The Porcine Humoral Immune Response against ... - Journal of Virology

JOURNAL OF VIROLOGY, Feb. 2000, p. 1752–1760 0022-538X/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 4

The Porcine Humoral Immune Response against Pseudorabies Virus Specifically Targets Attachment Sites on Glycoprotein gC ¨ LLER,§ GU ¨ NTHER JUNG,§ BERTRAM T. OBER,† BERTHOLD TEUFEL,‡ KARL-HEINZ WIESMU ¨ LLER, AND HANNS-JOACHIM RZIHA* EBERHARD PFAFF, ARMIN SAALMU Federal Research Centre for Virus Diseases of Animals, Institute of Immunology, D-72 076 Tu ¨bingen, Federal Republic of Germany Received 11 August 1999/Accepted 22 November 1999

High titers of virus-neutralizing antibodies directed against glycoprotein gC of Pseudorabies virus (PRV) (Suid herpesvirus 1) are generally observed in the serum of immunized pigs. A known function of the glycoprotein gC is to mediate attachment of PRV to target cells through distinct viral heparin-binding domains (HBDs). Therefore, it was suggested that the virus-neutralizing activity of anti-PRV sera is directed against HBDs on gC. To address this issue, sera with high virus-neutralizing activity against gC were used to characterize the anti-gC response. Epitope mapping demonstrated that amino acids of HBDs are part of an antigenic antibody binding domain which is located in the N-terminal part of gC. Binding of antibodies to this antigenic domain of gC was further shown to interfere with the viral attachment. Therefore, these results show that the viral HBDs are accessible targets for the humoral anti-PRV response even after tolerance induction against self-proteins, which utilize similar HBDs to promote host protein-protein interactions. The findings indicate that the host’s immune system can specifically block the attachment function of PRV gC. Since HBDs promote the attachment of a number of herpesviruses, the design of future antiherpesvirus vaccines should aim to induce a humoral immune response that prevents HBD-mediated viral attachment. well as vaccination with DNA encoding gC were found to provide protection against PRV challenge (11, 16, 32). Several immune mechanisms directed against gC have been reported to combat PRV infection. PRV infection can induce MHC class I restricted, gC-specific, cytotoxic T cells (52). It was also shown that gC is involved in priming of porcine Thelper cells (17) and induces MHC class II restricted, PRVspecific memory T-helper cells (30). In addition to its role in the specific cell-mediated immune response, gC also is one of the major viral proteins that trigger the porcine humoral immune response. In pigs, the majority of virus-neutralizing antibodies induced by PRV infection or vaccination with attenuated live virus is directed against gC (1). Passive immunization of pigs with gC-specific mouse monoclonal antibodies was sufficient to confer protection (24), indicating that gCspecific antibodies, even in the absence of a cell-mediated immune response, play an essential role in the control of PRV infection. PRV gC, which is a constituent of the viral envelope (34, 35), is nonessential for virus growth in tissue culture (34, 44) but is required for maximum infectivity, as mutant virions lacking gC have been shown to be defective in attachment to target cells (25). The attachment defect of gC-negative virions is due to a failure to bind to heparan sulfate glycosaminoglycan chains on the target cell (28). It was further shown that the heparanbinding protein gC is promoting the initial attachment of the virus (13) and that three functional heparin-binding domains (HBDs), located in the amino-terminal part of gC (Fig. 1), mediate this interaction (7, 8, 21). The single domains are each composed of a cluster of positively charged viral amino acids with homology to the HBD consensus sequence (2). Each separate HBD is able to promote attachment to target cells (8). In the present study, we identified an antigenic, antibody binding domain (ABD) on gC, and we showed that the HBDs, which are located in this ABD, are recognized by the humoral immune response of the natural host, swine. Antibodies binding to HBDs interfered with viral attachment and prevented

Herpesviruses persist in the natural host for the life of the host, and therefore, they must be able to evade the host’s immune mechanisms directed against the virus and virus-infected cells. Besides establishing a latent infection and thereby escaping direct immune attack, herpesviruses employ a variety of strategies to directly interfere with the immune system, e.g., expression of functional homologues or receptors of cytokines, chemokines, and the major histocompatibility complex (MHC) class I molecules (5). Such sophisticated evasion strategies can explain why vaccination against herpesviruses has only limited success. Although vaccines can prevent symptoms of the disease, they do not eliminate the virus from the infected host and do not prevent the establishment of a latent infection. Pseudorabies virus (PRV), a member of the Alphaherpesvirinae, is the causative agent of Aujeszky’s disease. In pigs, the natural host of PRV, mortality decreases with increasing age, and upon primary infection a latent infection that can be reactivated is generally established (45). Since outbreaks of Aujeszky’s disease cause important economic losses, attenuated live viruses or inactivated vaccines are widely used for vaccination of pigs. Glycoprotein gC of PRV plays an important role in eliciting a protective immune response in pigs. Vaccinations with protein fractions enriched for PRV gC, recombinant gC subunit vaccine, and gC vaccinia virus recombinants as

* Corresponding author. Mailing address: Federal Research Centre for Virus Diseases of Animals, Institute For Immunology, Paul-Ehrlich-Strasse 28, D-72 076 Tu ¨bingen, Federal Republic of Germany. Phone: 49-7071-967 253. Fax: 49-7071-967 303. E-mail: achim.rziha @tue.bfav.de. † Present address: Gwen Knapp Center for Lupus and Immunology Research, Department of Pathology, The University of Chicago, Chicago, IL 60637. ‡ Present address: Pathogenix Laboratories GmbH, D-82152 Martinsried, Federal Republic of Germany. § Present address: Institute for Organic Chemistry, University of Tu ¨bingen, D-72076 Tu ¨bingen, Federal Republic of Germany. 1752

VOL. 74, 2000

SERUM ACTIVITY AGAINST PRV TARGETS HBDs ON gC

1753

FIG. 1. Schematic representation of the gC protein. The location of the three fusion proteins gC-N-term, gC-middle, and gC-C-term is indicated by the respective amino acid positions. The positions delineating the HBDs HBD 1 (H1), HBD 2 (H2), and HBD 3 (H3) are shown as described previously (8), as is that of the identified antigenic ABD. Segments found to react with the porcine immune sera are shaded.

