Residues Critical for Duck Hepatitis B Virus Neutralization Are Involved in Host Cell Interaction

Abstract

To date, no detailed analysis of the neutralization properties of duck hepatitis B virus (DHBV) has been reported, and it is not clear whether any of the known neutralization epitopes correspond to the viral receptor binding site or to sequences involved in the cell entry pathway. We demonstrate here that antibodies directed against two overlapping peptides (amino acids 83 to 97 and 93 to 107), covering the sequences of most DHBV pre-S neutralizing epitopes, both inhibit virus binding to primary duck hepatocytes and neutralize virus infectivity. An extensive mutagenesis of the motif ⁸⁸WTP⁹⁰, which is the shortest sequence of the epitope recognized by the virus-neutralizing monoclonal antibody (MAb) 900 was performed in order to define the amino acids involved in these interactions. Single point mutations within this epitope affected neither virus replication nor infectivity but abolished virus neutralization by MAb 900 completely. Interestingly, mutants with two and three consecutive residue replacements (SIP and SIH) within this epitope retained replication competence but were no longer infectious. The loss of infectivity of SIH and SIP mutant particles was associated with significantly reduced binding to primary duck hepatocytes and could be rescued by trans complementation with wild-type pre-S protein. Taken together, these results indicate that each amino acid of the DHBV pre-S sequence ⁸⁸WTP⁹⁰ is critical for recognition by the neutralizing MAb 900 and that replacement of the first two or all three residues strongly reduces virus interaction with hepatocytes and abrogates infectivity. These data imply that the motif ⁸⁸WTP⁹⁰ contains key residues which are critical for interaction with both the neutralizing MAb and the host cell.

Duck hepatitis B virus (DHBV) provides a very useful model system with which the role of the hepadnavirus envelope proteins in infection, as well as in virus neutralization by antibodies, can be studied. Two major cocarboxy-terminal proteins have been identified within the DHBV envelope: a large (L, 36 kDa) and a small (S, 18 kDa) surface protein, both required for viral infectivity. The major DHBV neutralization epitopes map within the pre-S region of L protein, which is thought to mediate virus-hepatocyte interaction. During the last few years, a number of studies have shown that this pre-S domain takes part in the early steps of DHBV infection and that a 180-kDa carboxypeptidase D protein (gp180), a putative component of the cell receptor protein complex, is suspected of mediating this interaction (3, 13, 14, 16, 28, 30). A domain spanning about one-half of the DHBV pre-S region, mapped to amino acids (aa) 43 to 108 by Ishikawa et al. (13) and to aa 30 to 115 by Urban et al. (30), has been identified as being involved in binding to gp180 and in interaction with the cell surface receptor. In addition, an inner subdomain, the deletion of which abolishes virus interaction with gp180, has been defined within aa 87 to 102 by Tong et al. (28) and aa 85 to 115 by Breiner et al. (3). This region is a part of the highly conserved DHBV envelope protein domain ranging from aa 58 to 107, which contains several neutralization epitopes, i.e., type II (5, 32), type IV (5, 32), M900 (4), and SD20 (4, 17). However, no detailed analysis of the neutralization properties of DHBV has been reported, and it is not clear whether the same residues are involved in both virus neutralization and interaction with the host cell receptor. In this regard, neutralization of viral infectivity by antibodies is a complex and as yet poorly understood phenomenon. Recent studies on the functional domains of proteins suggest that neutralization sites and virus attachment sites are often distinct. Therefore, the binding of neutralizing antibodies does not necessarily directly block attachment to the host cell, as shown for a number of different viruses such as feline immunodeficiency virus (23), feline leukemia virus (26) and influenza virus (22). The neutralizing antibody may interfere with recognition of a host cell by a steric hindrance, by aggregation of virions, or by inducing a conformational change. There are only a few documented examples of residues within neutralization epitopes which are also involved in the attachment of the virus to its cellular receptor (for a review, see reference 10).

