Characterization of a hepatitis B and hepatitis delta virus receptor binding site


  • Potential conflict of interest: Nothing to report.


Insights into the early infection events of the human hepatitis B (HBV) and hepatitis delta virus (HDV) have been limited because of the lack of a cell culture system supporting the full replication cycle for these important pathogens. The human hepatoma cell line HepaRG allows the experimental induction of a differentiated state, thereby gaining susceptibility toward HBV and HDV infection. We recently identified HBV envelope protein–derived lipopeptides comprising amino acids 2 though 48 of the preS-domain of the L-surface protein, which block infection already at picomolar concentrations. To map the responsible sequence for the peptides' activity we describe an Escherichia coli expression system that permits myristoylation and investigated recombinant HBVpreS-GST fusion proteins with deletion- and point-mutations for their ability to prevent HBV and HDV infection. We found that (1) a myristoylated HBVpreS/2-48-GST fusion protein efficiently interferes with HBV infection of HepaRG cells; (2) deletions and point mutations in the highly conserved preS1 sequence between amino acids 11 through 21 result in the loss of infection inhibition activity; (3) hepatitis B viruses carrying single amino acid exchanges within this region lose infectivity; and (4) HDV infection of HepaRG cells can be inhibited by myristoylated HBVpreS peptides with the same specificity. In conclusion, HBV and HDV use at least one common step to enter hepatocytes and require a highly conserved preS1-sequence within the L-protein. This step is exceptionally sensitive toward inactivation by acylated HBVpreS1 peptides, which therefore represent a novel group of entry inhibitors that could be used for the treatment of hepatitis B and D. (HEPATOLOGY 2006;43:750–760.)

Hepadna- or hepatitis B viruses (HBV) are small enveloped DNA viruses that cause acute and chronic liver infections in mammals and birds. In the case of HBV, progression to the chronic state significantly increases the risk of developing liver cirrhosis and hepatocellular carcinoma. Because 400 million people worldwide suffer from chronic HBV infection, HBV is one of the most important human pathogens. Hepadnaviruses possess remarkable host specificities and preferentially replicate in hepatocytes of their respective hosts.1 Until recently, in vitro HBV infections were only successful in primary human hepatocytes (PHH) or hepatocytes of related primates.2, 3 Systematic investigations of HBV early infection events were thus difficult and varied with the quality of the hepatocyte preparation. The observation that hepatocytes of Tupaia belangeri are susceptible for HBV4, 5 and the recent establishment of the HepaRG cell line, which also supports the full HBV replication cycle,6 resolved this limitation and facilitated detailed studies on the contribution of HBV envelope protein domains in virus attachment and entry.

The HBV envelope consists of three membrane proteins termed large (L-), middle (M-), and small (S-) proteins. They are encoded in one open reading frame with three in-phase start codons.1 L- and M- share the S-domain, which serves as a membrane anchor but accomplishes other important roles, such as virus assembly and particle secretion.7 M is defined by an N-terminal extension of S by 55 hydrophilic amino acids (called preS2), whereas further extension with 108 (genotype D) or 119 (genotypes A and C) hydrophilic N-terminal amino acids (designated preS1) classify the L-protein. L displays remarkable features: it assumes a dual topology with approximately 50% of preS being translocated to the outside of the virion (eL = external L) and 50% remaining inside (iL = internal L). EL participates in receptor recognition whereas iL serves as a matrix protein responsible for nucleocapsid envelopment.8 L becomes myristoylated at glycin-2 of preS1 before translocation to the ER-lumen. This leads to an N-terminally myristoylated iL and, amazingly, with respect to the unusual membrane orientation, myristoylated eL. Although myristoylation of L is not important for virus assembly, it is indispensable for HBV infectivity.9, 10 Compelling evidence indicates that the preS1-domain of eL is involved in the recognition of (a) virus receptor(s) on hepatocytes: (1) HBV bearing deletions of 5 amino acids throughout the first 77 preS1 amino acids are non-infectious in PHH,11 whereas deletions of amino acids in preS2 that were tolerated during assembly were dispensable for infection.12 (2) Antibodies recognizing epitopes in the N-terminal third of preS1, but not preS2, neutralize HBV infection in vitro.4 (3) Myristoylated or otherwise acylated peptides representing the N-terminal 47 amino acids of preS1 (genotype D) block HBV infection of HepaRG cells, PHH, and primary tupaia hepatocytes even at picomolar concentrations. The sustained antiviral activity of the peptides after pre-incubation of the cells supported the idea of inactivating a specific receptor.13, 14

We here describe an expression system that allows large-scale production of highly active myristoylated HBVpreS-GST fusion proteins in Escherichia coli. We determined the inhibitory activities of HBVpreS1 mutants carrying internal deletions and identified a conserved amino acid sequence that is mandatory for receptor interaction of HBV and hepatitis D virus (HDV). This sequence plays an essential role in the HBV and HDV entry pathway, because single amino acid exchanges abrogated the antiviral activity of the peptides and rendered the respective recombinant HBV virions non-infectious.