infection of target cells. The results demonstrate that the gCspecific neutralization of PRV by the humoral immune response is linked to the known attachment function of this glycoprotein. Blocking of HBDs is therefore one of the strategies the host immune system utilizes to combat PRV infection. MATERIALS AND METHODS Viruses and cell culture. The porcine kidney cell line PSEK and the MadinDarby bovine kidney cell line MDBK were used for virus propagation and titration as described recently (30). The ␤-galactosidase-expressing recombinant PRV PHY-B111 is derived from the mutant PHY-EBG, which carries the lacZ gene in a nonessential, intergenic region of the wild-type (WT) PRV strain Phylaxia (9). To improve the lacZ gene expression in PHY-B111, the simian virus 40 polyadenylation signal was inserted at the 3⬘ end of the lacZ gene. In vitro and in vivo, both lacZ mutants of PRV demonstrate the same growth characteristics and virulence as the parental strain Phylaxia (H.-J. Rziha and A. Six, unpublished results). The gE-negative and thymidine kinase-negative live vaccine of PRV (Begonia, Nobi-Porvac live; Intervet, Boxmeer, The Netherlands), an virulent derivative of PRV strain NIA3 (29), was kindly provided by N. Visser (Intervet), and it has been reported many times that the induced serum antibodies efficiently neutralize various WT PRV strains. WT PRV strain Kapan (Ka) and the isogenic gC deletion mutant (38) were provided by T. C. Mettenleiter (Federal Research Centre For Virus Diseases of Animals, Insel Riems, Federal Republic of Germany). Construction of gC expression plasmids. Fragments of the open reading frame of gC were cloned by standard procedures into the expression vector pEV40, a derivative of plasmid pEx34, as described previously (10, 40). Nucleotide sequence comparison demonstrated the high preservation of the gC gene of various PRV strains (including strains NIA3 and Becker), though some minor base differences exist (12). Similarly, in contrast to the published sequence of the PRV strain Becker (34), PRV strain Ka carries a silent nucleotide exchange at nucleotide 330 (G to A), which generates an additional SacI restriction site. This allowed cloning of the following three gene segments, covering 95% of the gC coding region, as shown in Fig. 1: the 335-bp N-terminal SacI fragment (pOBEB) comprising amino acids 44 to 156 and designated gC-N-term; the 402-bp SacI fragment (pG32-40) comprising amino acids 156 to 290 and designated gC-middle; and the 448-bp C-terminal SmaI-BamHI fragment (pOB-SB) comprising amino acids 309 to 459 and designated gC-C-term. DNA sequencing proved the correct insertion of each gC fragment (data not shown). Expression, purification, and Western blot analysis of gC fusion proteins. After heat induction, the MS-2 polymerase-gC fusion proteins were isolated, and the insoluble protein fraction was dissolved in 7 M urea and used for Western blot analysis as described previously (10). Synthesis and analysis of synthetic peptides and lipopeptides. Overlapping peptide amides (15 amino acids in length, overlapping each other by 10 amino acids) and the lipopeptide {N-palmitoyl-S-[2,3-bis(palmitoyloxy)-(RS)-propyl](R)-cysteinyl} of peptide 65-79 (covering amino acids 65 to 79 of gC) were synthesized according to the gC sequence of PRV strain Ka (34) by using Fmoc chemistry (multiple peptide synthesizer SMPS 350 A; Zinsser, Frankfurt, Germany) as described recently (30). The synthesized peptides and lipopeptides were analyzed by amino acid analysis (ABI 420A; Applied Biosystems, Weiterstadt, Germany), analytical high-performance liquid chromatography (System

Gold; Beckman, San Ramon, Calif.) on a Nucleosil C18 column (Grom, Herrenberg, Germany), and ion-spray mass spectrometry (API III Triple-Quatrupol Ionspray-MS; Grom). The peptides displayed a purity of at least 80%. Animals and antisera. Porcine anti-PRV sera were obtained by immunization of three d/d haplotype National Institutes of Health minipigs (inbred 1, 2, and 3) (37) and two outbred pigs (German landrace) (outbred 1 and 2) with the PRV live vaccine Nobi-Porvac live (Intervet). The vaccine was administered twice at an interval of 10 days, as recommended by the manufacturer. Mouse antisera directed against the region comprising amino acids 65 to 79 of PRV gC (peptide 65-79) were generated using two different protocols. First, BALB/c mice were immunized with 100 ␮g of lipopeptide 65-79 together with lipopeptide adjuvant three times at 2-week intervals as described previously (6). Second, peptide 65-79 was coupled through an additional cysteine at the N terminus to the carrier protein keyhole limpet hemocyanin (KLH) using mmaleimidobenzyl-N-hydroxysuccinimide ester (23). The mice were initially primed with 50 ␮g of peptide-coupled protein in complete Freund’s adjuvant and boosted with 50 ␮g of peptide-coupled protein in incomplete Freund’s adjuvant at 2-week intervals. Production and characterization of the polyclonal goat antiserum against PRV, which neutralized PRV in the presence of complement up to a serum dilution of 1:20,000 and in the absence of complement in a serum dilution of 1:5,000, has been described previously. Serum samples of the immunized pigs were collected before and 1 week after each immunization. Prior to use, serum complement was inactivated by heating of the sera for 1 h at 56°C. ELISA. Microtiter plates (96-well) were coated with synthetic peptides and purified PRV at concentrations of 200 and 5 ␮g per ml of phosphate-buffered saline, respectively, and standard enzyme-linked immunosorbent assay (ELISA) was performed as described previously (22). Bound antibodies were detected with rabbit anti-porcine, anti-goat, or anti-mouse immunoglobulin-coupled horseradish peroxidase (Dianova, Hamburg, Federal Republic of Germany). The ELISA titer of the sera was determined as the highest dilution of the serum displaying an optical density at a wavelength of 490 nm (OD490) that was approximately twice that of the corresponding preimmune serum. Virus neutralization test. The virus-neutralizing activity of the anti-PRV sera was tested after preincubation with PRV for 1 h at 37°C in the absence of complement on MDBK cells (26). Virus adsorption was allowed for 1 h on ice to delay virus penetration (50). After removal of excess virus by washing with media, the cells were incubated at 37°C. The neutralization titer was determined from triplicates and expressed as the dilution of serum resulting in 95 to 100% plaque reduction. Dilutions of preimmune sera from the inbred and outbred pigs as well as medium were used as controls. Competitive attachment inhibition assay. The heat-inactivated goat anti-PRV serum was diluted 1:50 in medium, the peptide was added at the indicated concentrations, and the mixtures were incubated overnight at 4°C. Thereafter, the lacZ gene-expressing PRV strain PHY-B111 (1.5 ⫻ 107 PFU) was added, and the mixture was incubated at 37°C for 30 min and then used for infection of 106 MDBK cells at 2°C. After 1 h of adsorption, the cells were washed with phosphate-buffered saline and incubated for an additional 4 h at 37°C. After trypsinization, the cells were hypotonically treated with 2 mM fluorescein-di-(␤D-galactopyranoside) (Sigma-Aldrich, Munich, Federal Republic of Germany), as described recently (36), which treatment allows quantitation of ␤-galactosidase-expressing cells by flow cytometry.