One experimental approach for defining residues involved in neutralization is to select mutant viruses which are able to escape neutralization by monoclonal antibodies (MAbs) and to identify the modified amino acids. Using this approach, we have previously reported the in vivo selection of DHBV pre-S neutralization escape variants by using a murine MAb (MAb 900) which recognizes an epitope mapped between aa 83 and 90 within the DHBV pre-S region (27). These variants harbored point mutations both at proline 90, within this epitope, and at a distance, at position 5, which almost completely abolished virus recognition by neutralizing MAb without affecting viral infectivity (27). However, the selection of neutralization-resistant variants is a strategy which cannot characterize every amino acid in an epitope, since any mutations which interfere with receptor recognition would lead to noninfectious virus and would thus not be detected in such a study.

In the present study, we have investigated the involvement of specific DHBV pre-S residues in both antibody-mediated neutralization and interaction with the host cell by using two complementary approaches. First, with a binding test and antisera to synthetic peptides covering the sequence of the clustered DHBV immunodominant epitopes (aa 83 to 107), we show that this domain is involved in virus neutralization and interaction with hepatocytes. Second, site-directed mutagenesis within the minimum sequence (⁸⁸WTP⁹⁰) recognized by the well-characterized DHBV-neutralizing MAb 900 shows that the substitution of a single amino acid is sufficient to abolish virus neutralization but not infectivity, whereas the replacement of two or three residues strongly reduces virus-hepatocyte interaction and abolishes infectivity.

MATERIALS AND METHODS

Antibodies.

The rabbit polyclonal antiserum (DPS) used in this study was raised against the first 131 aa of the recombinant DHBV L envelope protein (17). The MAb 900, generated with the same DHBV pre-S domain polypeptide, recognizes a neutralization epitope which was previously characterized and mapped between amino acids I⁸³ and Q⁹⁰ (3). The MAb 7C-12, kindly provided by J. Pugh, is directed against the aa 266 to 275 on the S domain of the DHBV surface proteins (23, 24).

Generation of pre-S mutants.

Mutagenesis was performed by two rounds of PCR as previously described (1) by using the wild-type DHBV genome sequenced by Mandart et al. (21). The mutations introduced led to the replacement of one, two, or three residues within the epitope ⁸⁸WTP⁹⁰ (Fig. 1), which is the minimum sequence recognized by MAb 900 in PEPSCAN analysis (4). All mutations (except P90T) were designed to avoid introduction of modifications in the overlapping polymerase amino acid sequence. To analyze the impact of mutations on viral replication and infectivity, PCR-mutated fragments were subcloned into an infectious pCMV-DHBV genome (7) by using the appropriate restriction enzyme sites. All mutations were confirmed by sequencing with the Sequenase Version 2.0 DNA sequencing Kit (USB; Amersham). For envelope protein complementation, a pCI–pre-S/S vector was constructed in which the expression of the entire pre-S/S sequence is driven by the early cytomegalovirus promoter. The DHBV DNA fragment (nucleotides 801 to 1785) encoding the pre-S/S gene was PCR amplified and cloned into the polylinker of the pCI expression vector (Promega, Lyon, France).

FIG. 1.

Schematic representation of amino acid substitutions in the pre-S domain of the DHBV L envelope protein. The mutant names refer to the location and the nature of the amino acid changes within the ⁸⁸WTP⁹⁰ neutralization epitope recognized by MAb 900.

LMH cell transfection and detection of mutant virus.

The avian hepatoma LMH cell line was transfected by the calcium phosphate method with 20 μg of DNA plasmid per 100-mm dish as described previously (27). For cotransfection assays the vector pCI–pre-S/S, which expresses DHBV L protein, was used as a helper plasmid to provide wild-type L protein. A total of 20 μg of mutated pCMV-DHBV and pCI–pre-S/S plasmids (4:1 ratio) was used. Culture media were collected daily between days 5 and 7 posttransfection, clarified by centrifugation at 1,000 × g for 10 min, and concentrated by polyethylene glycol (PEG) precipitation or ultracentrifugation on 10 to 20% sucrose gradients as previously described (27). The presence of mutations was confirmed by PCR and direct sequencing of DHBV DNA extracted from DNase I-treated and immunoprecipitated viral particles as described previously (27).

Sedimentation analysis of viral particles in CsCl gradient.

Supernatants of LMH cells transfected with different plasmids were first clarified and then ultracentrifuged for 2 h at 35,000 rpm (SW41 rotor) at 4°C. The concentrated viral particles were then placed on the top of a linear 1.14- to 1.36-g/ml CsCl gradient and centrifuged to equilibrium in an SW41 rotor at 35,000 rpm for 20 h at 4°C as described previously (2). The fractions containing enveloped viral particles and nonenveloped cores were identified by analysis of DHBV DNA by dot blotting and of pre-S protein by immunoblotting.