HBV, hepatitis B virus; PHH, primary human hepatocytes; HDV, hepatitis delta virus; HBVpreS, hepatitis B virus preS polypeptide; eL, external L; iL, internal L; L-protein: hepatitis B virus large surface protein; HBsAg, hepatitis B surface antigen; HBeAg, hepatitis B e antigen; IPTG, iso-propyl-1- thio-β-D-galactosylpyranoside; GSH, glutathione; PEG, polyethylene glycol; NMT, N-myristoyltransferase; WMHBV, woolly monkey hepatitis B virus; DHBV, duck hepatitis B virus; WHV, woodchuck hepatitis B virus; MAb, monoclonal antibody.

Materials and Methods

Cultivation of HepaRG cells, preparation, and quantification of HBV and HDV stocks is described in the supplementary section. Supplementary material can be found on the HEPATOLOGY website (http//

Infection Competition Assays.

Differentiated HepaRG cells were incubated with the concentrated infectious virus,15 is 5- to 25-fold diluted in culture medium containing 4% PEG 8000, for 14 hours at 37°C. Cells were washed 3 times and maintained in the presence of 2% dimethylsulfoxide and 5 × 10−5 mol/L hydrocortisone. Supernatants were analyzed by hepatitis B surface antigen (HBsAg) or hepatitis B e antigen (HBeAg) ELISA. As shown in Supplementary Fig. 1 at the HEPATOLOGY website, HBsAg is suitable to measure infection at time points later than day 6 post-infection. For competition experiments, approximately 2 × 106 cells/6-well were pre-incubated for 30 minutes with recombinant preS-GST fusion proteins or synthetic myristoylated preS-peptides13 followed by a co-incubation of cells with peptide and virus for 20 hours.

Figure 1.

Expression, characterization, and infection inhibition activity of a myristoylated HBVpreS/2-48XaGST fusion protein derived from E. coli. (A) Schematic illustration of the expected myristoylated HBVpreS fusion protein. The coding sequences of preS1 (aa 1-48), the recognition site of endoprotease factor Xa (IEGR), and GST were merged by polymerase chain reaction (PCR) and ligated into plasmid pBB132 to permit expression of HBVpreS/2-48XaGSTmyr. Numbers in the figure indicate amino acid positions, beginning with Gly 2 of preS1. Note that Met 1 becomes posttranslationally removed from the N-terminus (see Table 1). (B) Coomassie-stained SDS-PAA gel (left) and the corresponding Western blot using the polyclonal antiserum H863 (right) of the GSH-agarose purified cytosolic extracts from E. coli, expressing HBVpreS/2-48XaGST in the presence (+NMT) or absence (-NMT) of yeast N-myristoyltransferase. Molecular weights are indicated. HBVpreS/2-48XaGST on the left assigns the myristoylated and non-myristoylated full-length fusion protein, both running at an apparent molecular weight of 32.3 kDa. NMT indicates yeast N-myristoyltransferase, which co-purifies with HBVpreS/2-48XaGSTmyr in NMT-expressing cells. The two bands marked by the stars point to the two N-terminally shortened fragments HBVpreS/15-48XaGST and HBVpreS/25-48XaGST as resolved by mass spectrometry (Table 1). (C) Comparison of calculated and experimentally determined molecular weights of preS-specific tryptic fragments obtained after hydrolysis of HBVpreS/2-48XaGSTmyr. (D) Comparative HBV infection competition assay in HepaRG cells using 0 nmol/L, 10 nmol/L, 40 nmol/L, 160 nmol/L, and 640 nmol/L HBVpreS/2-48XaGSTmyr (black bars) and, as a control, HBVpreS/2-48XaGST, purified from an E. coli culture that did not co-express NMT. Secreted hepatitis B surface antigen (HBsAg) between days 6 and 12 post-infection was detected and represented as percentages of the uncompeted infection. The sensitivity of the ELISA is indicated by the dotted line marked with cutoff. Note that the concentrations given below the bars are total protein concentrations, including NMT and the two N-terminally truncated (and therefore inactive) fragments. The effective concentration of the full-length HBVpreS/2-48XaGSTmyr is lower and can be calculated by multiplication with the factor 0.27, as given in Table 2.