1754

OBER ET AL.

J. VIROL. TABLE 1. Reactivity of anti-PRV sera

Anti-PRV serum

Inbred pig 1 2 3 Outbred pig 1 2 Goat

Result of:

Virus neutralizationa

Western blot analysis

b

Peptide ELISAc

WT

gC-negative

gC-N-term

gC-middle

gC-C-term

65-79

80-94

85-99

960 960 NDd

120 20 ND

⫹⫹⫹ ⫹ ⫹

⫺ ⫹⫹⫹ ⫹⫹

⫺ ⫺ ⫺

⫹⫹⫹ ⫺ ⫺

⫺ ⫹⫹⫹ ⫺

⫺ ⫹⫹⫹ ⫺

480 480

120 60

⫹ ⫹⫹

⫹⫹ ⫹

⫺ ⫺

⫹ ⫺

⫹⫹ ⫹⫹⫹

⫹⫹ ⫹⫹⫹

5,000

640

⫹⫹⫹

⫹

⫺

⫹⫹⫹

⫹

⫹

a

Values are the reciprocal serum dilutions which resulted in complete virus neutralization (100% plaque reduction) of WT or gC-negative PRV. Western blot analysis of gC fusion proteins, which were not (⫺), weakly (⫹), moderately (⫹⫹), or strongly (⫹⫹⫹) recognized by the different immune sera. The goat anti-PRV serum was used in a dilution of 1:5,000. c Results from ELISAs with peptides 65-79, 80-94, and 85-99, showing the difference in OD490 between preimmune and immune serum (1:100 dilution), as follows: ⬍0.1 (⫺), 0.1 to 0.25 (⫹), 0.25 to 0.5 (⫹⫹), and ⬍0.5 (⫹⫹⫹). The goat anti-PRV serum was diluted 1:5,000. d ND, not determined. b

RESULTS PRV gC represents a target for virus-neutralizing antibodies. Several reports indicate the importance of gC and gCspecific antibodies in PRV neutralization. To gain a better understanding of the natural humoral immune response, which is directed against gC during PRV infection, we generated anti-PRV sera of three d/d haplotype inbred pigs (Table 1, inbred pigs 1 to 3) and two German landrace pigs (Table 1, outbred pigs 1 and 2). All immune sera displayed PRV-specific antibody titers of more than 1:10,000 as determined by ELISA (data not shown). The presence of gC-specific virus-neutralizing antibodies was tested by comparing the capacities of the different antisera to neutralize WT PRV and an isogenic gC deletion mutant in a plaque-reduction assay. Since gC is involved in viral attachment, virus adsorption was performed in the cold in order to prevent premature penetration of PRV during the infection. Excess virus was subsequently removed and incubated at 37°C to allow virus multiplication. The swine anti-PRV sera neutralized WT PRV in the absence of complement up to a serum dilution of 1:480 or 1:960 (Fig. 2; Table 1). In contrast, a reduction in serum neutralization titer was reproducibly found when the gC-negative PRV was used (Fig. 2; Table 1). The reduction of virus-neutralizing serum antibody titers ranged from 4-fold (serum from outbred pig 1) to 48-fold (serum from inbred pig 2). Using the preimmune sera from inbred and outbred pigs, no plaque reduction was found. These results therefore demonstrate that the immunization led to the induction of a humoral virus-neutralizing antibody response directed against gC. Major B-cell epitopes are located in the N-terminal part of gC. To elucidate the gC-specific reactivity of the different antiPRV sera, three fragments of gC encoding the amino acids 44 to 156 (gC-N-term), 156 to 290 (gC-middle), and 309 to 459 (gC-C-term) (Fig. 1) were expressed as insoluble bacterial MS-2 polymerase-gC fusion proteins (see Materials and Methods). The Coomassie blue-stained sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE) gel showed that the fusion proteins gC-C-term and gC-middle were abundantly expressed, whereas gC-N-term (23 kDa in size) could be induced only to moderate amounts (Fig. 3A). These specific protein bands were not detected in bacteria containing the vector plasmid without insertion or in noninduced bacterial lysates (data not shown). To map the gC-specific humoral

antibody response, the three fusion proteins were tested by Western blot analysis with the different antisera. The results are shown representatively in Fig. 3B to D, which were derived from different Western blotting experiments. Whereas the swine preimmune sera did not react with the gC-specific fusion proteins (Fig. 3B), all tested immune sera displayed two distinct patterns of reactivity with gC-N-term and gC-middle, respectively. First, inbred 1 and outbred 2 reacted strongly with gC-N-term and only weakly or not at all with gC-middle (Fig. 3C; Table 1). Second, inbred 2, inbred 3, and outbred 1 pig sera showed a strong reactivity with gC-middle but only weakly

FIG. 2. Porcine immune sera exhibit reduced capacity to neutralize gC-negative PRV compared to WT PRV. Immune sera from inbred pigs 1 and 2 (numbered circles) and from outbred pigs 1 and 2 (numbered triangles) were diluted and used for plaque reduction assay (performed in triplicate) with WT PRV and an isogenic gC-negative mutant PRV (gC-minus) as described in Materials and Methods. The reciprocal serum dilutions resulting in 95 to 100% plaque reduction are depicted, as are the standard errors (error bars).