Infection of PDH cultures and DHBV neutralization.

Primary duck hepatocytes (PDHs) were isolated from 3-week-old Pekin ducklings by in situ double-step collagenase perfusion of livers as previously described (2) and seeded on six-well plates (2.5 × 10⁵ cells/well). Cells were infected 1 day after seeding with 100 μl of virus samples concentrated from transfected LMH cells and containing about 5 × 10¹⁰ viral genomes (VGE)/ml. For the neutralization tests, PDHs were infected with 2 × 10⁸ VGE/well of wild-type virus or DHBV pre-S mutants, preincubated overnight either with L15 culture medium (controls) or with murine or rabbit antibody as described earlier (18). The release of viral particles into cell culture media was monitored for 8 days by quantitative dot blot hybridization. At the end of the culture, PDHs were harvested, and intracellular viral DNA was analyzed by Southern blot as previously described (18). All infection and neutralization assays were performed in duplicate and reproduced twice by using two series of inocula prepared independently.

In vivo infectivity of DHBV mutants.

Three-day-old DHBV-free Pekin ducklings were intravenously inoculated with quantified wild-type or mutant DHBV inocula concentrated from transfected LMH cells (2 × 10⁸ VGE/duck). Viremia was followed every day during the first week and then every 2 days for the two following weeks as described previously (8, 27).

Immunoblotting.

The expression of DHBV L protein was monitored in transfected cells or PDHs by Western blotting with rabbit polyclonal anti-DHBV pre-S serum (diluted 1:1,000), anti-peroxidase conjugate (diluted 1:1,000; Biosys), and the ECL chemiluminescence detection kit (Amersham) as reported previously (17, 27).

Analysis of DNA from infected primary hepatocytes and duck livers.

Primary hepatocytes were collected in lysis buffer containing 0.5% sodium dodecyl sulfate (SDS) at day 10 postinfection and assayed for viral replicative intermediates. After proteinase K digestion (100 μg/ml, 4 h at 37°C), the total DNA was extracted as previously described (2, 27). The same procedure was applied to homogenized duck liver samples. Equal amounts of DNA from each sample were analyzed by electrophoresis through a 1% agarose gel, transferred to a nylon membrane, and hybridized with a ³²P-labelled full-length genomic DHBV DNA probe as previously described (2, 11).

Peptide synthesis and preparation of anti-peptide antibodies.

Two overlapping peptides PS1 (aa 83 to 97; IPQPQWTPEEDQKAR) and PS2 (aa 93 to 107; DQKAREAFRRYQEER) were synthesized in solid phase by using 9-fluoroenylmethoxycarbonyl technology and chemically coupled to a carrier protein, keyhole limpet hemocyanin, by the method of Liu (20). Two rabbits were immunized separately with 200 μg of each peptide followed by three boosts at days 14, 28, and 56 by using standard protocols (Neosystem, Strasbourg, France). The immune response to the peptide was evaluated by determining the endpoint titer of antibodies that reacted with recombinant DHBV pre-S protein in an enzyme-linked immunosorbent assay (ELISA) as described previously (4).

Assay to detect DHBV particles bound to PDHs.

The assay we have used is essentially the one described by Pugh et al. (25). Briefly, PDHs were prepared as described above and plated at high density in six-well plates. Two or three days after plating, cells were washed with phosphate-buffered saline (PBS) and incubated (1 h 30 min at room temperature) with DHBV diluted in Opti-MEM medium (Life Technologies) or preincubated overnight with different rabbit anti-preS sera (DPS, anti-PS1, and anti-PS2) diluted to a similar anti-DHBV pre-S titer as assessed by ELISA. After three washes, anti-S MAb 7C-12 (diluted 1/100 in PBS–5% normal duck serum [NDS]) was added to the cells and incubated under the same conditions. Cells were washed again and then incubated for 30 min with ¹²⁵I-labeled goat anti-mouse immunoglobulin G (IgG) (Biosys) diluted in PBS–5% NDS. After three washes, cells were lysed with PBS–1% SDS, and the cell-associated radioactivity was determined in a gamma counter.