Table 1. Molecular Weights of Endopeptidase Xa Fragments of HBVpreSXaGST Fusion Proteins
∴HBVpreSXaGST Fusion ProteinMr Calculated [Da]Mr Detected [Da]Matching PreS Fragment
HBVpreS/2-48XaGSTmyrΔ41-455311.925310.74preS/2-48Xa Δ41-45myr
 3735.783735.60preS/15-48Xa Δ41-45
 2514.162514.39preS/25-48Xa Δ35-39myr
HBVpreS/2-48XaGSTmyrΔ29-335267.915265.62preS/2-48Xa Δ29-33myr
HBVpreS/2-48XaGSTmyrΔ23-275305.875304.85preS/2-48Xa Δ23-27myr
HBVpreS/2-48XaGSTmyrΔ17-215304.885304.00preS/2-48Xa Δ17-21myr
 2178.352178.16preS/2-24 Δ17-21myr
 3728.733728.71preS/15-48Xa Δ17-21
HBVpreS/2-48XaGSTmyrΔ11-155333.845332.99preS/2-48Xamyr Δ11-15
HBVpreS/2-48XaGSTmyrL11R2811.652810.86preS/2-24 L11Rmyr
HBVpreS/2-48XaGSTmyrG12E2840.662839.81preS/2-24 G12Emyr
HBVpreS/2-48XaGSTmyrF13S2708.602707.83preS/2-24 F13Smyr
HBVpreS/2-48XaGSTmyrFF13/14SS2648.562648.17preS/2-24 FF13/14SSmyr
Table 2. Factors to Calculate Effective Concentrations of Myristoylated PreS-GST Fusion Proteins
preS-GST Fusion ProteinCorrection Factor

Infection of HepaRG Cells With Hepatitis D Virus.

HDV 5 × 106 particles per 12-well (multiplicity of genome equivalents ≈ 10) were incubated in the presence of 4% PEG for 24 hours with differentiated HepaRG cells. The cells were washed and supplied with fresh medium, which was exchanged every 72 hours. The supernatants were collected between days 8 and 11 post-infection. Cells were lysed and HDV RNA was quantified by real-time quantitative polymerase chain reaction (PCR) as described later.

Generation of HBV Mutants.

HBV particles bearing the preS1-mutations L11R, G12E, F13S, and FF13/14SS in their L-protein were generated by introducing nucleotide exchanges into the HBV-expression plasmid pCH-9/3091 by primer mutagenesis.16 pCH-9/3091 harbors a 1.1× overlength genome of HBV (ayw) under the control of a cytomegalovirus promoter. Details are described in the Supplementary Material. pCH-9/3091 was used in parallel to obtain WT-HBV particles. Sixteen hours and 2 days after transfection, cells were washed and supplied with fresh medium. At day 6, viral particles were 100× enriched either by PEG precipitation or by affinity chromatography on heparin columns. The virion concentration in these preparations was quantified by analytical cesium chloride density gradient centrifugation and subsequent DNA–dot-blot hybridization of the gradient fractions using an HBV-specific 32P-labeled probe. The same virus stock were used for every set of experiments

Generation and Subcloning of HBVpreSXaGST Fusions.

Subcloning of the HBVpreS-GST fusion proteins used in this study was performed by standard molecular biology techniques. A detailed description is provided in the Supplementary Material section.

Expression, Purification, and Characterization of Myristoylated preS-GST Fusion Proteins.

E. coli-DH5α (Invitrogen, Carlsbad, CA) transformed with the yeast-NMT expression plasmid pBB131 NMT17 and the respective preS expression plasmid (e.g., pBB132 HBVpreS/1-48XaGST) were grown at 37°C in 1 l TB medium containing 100 μg/mL ampicillin and 25 μg/mL kanamycin sulfate. At an OD600 nm of 0.5, 10 mL stock solution containing 5 mmol/L sodium myristate (Sigma, St. Louis, MO) and 0.6 mmol/L bovine serum albumin “fatty acid free” (Roche, Mannheim, Germany) was added and cells were grown for another hour. Expression of NMT was induced by the addition of iso-propyl-1- thio-β-D-galactosylpyranoside (IPTG) at a final concentration of 1 mmol/L. Thirty minutes later, expression of the preS-GST fusion protein was induced by nalidixic acid (50 μg/mL). After 3 hours at 27°C, cells were centrifuged (4,000g, 10 minutes, 4°C) and either stored at −20°C or resuspended in 30 mL phosphate-buffered saline supplemented with 2 mmol/L phenylmethylsulfonyl fluoride. Bacteria were homogenized (EmulsiFlex C5, Avestin) and centrifuged (14,000g, 30 minutes, 4°C). The supernatant was applied to a 2-mL glutathione (GSH)-agarose column (Amersham, Buckinghamshire, UK). After washing, bound proteins were eluted in fractions of 2 mL with phosphate-buffered saline (pH 7.4), containing 25 mmol/L GSH (Sigma). Protein concentrations were determined by measuring the OD at 280 nm. The purity was controlled by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Digestion of HBVpreS Fusion Proteins With Endoprotease Factor Xa.

