was unable to block DNA fragmentation from gD^−/− virus infection (Fig. 3A, lanes 6, 7, and 8).

FIG. 3.

Agarose gels containing electrophoretically separated low-molecular-weight DNA from lysates of mock-infected cells or cells infected first with recombinant baculoviruses and then with the indicated gD^−/− virus stocks. Replicate cultures of subconfluent SK-N-SH cells in 25-cm² flasks were infected with 10 PFU of the indicated recombinant baculovirus per cell. After 2 h, the cells were infected with 100 PFU equivalents of gD^−/− per cell. The cells were harvested 24 h after gD^−/− infection and processed as described in Materials and Methods.

The second series of experiments was designed to determine whether HA-tagged gD-D construct retain the key properties of gD. Replicate cultures of SK-N-SH cells were exposed to 10 PFU each of recombinant baculoviruses indicated in Fig. 3B. The cells were infected with gD^−/− stocks and processed as described above. The results (Fig. 3B) show that the mixture of gD-B plus HA-gD-D blocked apoptosis (Fig. 3B, lane 3) whereas the mixture gD-B plus HA-mgD-D did not.

We conclude from this series of experiments the following.

(i) Full-length gD carrying the substitution Cys7Ser can block apoptosis, as does wild-type gD.

(ii) The mixture of the gD-B and gD-D can block apoptosis, but substitution of either component of the mixture with a mutant carrying the Cys7Ser no longer blocks apoptosis.

(iii) Insertion of the HA tag into the carboxyl terminus of gD-D construct has no effect on its function in the assays performed as described above.

The truncated forms gD-B and gD-D can form a heterodimer that is dependent on the presence of Cys7 in both constructs.

To verify the interaction of the truncated forms of gD-B and gD-D and to investigate further the role of Cys7 in the interaction between these constructs, SK-N-SH cells were mock infected or were infected with baculoviruses expressing intact gD or gD-B or mixtures of baculoviruses expressing gD-B and HA-gD-D, mgD-B and HA-gD-D, gD-B and HA-mgD-D, or mgD-B and HA-mgD-D. After 24 h of incubation at 37°C, the cells were harvested, lysed, and reacted with anti-HA antibody. The precipitates were collected, subjected to electrophoresis on a denaturing polyacrylamide gel, transferred to a nitrocellulose sheet, and reacted with anti-HA antibody. The salient features of the results shown in Fig. 4 were as follows.

FIG. 4.

Detection of gD-B/gD-D heterodimers in cells doubly transduced by recombinant baculoviruses encoding the individual polypeptides. SK-N-SH cells in 25-cm² flasks were either mock infected or doubly infected with 10 PFU of indicated baculoviruses per cell (lanes 1 to 5). After 24 h, the cells lysates were reacted with anti-HA antibody. The precipitated proteins were separated on denaturing sodium dodecyl sulfate-12% polyacrylamide gel and reacted with antibody to gD as detailed in Materials and Methods. In addition, SK-N-SH cells were infected with either recombinant baculoviruses expressing gD or recombinant baculoviruses expressing gD-B and then reacted with antibody to gD (lanes 6 and 7). The arrow points to gD-B coprecipitated with HA-tagged gD-D. Note that the coprecipitated gD-B migrated more slowly than predicted most likely because it was more extensively glycosylated.

(i) The anti-HA antibody precipitated both HA-tagged gD-D and gD-B from lysates of cell infected with the mixture of baculoviruses expressing the two truncated forms of gD (Fig. 4, lane 2).

(ii) The anti-HA antibody did not precipitate gD-B from lysates of cells exposed to baculoviruses in which either one or both constructs contained the Cys7Ser substitution.

We conclude that gD-B and gD-D truncated forms of gD can form heterodimers that is dependent on the presence of Cys7 in both constructs.

The gD-B/gD-D heterodimer partly complements gD with respect to virus entry.

As stated above, gD^−/+ stocks can infect cells and produce progeny virions. However, infection of noncomplementing cells with gD^−/+ stocks results in the production of gD^−/− virions. Infection of complementing cells, i.e., cells that express gD, results in the generation of gD^−/+ progeny virus. Although gD^−/− virions can be taken up by endocytosis, they are ultimately destroyed, and no infectious progeny ensues (22). The objective of the experiments described below were to determine whether the heterodimer gD-B plus gD-D can substitute for gD to confer infectivity on the progeny of gD^−/+ virus stocks in noncomplementing cells. Briefly, as described in Materials and Methods, SK-N-SH cells were either mock infected or exposed to baculoviruses (15 PFU/cell) expressing gD only, gD-B plus gD-D, mgD-B plus gD-D, gD-B plus mgD-D, or mgD-B plus mgD-D. After 2 h of incubation, the cells were exposed to PFU of gD^−/+ virus per cell. After three additional hours of incubation the cells were thoroughly rinsed with 1 M citrate (pH 3.0) to remove unattached virus. After an additional 18 h of incubation, the cells were harvested and lysed by freeze-thawing and sonication. The cell lysates were then used to infect SK-N-SH cells grown in four-well slide cultures. After 12 h of incubation, the slide cultures were reacted with anti-ICP0 antibody and examined with the aid of a Zeiss confocal microscope. The results shown in Fig. 5 indicate that only SK-N-SH exposed to lysates of cells transduced with wild-type gD and infected with gD^−/+ stock or transduced with gD-B plus gD-D and infected with gD^−/+ stock expressed ICP0. The number of cells expressing ICP0 in cultures exposed to lysates of cells transduced with the heterodimer and infected with gD^−/+ was approximately 2% of the total. None of the other lysates tested in this experiment yielded infectious progeny capable of infecting SK-N-SH cells, as evidenced by the expression of ICP0.