RESULTS

Antibodies to DHBV neutralization epitopes inhibit DHBV binding to PDHs.

First, we determined whether virus binding to PDHs is affected by anti-DHBV pre-S antibodies against recombinant pre-S protein or synthetic peptides. The binding assay was performed by measuring the amount of PDH-associated viral particles after incubation with DHBV-positive duck serum, detected by using the MAb 7C-12, which is known to specifically detect DHBV S protein (25) and I¹²⁵-labeled goat anti-mouse IgG. The results found with PDHs incubated with DHBV-positive serum showed that DHBV binds to PDH, whereas no significant binding was observed when hepatocytes were incubated with DHBV-negative (mock) serum (Fig. 2). In addition, the cell-associated radioactivity was proportional to the amounts of viral particles used (results not shown). Taken together, these results indicate that this assay allows the detection of PDH-associated DHBV particles.

FIG. 2.

Binding of DHBV to PDHs is inhibited by antibodies to neutralization epitopes. PDH monolayers were incubated with concentrated DHBV-positive duck serum (DHBV) or uninfected duck serum (Mock). To test the capacity of DHBV pre-S antisera to inhibit binding, virus was preincubated with sera from a nonimmunized rabbit, DHBV pre-S polypeptide (aa 1 to 131) rabbit antiserum (DPS), or ⁸³I-R⁹⁷ (PS1), or/and ⁹³D-R¹⁰⁷ (PS2) peptide antisera. All antisera were diluted to similar anti-DHBV pre-S titers, as assessed by ELISA. Bound virus was detected with anti-S MAb 7C-12 and ¹²⁵I-labeled goat anti-mouse IgG as described in Materials and Methods. The counts obtained for control DHBV are represented as 100%.

Preincubation of viral particles with nonimmunized rabbit serum has no effect on binding, whereas preincubation with the recombinant DHBV pre-S polypeptide (aa 1 to 131) antiserum (DPS) reduced binding by 75% (Fig. 2). Both of the two antipeptide sera, which recognize part of the pre-S region thought to be involved in recognition of the hepatocyte receptor, namely, aa 83 to 97 (PS1) and aa 93 to 107 (PS2), reduced binding by about 65%. Simultaneous addition of both antipeptide sera reduced the binding of viral particles even further to levels similar to those observed with the recombinant DHBV pre-S polypeptide antiserum directed against almost the entire pre-S region. These data suggest that either the major hepatocyte binding domain of the pre-S protein or a critical part of it is located between aa 83 and 107.

Next, we tested whether the polyclonal antibodies against the recombinant pre-S protein or the two peptides are neutralizing. The in vitro neutralizing test showed that 100% neutralization was achieved by using DPS, whereas serum from a nonimmunized rabbit was not neutralizing (Fig. 3). Both PS1 (aa 83 to 97) and PS2 (93 to 107) antisera are highly neutralizing, since a decrease by 89 and 94%, respectively, of released virions was observed in PDHs infected with DHBV preincubated with these sera (Fig. 3). Taken together, these results indicate that antibodies elicited to the aa 83 to 107 domain within DHBV pre-S both inhibit virus binding to PDHs and neutralize DHBV infectivity.

FIG. 3.

DHBV infectivity is neutralized in vitro by PS1 and PS2 peptide antisera. PDHs were infected with DHBV-positive duck serum (2 × 10⁸ VGE/well), preincubated with L15 medium (infection controls), unimmunized rabbit serum, DPS, or PS1 (⁸³I-R⁹⁷), PS2 (⁹³D-R¹⁰⁷), or PS1 and PS2 rabbit antisera. The release of viral particles into the supernatants was quantified by dot blot hybridization as described in Materials and Methods. The means of duplicate determinations are presented.

DHBV pre-S mutants are replication competent but differ in their infectivity.