VOL. 74, 2000


1755

FIG. 3. Western blot analysis of gC fusion proteins. The fusion proteins gC-N-term (N), gC-middle (M), and gC-C-term (C) were obtained by urea extraction and analyzed by SDS–15% PAGE followed by Coomassie blue staining (A) or Western blot analysis (B to D). The shown results of the Western blot analysis are representative of the reactivities of all sera tested (Table 1), and the porcine preimmune serum (B) and immune sera from inbred pig 1 (C) and inbred pig 2 (D) were diluted 1:500. The specifically induced proteins are marked by arrowheads. Migration of the molecular weight (MW) standards (in thousands) is indicated to the left.

recognized gC-N-term (Fig. 3D; Table 1). In contrast, none of the tested immune sera reacted specifically with the fusion protein gC-C-term (Fig. 3C and D; Table 1), containing the transmembrane region of gC. Additional, nonspecific protein bands reacted in some Western blotting tests; a protein with an apparent molecular mass of approximately 21 kDa (Fig. 3C and D) variably appeared in different SDS-PAGE gels and reacted in more and less pronounced manners with nonimmune (data not shown) and immune swine sera (Fig. 3C and D), respectively. Taken together, these results indicate that the N-terminal part of gC is predominantly recognized by gCspecific antibodies of the immunized pigs, which is consistent with the location of the cytoplasmic and transmembrane region in the C-terminal part of gC (34). Identification of an ABD in the N-terminal part of gC. To further characterize the humoral immune response against the N-terminal part of gC, overlapping synthetic peptides of gC (amino acids 1 to 310) were tested by ELISA using the porcine anti-PRV sera or the respective preimmune sera. The peptides were 15 amino acids in length and overlapped each other by 10

amino acids. No PRV-specific reaction was found with peptides covering amino acids 1 to 54 and 105 to 310. However, between amino acids 54 and 104, three reactive peptides could be identified. While all peptides resulted in ELISA OD490 values of approximately 0.2 with the preimmune sera (Fig. 4), the peptides 65-79, 80-94, and 85-99 showed a significant increase in reactivity with several of the tested serum samples (Fig. 4; Table 1). Peptide 65-79 (65-STPPVPPPSVSRRKP-79) showed a strong reaction with the serum from inbred pig 1 (OD490, 0.81) (Fig. 4A), a weak reaction with the serum from outbred pig 1 (OD490, 0.3), and no reaction with the sera from inbred pigs 2 and 3 and from outbred pig 2 (Table 1). Peptides 80-94 and 85-99 (80-PRNNNRTRVHGDKATAHGRK-99) reacted strongly with the serum from inbred pig 2 (OD490 values, 0.62 and 0.58, respectively) (Fig. 4B) and outbred pig 2 (OD490 values, 2.16 and 1.98, respectively), a weaker reaction was found with serum from outbred pig 1 (OD490 values, 0.48 and 0.41, respectively), and no specific reaction was found with sera of inbred pigs 1 and 3 (Table 1). All three reacting peptides map between amino acids 65 and 99 of gC. This peptide

1756

OBER ET AL.

J. VIROL.

FIG. 5. The identified ABD is exposed on the surface of PRV virions. Mouse antisera against peptide 65-79 (open symbols) and preimmune sera (closed symbols) were tested by ELISA for reactivity with peptide 65-79 (insert, squares) and an irrelevant control peptide (insert, circles), or with purified PRV (squares) and bovine serum albumin (circles). Each experiment was performed in triplicate, with the error bars indicating the standard deviations. The results are representative for five different mouse anti-peptide 65-79 serum samples.

FIG. 4. Identification of linear B-cell epitopes in the N-terminal part of gC. Overlapping synthetic peptides spanning amino acids 55 to 114 of gC were coated onto 96-well plates and tested by ELISA with diluted (1:200) immune sera (open columns) of inbred pig 1 (A) and inbred pig 2 (B) and with the respective preimmune sera (gray columns). Peptide numbers correspond to the first amino acid position of each peptide in the gC sequence. Each experiment was performed in triplicate, with the standard deviations shown by error bars; no error bar is shown where the standard deviation was less than the extent of the bar.

screening approach therefore allowed us to identify an ABD in the N-terminal region of gC comprising at least amino acids 65 to 99 (Fig. 1). ABD is accessible to humoral immune responses on the surface of PRV. The peptide screening approach did not answer the question of whether the identified ABD is exposed on the virion surface and accessible to the humoral immune response. To answer this question, BALB/c mice were immunized with the synthetic peptide 65-79, either as a lipopeptide

or coupled to the KLH protein as a carrier. The specificity of the obtained antisera was shown by ELISA as depicted representatively in Fig. 5 for one of the mouse anti-peptide 65-79 serum samples. As expected, the peptide antisera strongly reacted with peptide 65-79 in serum dilutions higher than 1:1,000 (up to 1:10,000), whereas the reaction with the control peptide was as negative as the preimmune serum reaction (Fig. 5, insert). More importantly, the generated anti-peptide 65-79 antisera, but not the preimmune sera, reacted specifically with purified PRV (Fig. 5). This priming of anti-PRV specific antibodies by peptide 65-79 immunization demonstrated that the identified ABD on gC is exposed on the virion surface and accessible to gC-specific antibodies. This is corroborated by computer-aided prediction of the secondary structure of gC, indicating that the region of the ABD (amino acids 65 to 99) forms a loop structure with a high surface probability (data not shown). Amino acids of the HBD consensus sequence are directly recognized by anti-PRV sera. The identified ABD overlapped with two of the three known HBDs of gC (Fig. 1). Since the HBD consensus sequence was shown to mediate viral attachment, we asked whether amino acids of ABD can be directly recognized by the anti-PRV sera. A direct involvement of HBD 2 in binding of the antisera appears unlikely, since both peptides 80-94 and 85-94 reacted with most immune sera (Table 1), but only peptide 85-94 is overlapping HBD 2. To test for direct binding of the serum antibodies to HBD 1, shortened and mutated versions of peptide 65-79, which includes 4 of the 6 amino acids of HBD 1 (76-RRKPPR-81), were synthesized (Fig. 1). Due to the strong reactivity with peptide 65-79, the antiserum of inbred pig 1 was used for ELISA with the different mutated peptides. The removal of amino acids 65 and 66, which do not overlap with the HBD, led to a markedly decreased serum reactivity (Fig. 6, peptide 67-79). However, the

VOL. 74, 2000


FIG. 6. Amino acids corresponding to HBD 1 are essential for specific antibody recognition of ABD. Ninety-six-well plates were coated with synthetic peptides 65-79, 65-77, and 67-79 (designated as such because of the amino acids of gC that they comprise) and peptides with mutated amino acid 66 (T to G; 65-79/66G), amino acid 71 (P to G; 65-79/71G), or amino acid 77 (R to E; 65-79/77E) and tested by ELISA with diluted (1:200) immune serum from inbred pig 1 (open columns) and the respective preimmune serum (gray columns). Each experiment was performed in triplicate, with the error bars indicating the standard deviations; no error bar is evident where the standard deviation was less than the extent of the bar. Very similar results were obtained with the goat anti-PRV serum diluted at 1:2,000 (data not shown).

