A Metastable Form of the Large Envelope Protein of Duck Hepatitis B Virus: Low-pH Release Results in a Transition to a Hydrophobic, Potentially Fusogenic Conformation

Western blot analysis.

Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12.5% acrylamide) and transferred to a nitrocellulose membrane (Schleicher and Schuell) using a Trans-Blot SD semidry transfer cell (Bio-Rad). Membranes were blocked for 1 h with 3% skim milk in TBST (100 mM Tris-HCl [pH 7.4], 150 mM NaCl, 0.3% Tween 20). L and L protein products were detected with polyclonal or monoclonal antiserum to pre-S or S domains as indicated by amino acid numbers in the diagram in Fig. 4A. Anti-S domain antiserum detects both L and S proteins due to their shared carboxy-terminal S domains. Membranes were probed with specific antiserum for 1 h in 1% skim milk–TBST, washed with TBST, and probed with horseradish peroxidase-conjugated species-specific antibodies (Amersham) in 1% skim milk–TBST. After a final wash in TBST (three times for 10 min each), protein bands were visualized by enhanced chemiluminescence (ECL; Amersham).

Treatment of particles with deforming agents and digestion with trypsin.

SVPs were prepared from DHBV-positive serum by a sucrose step gradient as described previously (11). SVPs were stored in small aliquots at −20°C for single use only.

(i) Low-pH treatment.

Ten microliters of SVP preparation (equivalent to approximately 500 ng of L protein and 2 μg of S protein), serum virus, or serum virus pelleted through a 20% sucrose cushion (0.2 ml) was diluted with an equal volume of 50 mM Na acetate–150 mM NaCl adjusted so as to achieve a final pH of either 4 or 5 as required. Particles were incubated at pH 4 for 15 min or pH 5 for 16 h at either 4 or 37°C and at completion neutralized with 1.5 M Tris-HCl (pH 8.8)–150 mM NaCl.

(ii) Heat or DTT treatment.

Particles were incubated for 15 min at 50, 55, or 60°C and then briefly cooled on ice prior to trypsin digestion. Particles for direct trypsin analysis were incubated at 37°C for the prescribed times at a final concentration of 10 mM DTT. Serum virus, used for infection of primary duck hepatocyte (PDH) cultures, was treated with 5 mM DTT.

(iii) Trypsin digestion.

Trypsin efficiently cleaves the pre-S domain of particles at a cluster of lysine residues at amino acid (aa) 154 and 159 to 161 to generate a pre-S product of approximately 20 kDa. Another major trypsin site exists at lysines 204 and 206, between the first two transmembrane domains of the S domain, but is accessible to cleavage only in the presence of detergent (Fig. 1D; see also Fig. 4A). Treated and untreated particles were digested with trypsin (Sigma) (20 μg/ml) for 1 h at 37°C. Trypsin digestion was stopped by addition of aprotinin (final concentration, 2 μg/ml) and boiling in Laemmli buffer.

Infection of PDHs.

PDH cultures were prepared as described previously (20). PDHs seeded onto 12-well multiplates were infected for 4 h with a serum virus pool pretreated with low-pH buffers or DTT as described above. Samples were neutralized with an equal volume of 200 mM Tris-HCl (pH 8.8) prior to infection, and cells were harvested at day 9 postinfection for intracellular viral DNA analysis by dot blot hybridization as described previously (23).

Sucrose gradient separation of virions and SVPs.

Eight milliliters of DHBV-positive serum was pelleted for 3 h at 38,000 rpm in an SW40 rotor (Beckman) through 3 ml of 20% sucrose onto a 2-ml 70% sucrose cushion. Fractions were analyzed by SDS-PAGE and Coomassie blue staining, and 0.2 ml of the S protein-containing fraction was treated with deforming agents as described above. Samples were diluted in phosphate-buffered saline (PBS) to 4 ml and centrifuged for 4 h at 38,000 rpm through a 20 to 70% sucrose step gradient (11). Fractions were collected from the bottom of the gradient and analyzed for envelope proteins and viral DNA by SDS-PAGE and silver staining and DNA dot blot hybridization, respectively.

Liposome membrane binding.

Liposomes were obtained as a preliposome formulation (Sigma) composed of l-α-phosphatidylcholine β-oleoyl-V-palmitoyl (POPC), l-α-phosphatidyl-dl-glycerol,dioleoyl (PGDO), and cholesterol. The preparation was reconstituted in 1 ml of filtered PBS according to product instructions. Fifteen microliters of the preparation was mixed with 10 μl of SVP to which an equal volume of deforming agent was added. The reaction mixture was incubated for 15 min at 37°C and then resuspended with 67% (wt/vol) sucrose in PBS to a final volume of 200 μl and overlaid with a step gradient of 2 ml of 25% sucrose and 2 ml of 10% sucrose. Following centrifugation for 5 h at 45,000 rpm in an SW60 rotor (Beckman), the liposomes floated to the top of the gradient from where 0.25-ml fractions were collected. Fractions were methanol precipitated, and SVPs were detected by SDS-PAGE and Western blotting with an anti-pre-S antiserum.