To define which, if any, of the amino acids within this domain plays a role in virus infectivity, we chose to introduce into the pCMV-DHBV plasmid 1-, 2-, or 3-amino-acid replacements in the ⁸⁸WTP⁹⁰ motif recognized by the highly neutralizing MAb 900 (Fig. 1). From LMH cells transfected with all eight mutated plasmids, designated W88L, W88S, T89I, P90H, P90T, WIH, SIH, and SIP, and the wild type, a progressive release of viral DNA into the cell culture medium was observed (Fig. 4A). That this represents secreted viral particles was further confirmed by the detection of similar amounts of the 35- to 36-kDa DHBV pre-S/S proteins in the concentrated culture medium of cells transfected either with the mutants or with the wild-type plasmid (Fig. 4B). To ensure that all of the mutations were maintained and that all of the mutants are replication competent and really secreted enveloped DNA-containing viral particles, we have immunoprecipitated viral particles from supernatants of transfected cells after DNase I treatment and confirmed, by sequencing of the entire DHBV pre-S region (nucleotides 628 to 1485), that the DHBV DNAs had the same sequence as the respective constructions used for the transfection (data not shown).

FIG. 4.

DHBV pre-S mutations do not affect virus replication capacity. pCMV-DHBV DNA plasmids containing each mutation presented in Fig. 1 were used to transfect the LMH hepatoma cell line. (A) Virus release into the transfected cell supernatants was monitored by dot blot hybridization as described in Materials and Methods. Lane 1, wild type; lane 2, W88L; lane 3, W88S; lane 4, T89I; lane 5, P90H; lane 6, P90T; lane 7, WIH; lane 8, SIP; lane 9, SIH. (B) Viral particles from the culture supernatants were concentrated and analyzed by immunoblotting. Proteins were probed with rabbit anti-DHBV pre-S antibody (DPS). Lane 1, wild type; lane 2, W88L; lane 3, W88S; lane 4, T89I; lane 5, P90H; lane 6, P90T; lane 7, WIH; lane 8, SIP; lane 9, SIH; lane 10, negative control. For comparison, circulating viral particles concentrated from DHBV-positive duck serum were loaded into lane 11.

To test the infectivity of the secreted viral particles, transfected LMH cell culture media were concentrated with PEG and used to infect the PDHs. It was important to ascertain that the infection with different mutant or wild-type viruses was normalized to the same number of enveloped particles. To this end, we have separated the completely enveloped virions from unenveloped cores by CsCl equilibrium ultracentrifugation of LMH cell supernatants. Quantification of the DHBV DNA in all fractions by dot blot hybridization and pre-S protein by immunoblotting (data not shown) indicated that approximately 60% of the viral DNA sedimented at a density of 1.18 g/ml corresponding to the density of enveloped particles and 40% of the viral DNA sedimented as unenveloped cores at a density of 1.3 g/ml, a finding consistent with previous observations (2). The same respective proportions were found to be similar for wild-type and different mutant virus preparations. The single-amino-acid mutants W88L, W88S, P90T, and P90H and the double-mutant WIH were infectious in vitro, and the virus titers were similar to those obtained for the wild-type DHBV. The mutant T89I was also infectious, but the DNA levels in the cell culture medium were about fourfold lower (Table 1). However, the SIP and SIH mutants, which have 2- and 3-amino-acid changes, respectively, failed to infect hepatocyte cultures, as shown by the absence of detectable viral proteins and DNA signals in cell culture lysates and media (Table 1 [see also Fig. 7] and data not shown). Since the PDH infection assay has a limited sensitivity and may not reflect all aspects of an in vivo infection, we also examined the infectivity of these two mutants and of all the other mutants in ducklings. Nine groups of 3-day-old ducklings were infected with concentrated viral particles from transfected LMH cells. As expected from the in vitro results, all single mutants and the WIH double mutant were infectious in vivo (Table 1). In contrast, no DHBV DNA was detected in the sera of ducklings infected with the SIP and SIH mutants (Table 1). Moreover, no intrahepatic DNA was detected in the corresponding livers of these ducklings by Southern blot hybridization (Fig. 5A). This technique allows the detection of viral DNA in as little as 0.02 μg of total liver DNA (Fig. 5B). The direct sequencing of the DHBV pre-S DNA fragments amplified from viremic sera indicated that all mutations were maintained after one in vivo passage (data not shown). Altogether, these results demonstrate that all single-amino-acid mutants tested and one double mutant (WIH) with the last 2 aa changed in the WTP motif do not or only marginally affect viral infectivity in vitro or in vivo. In contrast, changing of the first 2 (SIP mutant) or all 3 aa (SIH mutant) of the WTP motif abrogates DHBV infectivity, which may imply a function of these amino acids in receptor recognition or downstream infection events.