loss of amino acids 78 and 79 of the HBD completely prevented binding of the antisera (Fig. 6, peptide 65-77). In addition, amino acid exchanges at position 66 (T to G, unrelated N-terminal amino acid) or at position 71 (P to G, central amino acid of the predicted loop structure in this region) had either no effect on antibody binding (Fig. 6, peptide 65-79/ 66G) or led to a serum reactivity only slightly reduced (Fig. 6, peptide 65-79/71G) compared to that of the authentic peptide 65-79. In contrast, a change of the central amino acid of HBD 1, amino acid 78, from K to E resulted in the complete loss of reactivity of the anti-PRV serum (Fig. 6, 65-79/78E). Therefore, these results demonstrate that amino acids constituting the HBD 1 domain are critical for the recognition of peptide 65-79 by the immune sera. Consequently, it indicates that the humoral immune response against gC can be directed against HBD 1 on the surface of PRV. Antibodies against ABD can block viral attachment to target cells. The findings that the ABD is expressed on the surface of PRV and that amino acids of HBD 1 are targets for the PRVspecific humoral immune response indicate that gC-specific antibodies directed against the ABD can interfere with attachment of PRV by blocking HBD 1. To test this assumption, MDBK cells were incubated with PRV, which was preincubated with different dilutions of mouse anti-peptide 65-79 serum. To prevent virus penetration, the adsorption was performed in the cold and thereafter shifted to 37°C to allow virus multiplication. The ability of the sera to interfere with viral attachment was measured in a plaque assay. However, even at very low dilution (1:5) of the anti-peptide 65-79 serum, no significant reduction of virus plaques could be observed, either by using the sera generated by lipopeptide injection or by KLH coupling. The failure to prevent PRV attachment with antibodies

1757

against peptide 65-79 might be explained by the described functional redundancy of the three HBDs of gC (8). For efficiently preventing viral attachment, probably all three HBDs have to be blocked simultaneously with an antiserum containing antibodies directed against all HBDs. Using the overlapping peptides, a linear B-cell epitope was not detected in HBD 3 (amino acids 133 to 141). However, this approach is unable to identify conformation-dependent epitopes. The presence of additional conformation-dependent ABDs on gC can be suggested from the Western blot reactivity of the porcine antiPRV sera with the fusion protein gC-middle (linear B-cell epitopes were not detected in this part of gC) (Table 1; Fig. 1). Thus, we wanted to test the involvement of the identified ABD 65-99 in binding of virus-neutralizing antibodies. To this end, we established an assay that allowed quantitation of the capacity of this peptide to interfere with complement-independent virus neutralization of anti-PRV sera. PRV strain PHY-B111, expressing the lacZ gene in a nonessential intergenic region, was used for infection of MDBK cells in the cold upon preincubation with peptide 65-79. After removing unattached virus, virus-infected cells could be detected by flow cytometry through their ␤-galactosidase activity (34). After 4 h of incubation at 37°C (in the absence of antibodies and peptide), 75% of cells were virus infected (Fig. 7A). To enable sensitive determination of peptide binding with virus-neutralizing antibodies, we used a goat anti-PRV serum which showed the same reactivity with the gC-fusion proteins and with the mutated gC peptide 65-79 as the serum from inbred pig 1 but exhibited a fivefold higher virus-neutralization titer in the absence of complement (Table 1). The addition of this antiserum alone neutralized the virus and reduced the concentration of infected cells to 11% (Fig. 7B). To test the ability of peptide 65-79 to interact with virusneutralizing antibodies, the anti-PRV serum was then preincubated with peptide 65-79 or with a control peptide and PHYB111. Preincubation with peptide 65-79 clearly reduced the virus-neutralizing capacity of the anti-PRV serum, depending on the concentration of the peptide added. At the highest peptide concentration (500 ␮g), 54% of the cells were PRV infected, and decreasing peptide concentrations resulted in a reduction of infected cells to 23% (Fig. 7C). In contrast, preincubation with the control peptide did not decrease the virusneutralizing activity of the serum. In this case, the percentage of virus-infected cells ranged from 15 to 18%, not exceeding 25% (Fig. 7D), demonstrating the specificity of peptide 65-79 to inhibit virus neutralization. These results could be also confirmed by the use of a conventional PRV plaque assay as a readout (data not shown). These results therefore demonstrate that antibodies directed against the ABD can interfere with the attachment of PRV. Since the ABD on gC was identified by screening of anti-PRV sera, these results indicate further that virus neutralization by gC-specific antibodies, which are induced after natural infection or vaccination, is at least partly due to antibody binding and thereby to blocking of HBDs on the viral surface. DISCUSSION The glycoprotein gC of PRV is essential for efficient primary attachment of the virus to heparan sulfate glycosaminoglycan chains on target cells (28, 49). This attachment is mediated through three independent HBDs located in the N-terminal part of the protein (7, 8). Among the virion glycoproteins, which are targets for the humoral immune response of pigs, PRV gC is suggested to represent a primary target for virusneutralizing antibodies (1, 45). The gC-specific virus-neutral-

1758

OBER ET AL.

J. VIROL.

FIG. 7. Antibodies directed against ABD interfere with the attachment of PRV. Medium (A) and diluted goat anti-PRV serum (1:50) without peptide (B), with the indicated concentrations of peptide 65-79 (C), or with an irrelevant control peptide (D) were incubated overnight before the lacZ gene-expressing PRV strain PHY-B111 was added. After adsorption to MDBK cells in the cold, excess virus was removed and the cells were incubated at 37°C and analyzed 4 h later by flow cytometry as described in Materials and Methods. The percentage of infected cells is given in the upper right corner of each quadrant. The results are representative of three independent experiments. The y axis shows the relative number of cells; the x axis indicates fluorescence intensity.

izing activity of anti-PRV sera implied that a humoral immune response which can prevent virus attachment by blocking the HBDs of PRV gC is primed. However, not only viral proteins but also self-proteins utilize HBDs to promote protein-target cell interactions. For instance, several chemokines, growth factors, and adhesion molecules use heparan sulfate glycosaminoglycan for attachment to target cells (for a review, see reference 39). Viral HBDs can therefore be targets of the host’s immune response only if they are structurally different from self-HBDs used by the host. Otherwise, tolerance induction, which deletes or anergizes self-reactive B-cells (for a review, see reference 4), not only eliminates autoreactive B cells against self-HBDs but also induces unresponsiveness against viral HBDs.