A trypsin cleavage site is exposed following treatment of DHBV particles with low pH, heat, or DTT.

In this study, we have used SVPs (11), which consist of the same ratio and topology of surface proteins as capsid-containing particles and which are produced in a 1,000-fold excess, making them a usable source of particles for study. Gradient-purified SVPs were examined for any conformational or topological changes in the envelope proteins caused by treatment at low pH or heat, under conditions of limited trypsin digestion. In this assay, a proportion of the full-length L (P36) protein remains resistant to protease attack in untreated intact control particles (Fig. 2A, lane 2) even after extended periods of digestion, which, in keeping with previous observations (25; E. V. L. Grgacic, unpublished data), represents the internal topology. Another topological form of L, the intermediate-topology form (Fig. 1C), is susceptible to trypsin cleavage at a cluster of lysine residues at aa 154 and 159 to 161, near the C terminus of the pre-S domain, generating a pre-S product of approximately 20 kDa (Fig. 2A, lane 2) (9, 25). However, a larger pre-S-reactive trypsin product of approximately 25 kDa was observed following treatment at low pH at 37°C (Fig. 2A, lane 4, and 2B, lane 2) or at temperatures above 50°C at neutral pH (Fig. 2C, lanes 2 and 3). Generation of the 25-kDa trypsin product occurred following a 15-min treatment at pH 4 (Fig. 2A, lane 4), while at pH 5 it did not occur after 15 min but required more prolonged periods (16 h, Fig. 2B, compare lanes 1 and 2). At pH 7, however, a 16-h incubation at 37°C did not result in an altered trypsin product (Fig. 2B, lane 3). Treatment at 55°C resulted in a mixture of the 20- and the 25-kDa cleavage products (Fig. 2C, lane 2), while at 60°C the 25-kDa product was the predominant product detected (lane 3). Treatment of particles with 4 M urea, an agent shown to result in changes mimicking those created by low pH and heat in influenza virus (5), resulted in total digestion of the DHBV envelope proteins (data not shown). Furthermore, binding to the primary attachment receptor, gp180, itself or together with low-pH or reducing conditions did not elicit a similar conformational change in L or any change in the trypsin cleavage pattern (data not shown).

Previously, we have shown that the cysteine-containing region between TM1 and TM2 forms a membrane-embedded loop (8), which may contribute to a metastable L structure and thus be susceptible to reducing agents. We therefore examined the effect that the reducing agent DTT would have on the trypsin cleavage pattern of L. Particles were incubated with 10 mM DTT for various times at 37°C as indicated (Fig. 3A) and then subjected to trypsin digestion as described above. DTT treatment for 15 min did not result in an altered trypsin cleavage pattern (Fig. 3A, lane 1). However, partial conversion to the 25-kDa cleavage product occurred after 4 h, and at 16 h, the 25-kDa product predominated (Fig. 3A, lanes 3 and 4). To assess whether DTT would also enhance the rate of this transition at pH 5, which required some 16 h, particles were incubated at pH 5 with 10 mM DTT at 37°C. DTT triggered a rapid change to the 25-kDa trypsin product when particles were incubated at pH 5 for 15 min (Fig. 3B, lane 1) or with the mildly acidic pHs of 6 and 6.5 (data not shown).

The 25-kDa pre-S trypsin product is generated by cleavage in the membrane-traversing loop between TM1 and TM2.

The origin of the 25-kDa trypsin product was mapped by sequential Western blotting using antisera recognizing specific epitopes in the L amino acid sequence (Fig. 4A). As shown in Fig. 4B, the 25-kDa product was detected using the antiserum to aa 10 to 29 and aa 94 to 107 in pre-S but did not react with an antiserum recognizing aa 207 to 226 in the S domain. This latter epitope lies immediately downstream of two lysine residues (aa 204 and 206) which are not surface exposed in any known topology of L (Fig. 1) and which require the presence of detergent for cleavage, generating a C-terminal, 12.9-kDa S product (8, 9). Accordingly, the 25-kDa product spans an extended region comprising all of pre-S and extending into the S domain, including the first transmembrane domain, to trypsin sites upstream of TM2. Cleavage may occur at a pair of either lysine or arginine residues within this domain (Fig. 4A), but the size of the pre-S product is more consistent with cleavage at the lysine residues. Following generation of this 25-kDa pre-S product, only a small amount of the expected 12.9-kDa carboxy-terminal fragment is detectable by ECL upon longer exposure (data not shown), suggesting that further cleavage at two potential trypsin sites (K267 and K289) in the external loop downstream of TM2 may be causing the complete degradation of this fragment.