TABLE 1.

Replication capacity, infectivity, and in vivo neutralization of DHBV pre-S mutants

Transfected DHBV DNA	DHBV detection in transfected LMH cellsa	In vitro infectivityb (virus titer [106 VGE/ml] of cell medium)	In vivo infectivityc (no. of DHBV-positive ducks/total no. of inoculated ducks)	In vitro neutralizationd
				DPS	MAb 900
Wild type	+	54	3/3	+	+
W88L	+	52	3/3	+	−
W88S	+	56	3/3	+	−
T89I	+	14	3/3	+	−
P90T	+	66	3/3	+	−
P90H	+	56	3/3	+	−
WIH	+	42	3/3	+	−
SIP	+	0	0/4	NTe	NT
SIH	+	0	0/3	NT	NT

^aFive days after LMH cell transfection with pCMV-DHBV plasmids, DHBV DNA and proteins were detected in cell lysates and supernatants as described in Materials and Methods. 

^bFollowing PDH infection with supernatants from transfected LMH cells, released virions were quantified in culture medium as described in Materials and Methods. 

^cNumber of ducks positive for DHBV DNA in serum and liver/total number of inoculated ducks. 

^dFor the in vitro neutralization assay, PDHs were infected with virus preincubated with rabbit anti-pre-S serum (DPS) or MAb 900. Significant neutralization (+) is taken to be a 50% or greater reduction in the supernatant viral DNA level. −, Absence of neutralization. 

^eNT, not tested for neutralization since found to be noninfectious. 

FIG. 7.

Wild-type (WT) DHBV L protein trans-complements SIP and SIH mutant infectivity. LMH cells were transfected with pCMV-DHBV plasmids bearing the double (SIP) or triple (SIH) substitutions in the presence or absence of pCI–pre-S/S vector, which expresses wild-type L protein. PDH cultures were infected with viral particles produced by transfected LMH cells. (A) The viral release into cell supernatants was monitored from days 3 to 8 postinfection by dot blot hybridization. (B) Cells were harvested 8 days postinfection, and the L protein was detected in the cell lysate by immunoblotting with rabbit anti-DHBV pre-S antibody. For comparison, circulating viral particles concentrated from DHBV-positive duck serum were loaded in the right lane. The arrow shows the position of the 36-kDa DHBV L protein.

FIG. 5.

SIP and SIH mutant viruses are not infectious in vivo. Viral particles produced by LMH cells transfected with pCMV-DHBV plasmids bearing the double (SIP) or triple (SIH) substitutions were used to infect ducklings. (A) Intrahepatic DHBV DNA replicative forms were analyzed by Southern blotting 3 weeks after inoculation in the liver from one duckling inoculated with wild-type DHBV (control) and two ducklings inoculated with either SIP or SIH mutant viruses. (B) Sensitivity of intrahepatic DHBV DNA detection in successive dilutions of the total liver DNA from a DHBV-positive liver. Arrows indicate the relaxed circular (RC), linear (L), and single-stranded (SS) DHBV DNA forms.

Neutralization of DHBV by MAb 900 is abolished by the WTP mutations.

To investigate the impact of DHBV neutralization by MAb 900 on those mutants that retain infectivity, we performed an in vitro neutralization assay. As illustrated in Fig. 6 and summarized in Table 1, all mutants were fully neutralized in vitro, when incubated with the highly neutralizing DPS polyclonal antibody but none were neutralized by MAb 900. From these results we conclude that each of the residues of the WTP motif is essential for the neutralization of virus infectivity by MAb 900. The neutralization of DHBV mutants by DPS is in agreement with our previous studies which showed that this polyclonal antibody recognizes, in addition to the MAb 900 epitope, other neutralization epitopes within the DHBV pre-S protein (27).

FIG. 6.

Mutations within the WTP epitope abolish in vitro neutralization of DHBV by MAb 900. PDH cultures were infected with wild-type DHBV or P90H mutant particles preincubated either with L15 culture medium (controls), with MAb 900 ascites fluid, or with DPS polyclonal antibody as described in Materials and Methods. Viral release into the culture medium was monitored from days 5 to 10 postinfection by dot blot hybridization.

Wild-type DHBV L protein trans-complements mutant infectivity.