The objective of the present study was a more detailed analysis of the porcine humoral immune response against gC of PRV. Western blot analysis using bacterial fusion proteins spanning almost the entire gC protein revealed that the humoral anti-gC immune response was directed against the Nterminal part of gC (amino acids 44 to 290), containing the three functional HBDs (8). The binding specificity of the antisera was further analyzed by ELISA tests with overlapping peptides covering the N-terminal antigenic region (amino acids 1 to 304). This approach identified three different antigenic peptides, which are adjacently located and define one antigenic ABD (amino acids 65 to 99). This domain comprises two of the three functional HBDs (HBD 1, amino acids 76 to 81; and HBD 2, amino acids 96 to 100) as described by Flynn and Ryan

VOL. 74, 2000


(8), and is exposed on the virion surface, since antisera specific for peptide 65-79 recognized purified virions. Furthermore, mutational analysis of peptide 65-79 revealed that gC-specific antibodies can bind to amino acids of HBD1. Therefore, this viral HBD appears to be structurally different from cellular HBDs and can be recognized by the porcine humoral immune response. This result implies that although self-proteins and viral proteins can share similar primary HBD consensus sequences (2), tolerance induction does not necessarily occur. This conclusion is in agreement with findings that the threedimensional structures of the heparin binding proteins antithrombin, lipoprotein lipase, and basic fibroblast factor show considerable structural differences in their arrangement of the heparin-binding sites (for a review, see reference 39). After immunization of pigs with a PRV live vaccine, sera containing a significant amount of complement-independent virus-neutralizing activity directed against gC were obtained. This activity could be demonstrated by serum neutralization tests, in which the different antisera displayed a significantly reduced capacity (4- to 48-fold reduction) to neutralize gCnegative PRV compared to that of the isogenic, gC-positive WT virus. Therefore, we asked whether complement-independent neutralization of PRV can be correlated with the binding of antibodies to the identified antigenic domain ABD of gC. Antisera directed against peptide 65-79, however, did not prevent viral attachment, although their binding capacity with PRV could be shown. This suggests that blocking of all three HBDs of gC is necessary to inhibit efficiently attachment of PRV, which is consistent with the existence of three functionally redundant HBDs in gC (8). HBD 2 and HBD 3 of gC are separated by more than 30 amino acids, and therefore, it is likely that antibodies directed against only one HBD are not sufficient to block attachment of the virus. Nevertheless, preincubation with a peptide covering almost the complete HBD 1 interfered with the complement-independent virus-neutralizing activity of a high-titer anti-PRV antiserum. The results of this competition assay showed that antibodies directed against HBDs can effectively impede attachment of PRV in vitro. Furthermore, the presence in sera from PRV-infected pigs of antibodies which recognize these HBDs allow the presumption that the humoral immune response against PRV interferes with the ability of the virus to attach to target cells. Mettenleiter et al. (27) observed an involvement of gC in the virulence of PRV, and attenuated viruses carrying a gC deletion or low-level expression of gC are candidates for anti-PRV vaccines (18). The attenuation of these mutant viruses is believed to be in part due to the loss of an efficient primary attachment of viruses with gC deleted to target cells in combination with an inability of these viruses to promote cell-tocell spread of the virus to secondary target cells by readsorption (33). Our finding that HBDs are neutralized by antibodies raised after vaccination with an attenuated gC-positive live vaccine (thymidine kinase-negative and gE-negative) indicates that blocking of viral attachment is one of the strategies of the host’s immune system to attenuate PRV infection in vivo. Although our results indicate that all the HBDs have to be blocked in order to prevent gC-mediated attachment of the virus in vitro, this might not be necessary in vivo. In a recent study, Trybala et al. (41) demonstrated that recombinant PRV mutants carrying only one functional HBD differ in their capacities to bind to different desulfated heparins. It can be argued that a humoral immune response against only one HBD might already be efficient to limit the range of cells susceptible for virus entry and therefore contributes to protection of the animal against PRV infection. This interpretation is supported by Zsak et al. (48), who demonstrated that a PRV

1759

mutant, which carries a deletion of amino acids 157 to 292 (a deletion which is spatially close to HBD 3 [amino acids 133 to 141]), showed different attachment properties of PRV to distinct cell types. In vivo this altered attachment resulted in a significant decrease in the virulence of this PRV mutant. We showed that blocking of HBD attachment sites of PRV leads in vitro to virus neutralization. However, in vivo blocking of viral HBDs seems not to be sufficient for protection, since high concentrations of gC-specific virus-neutralizing antibodies in the blood cannot prevent PRV infection. PRV does not completely lose its ability to attach to target cells even in the complete absence of gC (13, 14), and antibodies against viral envelope proteins other than gC have been suggested to interfere also with PRV adsorption (50). In addition, a humoral immune response against gC might be effective in preventing a virus infection only if PRV-specific antibodies are present at high concentration at the sites of entry, since spread of PRV into the central nervous system is not dependent on gC (19, 20). The protective effect of humoral antibodies blocking viral HBDs is therefore very likely limited to attenuating the virus. Blocking of the viral attachment sites on the host cell can impede virus entry and might lead to a delayed systemic infection, thereby ensuring that other immune effector mechanisms (e.g., antibody-mediated cellular cytotoxicity or cytotoxic T cells) have sufficient time to combat the virus. Taken together, the results of this study demonstrate that the success of vaccination strategies against PRV, which target the glycoprotein gC, is due to the induction of specific cytotoxic T lymphocytes (51) and to the ability of gC to prime a potent memory T-helper cell response (17, 30). Moreover, the induced gC-specific humoral immune response can also interfere directly with the attachment of PRV mediated by HBDs. The presented findings that HBDs themselves can be immunogenic and can prime a humoral immune response using synthetic peptides should be relevant for the design of improved antiPRV vaccines. It might be feasible to retain or supplement PRV vaccine strains that have gC deleted with HBD containing domains of gC in order to prime a more efficient humoral immune response but still maintain the reduced virulence of this strain. A variety of other herpesviruses (herpes simplex virus type 1 and type 2 [42, 46], human cytomegalovirus [3, 15], varicella-zoster virus [47], bovine herpesvirus 1 [21, 31], and bovine herpesvirus 4 [43]) use heparan sulfate glycosaminoglycan chains on target cells to promote primary attachment, which generally seems to be mediated by HBDs. The concept of combating or at least controlling viral infections by blocking attachment of the virus through the induction of HBD-specific antibodies therefore seems applicable to other members of the herpesvirus family. ACKNOWLEDGMENTS We thank B. Bauer and R. Bernhard for excellent technical assistance; W. Beck, M. Munari, and G.-J. Nicholson for performing the peptide spectrum analysis; and L. Stitz and M. Bu ¨ttner for critically reading the manuscript. This work was partially funded by grant DFG-Rz 2/1-5 from the Deutsche Forschungsgemeinschaft and by grant BRIDGE BIOTCT91 from the European Community. REFERENCES 1. Ben-Porat, T., J. M. DeMarchi, B. Lomniczi, and A. S. Kaplan. 1986. Role of glycoproteins of pseudorabies virus in eliciting neutralizing antibodies. Virology 154:325–334. 2. Cardin, A. D., and H. J. R. Weintraub. 1989. Molecular modeling of proteinglycosaminoglycan interactions. Arteriosclerosis 9:21–32. 3. Compton, T., D. Nowlin, and N. R. Cooper. 1993. Initiation of human cytomegalovirus infection requires initial interaction with cell surface heparansulfate. Virology 193:834–841.