Low pH induces a conformational change which increases the surface hydrophobicity, resulting in viral particle aggregation.

The detection of an altered trypsin cleavage pattern of the envelope proteins is indicative of conformational changes occurring following treatment with deforming agents. To assess whether such changes substantially altered virus particle structure, serum virus, pelleted through a 20% sucrose cushion, was either left untreated or exposed to pH 5 or 10 mM DTT or heated to 60°C before sedimentation over a preformed 20 to 70% sucrose gradient.

Gradient fractions were examined for virions by dot blot of viral DNA, and SVPs were detected by SDS-PAGE and silver staining for S protein (Fig. 5). Under the sedimentation conditions used (see Materials and Methods), DNA-containing particles and SVPs sediment according to their slightly differing densities, with substantial but incomplete separation of the two entities (Fig. 5A, pH 7.4, fractions 5 to 7 and 7 and 8, respectively). However, after pH 5 treatment for 16 h at 37°C (or pH 4 for 15 min at 37°C), both forms cosedimented, indicative of mixed aggregates of virus particles (Fig. 5B, fractions 6 and 7). The observed aggregation of these viral particles suggests that a major conformational change is occurring, which causes an increase in the hydrophobicity of the particle surface.

In contrast, virus particles treated with DTT (Fig. 5D), DTT in combination with pH 5 (Fig. 5E), or heat (Fig. 5C) did not aggregate but maintained a sedimentation profile similar to that of the untreated particles. Thus, although heat treatment or DTT facilitated the generation of the 25-kDa trypsin product, either alone or with the help of mildly acidic pH, such treatments failed to generate the hydrophobic conformation, indicating the absence of a major or further conformational rearrangement of L. To examine whether the conformation of the reduced particles still enabled low-pH aggregation to occur, reduced particles (DTT, 16 h at 37°C) were subsequently treated at pH 4 for 15 min and sedimented in a sucrose gradient. Such treatment resulted in virions and SVPs aggregating (Fig. 5F), indicating that reduced particles exhibit an alternative conformational change, primarily affecting the cysteine-containing loop region, while retaining competency for the low-pH conformational change.

Particles triggered to the hydrophobic L conformation bind to liposome membranes.

The exposure of hydrophobic domains and consequent particle aggregation lead us to speculate that these particles may be primed for membrane association or fusion. To examine whether such a potential existed, particles were incubated with liposomes under various conditions. The mixture was overlaid with a sucrose step gradient in which liposomes float to the top under centrifugation, and particles were detected in gradient fractions by Western blotting. Particles incubated with liposomes at pH 4 were detected in the top half of the gradient in association with the liposome fractions (Fig. 6A). In contrast, particles exposed to DTT at pH 5, like untreated particles, were located in the bottom half of the step gradient (Fig. 6B and C). Particles forced to aggregate through the low-pH conformational change prior to incubation with liposomes partially floated up the gradient, being detected in fractions within the 20 to 10% interface (Fig. 6D). Their partial association with liposomes appears to reflect their general hydrophobic character with a degree of aggregation also evident in a proportion of the untreated particles. In the absence of liposomes, aggregates banded as one fraction near the bottom of the gradient (Fig. 6E).

Formation of the hydrophobic conformation causes viral inactivation.

The structural changes resulting from various deforming treatments were also examined for their effect on viral infectivity (Table 1). Serum virus treated with low-pH conditions at 37°C, either pH 4 for 15 min or pH 5 for 16 h, failed to infect PDHs, as determined by the absence of production of intracellular viral DNA 9 days postinfection. Low-pH treatment at 4°C, a temperature not causing any detectable conformational changes of the envelope, did not affect viral infectivity. In contrast to low pH, DTT treatment, either alone or in combination with a short exposure to pH 5, had no effect on viral infectivity. These data (Table 1) are all in keeping with particle aggregation correlating with a loss in infectivity. Heat treatment of DHBV, previously known to result in a loss of infectivity (1), was not paralleled by particle aggregation. The mechanism by which viruses are inactivated by heat is not understood but may be the result of destabilization of the virus particle.

The loss of infectivity of viral particles which have aggregated, as observed here, following exposure to low pH is consistent with similar inactivation in several viruses by particle aggregation in the absence of a target membrane (5, 17).