The two mutants SIP and SIH were found to be noninfectious. We examined whether their infectivity could be restored by providing the wild-type DHBV L protein in trans by cotransfecting LMH cells with SIP- or SIH-encoding plasmids and a wild-type L protein expression vector. The virus stocks obtained were then used to infect PDHs as described above. Cotransfection of LMH cells with the SIH or SIP mutant and pCI–pre-S/S plasmid resulted in the release of viral particles which infected PDH, as judged by detection of both DHBV DNA in PDH media (Fig. 7A) and L proteins within cell lysates (Fig. 7B). This indicates that the infectivity of both mutants can be rescued by the wild-type envelope protein and implies that the corresponding mutations affect the pre-S protein domain function only.

Binding of noninfectious mutant virus to PDHs is affected.

The data presented above indicate that the lack of infectivity of the SIH and SIP mutants is due to the mutant DHBV L envelope protein. It is not clear whether this interferes negatively with the binding of the mutant viral particles to the cellular receptor or with a subsequent step in infection. In order to examine the first possibility, we performed a binding assay with SIP and SIH mutant viral particles. In PDHs treated with equivalent amounts of mutant or wild-type particles, 60% fewer mutant pre-S particles were bound (Fig. 8), indicating that the amino acid changes significantly reduce but do not abolish the binding of the mutant particles to PDHs.

FIG. 8.

Binding of SIP and SIH mutant virus to PDH is reduced. PDH monolayers were incubated with wild type (WT), SIP, or SIH viral particles produced by transfected LMH cells as described in Materials and Methods. The results are expressed as the percentages of counts obtained with wild-type DHBV particles, after substraction of the values obtained with concentrated medium from mock-transfected LMH cells.

DISCUSSION

In the present study, we have demonstrated that residues 88 to 90 within the DHBV L envelope protein are critical for virus neutralization and are involved in the interaction with the host cell. This is of particular interest, since the roles of hepadnavirus envelope protein in the earliest events of infection (i.e., virus interaction with the host cell membrane, penetration, and disassembly) and virus neutralization are still poorly understood. Our data are consistent with and extend previous data and speculation about the type of pre-S sequences involved in receptor interaction by pointing to the involvement of a short-amino-acid motif in the pre-S region. A large portion (aa 30 to 115) of DHBV L protein has recently been shown to interact with carboxypeptidase gp180, which is thought to be part of a cellular receptor for avian hepadnaviruses (3, 30). Within this pre-S domain, a recombinant polypeptide covering the residues 87 to 102 has been shown to bind specifically to a gp180 in an in vitro binding test (28). Since aa 87 to 102 harbor three major neutralization epitopes (4, 5, 6, 17, 32), it has been speculated that the sequences comprising the neutralization epitopes are also involved in the interaction with hepatocyte (28). However, the nature of viral epitopes involved in neutralization has been analyzed in detail for only some viruses, and it has been shown that they may be either identical to or distinct from the receptor binding site (10, 22, 23, 26). In this regard, our data are of particular interest since they provide the first experimental support for the hypothesis that some residues involved in DHBV neutralization are also involved in the early stages of interaction with the host cell.

We found that rabbit antibodies elicited by two overlapping peptides (I⁸³-R⁹⁷) and (D⁹³-R¹⁰⁷) covering the cluster of three DHBV pre-S neutralizing epitopes (aa 83 to 107) were able to both inhibit virus binding to primary duck hepatocytes and neutralize virus infectivity. To define these interactions at the amino acid level, we have performed an extensive mutagenesis of the motif ⁸⁸WTP⁹⁰, which is the shortest sequence of the epitope recognized by the previously characterized virus-neutralizing MAb 900. Single point mutations as well as the double substitution (mutant WIH) within this epitope affected neither virus replication nor infectivity but completely abolished virus neutralization by MAb 900. The single-mutant T89I exhibited a slight but reproducible decrease of infectivity in vitro. This residue forms part of an S/T-P motif that is a potential protein phosphorylation site (12). The T89I mutation could lead to a modification of the phosphorylation state of the DHBV L envelope protein. However, our laboratory has recently shown that the T89I mutation does not affect phosphorylation of the DHBV L protein (2). It is therefore probable that the reduction of viral secretion observed after infection of primary hepatocyte cultures with the mutant T89I is not due to an impact on phosphorylation.