1760

OBER ET AL.

4. Cornall, R. J., C. C. Goodnow, and J. G. Cyster. 1995. The regulation of self-reactive B cells. Curr. Opin. Immunol. 7:804–811. 5. Davis-Poynter, N. J., and H. E. Farrell. 1996. Masters of deception: a review of herpesvirus immune evasion strategies. Immunol. Cell Biol. 74:513–522. 6. Deres, K., H. Schild, K. H. Wiesmu ¨ller, G. Jung, and H. G. Rammensee. 1989. In vivo priming of virus-specific cytotoxic T lymphocytes with synthetic lipopeptide vaccine. Nature 342:561–564. 7. Flynn, S. J., B. L. Burgett, D. S. Stein, K. S. Wilkinson, and P. Ryan. 1993. The amino-terminal one-third of pseudorabies virus glycoprotein gIII contains a functional attachment domain, but this domain is not required for the efficient penetration of Vero cells. J. Virol. 67:2646–2654. 8. Flynn, S. J., and P. Ryan. 1996. The receptor-binding domain of pseudorabies virus glycoprotein gC is composed of multiple discrete units that are functionally redundant. J. Virol. 70:1355–1364. 9. Fuchs, W., B. Bauer, T. C. Mettenleiter, and H.-J. Rziha. 1994. A novel intergenic site for integration and expression of foreign genes in the genome of pseudorabies virus. J. Virol. Methods 46:95–105. 10. Fuchs, W., H.-J. Rziha, N. Lukacs, I. Braunschweiger, N. Visser, D. Lu ¨tticken, C. S. Schreurs, H. J. Thiel, and T. C. Mettenleiter. 1990. Pseudorabies virus glycoprotein gI: in vitro and in vivo analysis of immunorelevant epitopes. J. Gen. Virol. 71:1141–1151. 11. Gerdts, V., A. Jons, B. Makoschey, N. Visser, and T. C. Mettenleiter. 1997. Protection of pigs against Aujeszky’s disease by DNA vaccination. J. Gen. Virol. 78:2139–2146. 12. Ishikawa, K., M. Tsutsui, K. Taguchi, A. Saitoh, and M. Muramatsu. 1996. Sequence variation of the gC gene among pseudorabies virus strains. Vet. Microbiol. 49:267–272. 13. Karger, A., A. Saalmu ¨ller, F. Tufaro, B. W. Banfield, and T. C. Mettenleiter. 1995. Cell surface proteoglycans are not essential for infection by pseudorabies virus. J. Virol. 69:3482–3489. 14. Karger, A., J. Schmidt, and T. C. Mettenleiter. 1998. Infectivity of a pseudorabies virus mutant lacking attachment glycoproteins C and D. J. Virol. 72:7341–7348. 15. Kari, B., and R. Gehrz. 1993. Structure, composition and heparin binding properties of a human cytomegalovirus glycoprotein complex designated gC-II. J. Gen. Virol. 74:255–264. 16. Katayama, S., N. Okada, K. Yoshiki, T. Okabe, and Y. Shimizu. 1997. Protective effect of glycoprotein gC-rich antigen against pseudorabies virus. J. Vet. Med. Sci. 59:657–663. 17. Kimman, T. G., T. M. De Bruin, J. J. Voermans, B. P. Peeters, and A. T. Bianchi. 1995. Development and antigen specificity of the lymphoproliferation responses of pigs to pseudorabies virus: dichotomy between secondary B- and T-cell responses. Immunology 86:372–378. 18. Kit, S., and M. Kit. 1991. Genetically engineered herpesvirus vaccines. Accomplishments in pigs and prospects in humans. Prog. Med. Virol. 38:128– 166. 19. Kritas, S. K., H. J. Nauwynck, and M. B. Pensaert. 1995. Dissemination of wild-type and gC-, gE- and gI-deleted mutants of Aujeszky’s disease virus in the maxillary nerve and trigeminal ganglion of pigs after intranasal inoculation. J. Gen. Virol. 76:2063–2066. 20. Kritas, S. K., M. B. Pensaert, and T. C. Mettenleiter. 1994. Invasion and spread of single glycoprotein deleted mutants of Aujeszky’s disease virus (ADV) in the trigeminal nervous pathway of pigs after intranasal inoculation. Vet. Microbiol. 40:323–334. 21. Liang, X., L. A. Babiuk, and T. J. Zamb. 1993. Mapping of heparin-binding structures on bovine herpesvirus 1 and pseudorabies virus gIII glycoproteins. Virology 194:233–243. 22. Lukacs, N., H. J. Thiel, T. C. Mettenleiter, and H.-J. Rziha. 1985. Demonstration of three major species of pseudorabies virus glycoproteins and identification of a disulfide-linked glycoprotein complex. J. Virol. 53:166–173. 23. Maloy, W. L., J. E. Coligan, and Y. Paterson. 1998. Production of antipeptide antisera, p. 9.4.1–9.4.11. In R. Coico (ed.), Current protocols in immunology. John Wiley and Sons Inc., New York, N.Y. 24. Marchioli, C., R. J. Yancey, Jr., J. G. Timmins, L. E. Post, B. R. Young, and D. A. Povendo. 1988. Protection of mice and swine from pseudorabies virusinduced mortality by administration of pseudorabies virus-specific mouse monoclonal antibodies. Am. J. Vet. Res. 49:860–864. 25. Mettenleiter, T. C. 1989. Glycoprotein gIII deletion mutants of pseudorabies virus are impaired in virus entry. Virology 171:623–625. 26. Mettenleiter, T. C., C. Schreurs, H. J. Thiel, and H.-J. Rziha. 1987. Variability of pseudorabies virus glycoprotein I expression. Virology 158:141– 146. 27. Mettenleiter, T. C., C. Schreurs, F. Zuckermann, T. Ben-Porat, and A. S. Kaplan. 1988. Role of glycoprotein gIII of pseudorabies virus in virulence. J. Virol. 62:2712–2717. 28. Mettenleiter, T. C., L. Zsak, F. Zuckermann, N. Sugg, H. Kern, and T. Ben-Porat. 1990. Interaction of glycoprotein gIII with a cellular heparinlike