FIG. 5.

Digitized images of SK-N-SH cells in four-well slides infected with lysates of cells transduced with baculoviruses and infected with gD^−/+ stock. The timeline and experimental design are shown in the top portion of the figure. The slide cultures were reacted with anti-ICP0 antibody and examined for the accumulation of the protein by immunofluorescence with the aid of a Zeiss confocal microscope. The results obtained with recombinant baculovirus expressing mgD or the mixture of baculoviruses expressing gD-A and gD-B were negative and are not shown.

We conclude that the heterodimer formed by gD-B and gD-D can complement gD^−/+ virus at a low efficiency.

DISCUSSION

The experimental results that led to these studies were as follows.

(i) The components of gD required to block apoptosis induced by gD^−/− and gD^−/+ mutants virus stocks differed. The implications of the finding that the site or compartment in which gD must act to ablate apoptosis are very different for the two virus stocks. In the case of the gD^−/+ stocks, the delivery of a gene (gD-A) encoding solely the ectodomain of gD was sufficient to block apoptosis (23). In the absence of a transmembrane domain, gD-A would be expected to enter the secretory pathway and be available to interact with proteins embedded in membranes of the exocytic pathway. On the basis of the observation that gD interacts with M6PR and that over expression of M6PR blocks apoptosis induced by gD^−/+ mutant virus stock (24), a current model envisions that the soluble ectodomain interacts in the exocytic pathway with M6PR and precludes the lysosomal enzymes from destroying the cell. In contrast, gD-A does not block apoptosis induced by gD^−/− virus stocks. Inasmuch as gD^−/− viruses (i) are taken up in endosomal vesicles and destroyed and (ii) chloroquine blocks apoptosis induced by this mutant virus stock, the data indicate that the induction of apoptosis is a consequence of the fusion of lysosomes with early endosomal vesicles carrying endocytosed virions. This model predicts that, to block apoptosis, the gD must be present in endocytic vesicles and that it arises in this compartment either by being brought with the virion or, in the case of the gD gene delivered in trans, by being incorporated into the membrane of the endocytic vesicle. Consistent with this model, the gD-A construct does not block apoptosis since the truncated form of gD consisting solely of the ectodomain would not be present in the endosomal compartment.

(ii) The gD constructs that blocked apoptosis induced by gD^−/− mutant virus stocks are intact gD or a mixture of gD-B and gD-D. As noted above, the observation that intact gD blocks apoptosis induced by gD^−/− stocks is consistent with the model that newly made gD would be transported through the exocytic pathway to the plasma membrane and become incorporated in the membranes of the endocytic vesicles. Less comprehensible is the evidence that, in the case of subunits of gD, both gD-B and gD-D are required to block apoptosis. The results suggest that in order to reach the target compartment necessary to block apoptosis, gD-B and gD-D forms a multimeric complex that acquires some of the attributes of intact gD.

The experiments described here show that this is in fact the case. Although assembly of multimeric complexes could arise through a number of different interactions between gD-B and gD-D, the evidence presented elsewhere indicated that only mixtures in which both constructs are flanked by a transmembrane domain block apoptosis induced by gD^−/− virus stocks (23). Other studies have also shown that of the seven cysteines contained in HSV-1 gD, the seventh cysteine (Cys7) located in the transmembrane domain is unpaired in the monomeric structure of the protein (11, 12). Moreover, mutation of the cysteine codon precludes dimerization of gD (21). In this report, we show that substitution of the cysteine with serine in even one of the bipartite components, either in gD-B or gD-D resulted in a loss of the capacity of the mixture to block apoptosis. Unambiguous evidence that gD-B and gD-D for a heterodimer emerged from the studies showing that gD-B was coprecipitated by antibody directed to the HA-tagged gD-D from lysates of cells transduced by mixtures of baculoviruses encoding gD-B and gD-D. The gD-B component was not precipitated if the Cys7 was replaced with serine in either one of the two interacting proteins. Finally, consistent with the model, we showed that intact gD or the heterodimer consisting of gD-B and gD-D was transported to the cell surface. The mixtures of gD-B or gD-D that were not able to form a heterodimer because of Cys7Ser substitution were not transported to the cell surface.

We conclude from these studies that gD-B and gD-D can form a heterodimer that is dependent on Cys7 for its formation, and that the heterodimer gD-B/gD-D has the attributes of intact gD with respect to the ablation of apoptosis and partial complementation of infectivity. The results also suggest that the enctodomain and the cytoplasmic domain function independently of each other. The data also indicate, as has been reported earlier, the gD can form homodimers dependent on Cys7. Our results, however, indicate that the formations of homodimers that are dependent on Cys7 are not essential for ablation of apoptosis. The antiapoptotic functions described and studied in this, and preceding reports are expressed by the monomeric functions of gD.