By contrast, analysis of the SIP and SIH mutants has revealed that these double and triple amino acid substitutions abrogate DHBV infectivity in vitro and in vivo without affecting their replication capacity. The wild-type-like replication capacity of these two mutant plasmids in transfected LMH cells is consistent with the fact that none of the mutations introduced altered the P-protein sequence. Moreover, the wild-type-like efficiency of viral particle secretion of these mutants indicates that no other transcription regulatory sequences are affected by these mutations. Therefore, our results strongly suggest that these mutations affect early stages of virus infection only. This suggestion is corroborated by complementation assays that showed that the DHBV L envelope protein provided in trans is able to restore the infectivity of the mutants.

The question remained whether the mutations affect the initial step of specific binding to hepatocyte membranes or subsequent steps in viral penetration into the duck hepatocytes. Recent data provide evidence that gp180, a Golgi-resident protein, is involved in the first step of avian hepadnavirus uptake (3). This initial step involves the interaction of gp180 with a defined subdomain of DHBV pre-S which has been mapped between aa 85 and 115 by using recombinant DHBV pre-S polypeptides in an in vitro infection competition assay (3, 30). Interestingly, within this pre-S region, a short sequence spanning residues 85 to 96 has been recently identified as an absolutely essential element for receptor interaction (31). Using a completely different approach, i.e., site-directed mutagenesis, we have demonstrated that mutants which harbor the replacement of residues 88 and 89 within this short amino acid stretch completely lose their infectivity in PDH cultures and in ducklings. These observations are of particular interest since they provide the first evidence that the portion of the DHBV pre-S protein which is prerequisite for gp180 recognition contains the residues which are essential for in vivo initiation of DHBV infection. However, the direct interaction of the noninfectious mutants SIP and SIH with gp180 was not investigated in the present study. The binding assay we have developed showed that these mutants bind to PDHs with an efficiency decreased by 60% compared to that of wild-type particles. In spite of this significant decrease in mutant binding, it is unclear why the 40% of residual binding does not allow viral infection. Considering that gp180 is a Golgi-resident protein which cycles to and from plasma membrane (3), it is difficult to evaluate the number of binding sites at the hepatocyte surface in our binding test. Further study of the interaction of SIP and SIH mutants with recombinant gp180 protein will provide a definite answer as to what extent, if at all, the lack of infectivity of the SIH and SIP mutants is due to their inability to recognize and/or strongly attach to the hepatocyte receptor via interaction with gp180. In addition, the motif ⁸⁸WT⁸⁹ may be important for interaction with other, as-yet-unidentified host-specific factors (coreceptors) which are thought to be required for full viral uptake and infection (3, 30).

Taken together, our data indicate that the neutralization epitope ⁸⁸WTP⁹⁰ is also involved in the cell entry pathway. The change of only one amino acid within this motif was sufficient to affect immunoreactivity with MAb 900, while the replacement of at least the first two residues was required for strong reduction of virus interaction with hepatocytes and a loss of infectivity. From these results we conclude that the motif ⁸⁸WTP⁹⁰ contains key residues which are important for interaction with both the host cell and the neutralizing MAb, although this does not mean that other portions of the DHBV L protein are not involved in the recognition of the cellular receptor or in virus neutralization. In this regard, it has been recently demonstrated that recombinant polypeptides covering a large portion (aa 30 to 115) of DHBV or heron hepatitis B virus (HHBV) pre-S region compete equally well for interaction with gp180, despite a 50% difference in their amino acid sequences, suggesting that the three-dimensional structure of DHBV pre-S is important for interaction with the cellular receptor (30). The three-dimensional structure of DHBV pre-S seems also important for virus neutralization, since we have previously shown that P⁹⁰ within the 83–90 epitope studied here and a distantly located residue, P⁵, are both involved in DHBV neutralization (27). For HBV it has been demonstrated, by using the random peptide library approach, that an anti-pre-S1 MAb which prevents virus binding to liver membranes is also able to recognize distantly located pre-S2 epitopes, suggesting that the hepatocyte-binding domain of the surface proteins is conformationally dependent (9). The use of combinatorial libraries to define mimotopes recognized by anti-DHBV pre-S MAbs will be of interest in defining the tertiary structures involved in DHBV neutralization and/or the cell entry pathway.