J. VIROL. substance mediates adsorption of pseudorabies virus. J. Virol. 64:278–286. 29. Moormann, R. J. M., T. de Rover, J. Briaire, B. P. H. Peters, A. L. J. Gielkens, and J. T. van Oirschot. 1990. Inactivation of the thymidine kinase gene of a gI deletion mutant of pseudorabies virus generates a safe but still highly immunogenic vaccine strain. J. Gen. Virol. 71:1591–1595. 30. Ober, B. T., A. Summerfield, C. Mattlinger, K.-H. Wiesmu ¨ller, G. Jung, E. Pfaff, A. Saalmu ¨ller, and H.-J. Rziha. 1998. Vaccine-induced, pseudorabies virus-specific, extrathymic CD4⫹CD8⫹ memory T-helper cells in swine. J. Virol. 72:4866–4873. 31. Okazaki, K., E. Honda, and Y. Kono. 1994. Heparin-binding domain of bovid herpesvirus 1 glycoprotein gIII. Arch. Virol. 134:413–419. 32. Riviere, M., J. Tartaglia, M. E. Perkus, E. K. Norton, C. M. Bongermino, F. Lacoste, C. Duret, P. Desmettre, and E. Paoletti. 1992. Protection of mice and swine from pseudorabies virus conferred by vaccinia virus-based recombinants. J. Virol. 66:3424–3434. 33. Robbins, A. K., J. P. Ryan, M. E. Whealy, and L. W. Enquist. 1989. The gene encoding the gIII envelope protein of pseudorabies virus vaccine strain Bartha contains a mutation affecting protein localization. J. Virol. 63:250– 258. 34. Robbins, A. K., R. J. Watson, M. E. Whealy, W. W. Hays, and L. W. Enquist. 1986. Characterization of a pseudorabies virus glycoprotein gene with homology to herpes simplex virus type 1 and type 2 glycoprotein C. J. Virol. 58:339–347. 35. Ryan, J. P., M. E. Whealy, A. K. Robbins, and L. W. Enquist. 1987. Analysis of pseudorabies virus glycoprotein gIII localization and modification by using novel infectious viral mutants carrying unique EcoRI sites. J. Virol. 61:2962– 2972. 36. Saalmu ¨ller, A., and T. C. Mettenleiter. 1993. Rapid identification and quantitation of cells infected by recombinant herpesvirus (pseudorabies virus) using a fluorescence-based beta-galactosidase assay and flow cytometry. J. Virol. Methods 44:99–108. 37. Sachs, D. H., G. Leight, J. Cone, S. Schwarz, L. Stuart, and S. Rosenberg. 1976. Transplantation in miniature swine. I. Fixation of the major histocompatibility complex. Transplantation 22:559–567. 38. Schreurs, C., T. C. Mettenleiter, F. Zuckermann, N. Sugg, and T. Ben-Porat. 1988. Glycoprotein gIII of pseudorabies virus is multifunctional. J. Virol. 62:2251–2257. 39. Spillmann, D., and U. Lindahl. 1994. Glycosaminoglycan-protein interactions: a question of specificity. Curr. Opin. Struct. Biol. 4:677–682. 40. Strebel, K., E. Beck, K. Strohmaier, and H. Schaller. 1986. Characterization of foot-and-mouth disease virus gene products with antisera against bacterially synthesized fusion proteins. J. Virol. 57:983–991. 41. Trybala, E., T. Bergstrom, D. Spillmann, B. Svennerholm, S. J. Flynn, and P. Ryan. 1998. Interaction between pseudorabies virus and heparin/heparan sulfate. Pseudorabies virus mutants differ in their interaction with heparin/ heparan sulfate when altered for specific glycoprotein C heparin-binding domain. J. Biol. Chem. 273:5047–5052. 42. Trybala, E., T. Bergstrom, B. Svennerholm, S. Jeansson, J. C. Glorioso, and S. Olofsson. 1994. Localization of a functional site on herpes simplex virus type 1 glycoprotein C involved in binding to cell surface heparan sulphate. J. Gen. Virol. 75:743–752. 43. Vanderplasschen, A., M. Bublot, J. Dubuisson, P. P. Pastoret, and E. Thiry. 1993. Attachment of the gammaherpesvirus bovine herpesvirus 4 is mediated by the interaction of gp8 glycoprotein with heparinlike moieties on the cell surface. Virology 196:232–240. 44. Wathen, M. W., and L. M. Wathen. 1986. Characterization and mapping of a nonessential pseudorabies virus glycoprotein. J. Virol. 58:173–178. 45. Wittmann, G., and H.-J. Rziha. 1989. Aujeszky’s disease (Pseudorabies) in pigs, p. 230–325. In G. Wittmann (ed.), Herpesvirus disease of cattle, horses, and pigs. Kluwer Academic Publishers, Boston, Mass. 46. WuDunn, D., and P. G. Spear. 1989. Initial interaction of herpes simplex virus with cells is binding to heparan sulfate. J. Virol. 63:52–58. 47. Zhu, Z., M. D. Gershon, R. Ambron, C. Gabel, and A. A. Gershon. 1995. Infection of cells by varicella zoster virus: inhibition of viral entry by mannose 6-phosphate and heparin. Proc. Natl. Acad. Sci. USA 92:3546–3550. 48. Zsak, L., N. Sugg, and T. Ben-Porat. 1992. The different interactions of a gIII mutant of pseudorabies virus with several different cell types. J. Gen. Virol. 73:821–827. 49. Zsak, L., N. Sugg, T. Ben-Porat, A. K. Robbins, M. E. Whealy, and L. W. Enquist. 1991. The gIII glycoprotein of pseudorabies virus is involved in two distinct steps of virus attachment. J. Virol. 65:4317–4324. 50. Zuckermann, F., L. Zsak, L. Reilly, N. Sugg, and T. Ben-Porat. 1989. Early interactions of pseudorabies virus with host cells: functions of glycoprotein gIII. J. Virol. 63:3323–3329. 51. Zuckermann, F. A., L. Zsak, T. C. Mettenleiter, and T. Ben-Porat. 1990. Pseudorabies virus glycoprotein gIII is a major target antigen for murine and swine virus-specific cytotoxic T lymphocytes. J. Virol. 64:802–812.