The proteolytic digestion was initiated by the addition of 1/100 of the substrate of factor Xa (Roche, sequencing grade). Incubation was performed at 4°C for 18 hours.

Immunological Techniques.

After electrophoresis, proteins were transferred to a nitrocellulose filter for immunological analysis. As a primary antibody, we used monoclonal antibody (MAb) Ma18/718 or 5a1919 recognizing epitopes in preS1 or the rabbit polyclonal antiserum H863, generated against a recombinant HBVpreS1/preS2 protein. Detection was done by enhanced chemoluminescence (Amersham) according to the manufacturer's instructions.

Mass Spectrometric Analysis of HBVpreS-GST Fusion Proteins.

See Supplementary section. Supplementary material can be found on the HEPATOLOGY website (


Purification and Characterization of Myristoylated HBVpreS/2-48XaGST Fusion Proteins From E. coli.

To analyze the inhibitory activities of HBVpreS mutants without surrendering the opportunity for high-level expression in E. coli, we forced posttranslational myristoylation by co-expression of yeast N-myristoyltransferase (NMT) and the addition of myristic acid to the cell culture medium.17 Therefore, we transformed DH5α harboring the NMT expression vector pBB131-NMT with the vector pBB132-HBVpreS/1-48XaGST encoding the first 48 HBVpreS1 amino acids followed by the factor Xa cleavage site and the gene of gluthathion-S-transferase (Fig. 1A). After induction, the cytosolic fraction was applied to a GSH-affinity column. A control expression was performed in the absence of the NMT expression vector. Fig. 1B shows the SDS-PAA gel stained with Coomassie Blue (left) and the HBVpreS-specific Western blot (right) of the purified fusion proteins. In both cases preS-specific bands with the expected molecular weight of 32.3 kDa were detectable. Two smaller preS-specific bands at 30.9 and 28.4 kDa (indicated by stars) were also visible. Western blot analysis using the antibodies Ma18/7 (Fig. 2D) and 5a19 (not shown) as well as mass spectrometric analysis (Table 1) showed that they represent the two N-terminally shortened variants HBVpreS/15-48XaGST and HBVpreS/25-48XaGST, which presumably result through cleavage by (an) endogenous E. coli protease(s). The purified fraction from NMT-expressing bacteria showed an additional band at 56 kDa in Coomassie-stained SDS-PAA gels that did not react with preS-1 specific antibodies (Fig. 1B). Microsequencing identified it as yeast NMT, which apparently co-purifies with HBVpreS-GST proteins.

Figure 2.

Expression, characterization, and determination of the infection inhibition activity of myristoylated internal deletion mutants of HBVpreS/2-48XaGST. (A) Schematic illustration of myristoylated preS-GST fusion proteins carrying internal deletions in the N-terminal 48 preS1 amino acids. The full-length protein (WT) consisting of the N-terminal myristoyl moiety, followed by the 47 preS1-amino acids (black), the factor Xa protease cleavage site (gray), and GST is depicted on top. Below are enlarged preS/2-48 (WT) and the 7 deletions Δ5-9, Δ11-15, Δ17-21, Δ23-27, Δ29-33, Δ35-39, and Δ41-45. (B) Coomassie-stained SDS-PAA gel of HBVpreS/2-48XaGSTmyr (WT), HBVpreS/2-48XaGSTmyrΔ29-33, HBVpreS/2-48XaGSTmyrΔ35-39 and HBVpreS/2-48XaGSTmyrΔ41-45 (left) and a corresponding Western blot using monoclonal antibody (MAb) Ma18/7 (right) of the glutathione (GSH)-agarose purified proteins. Molecular weights are shown at the right. HBVpreS/2-48XaGSTmyr at the left designates the myristoylated full-length fusion protein; NMT indicates the co-purifying N-myristoyltransferase. Note the reduced molecular weights and the proteolytic fragments in all three deletions. Due to the specificity of MAb Ma18/7 (20-DPAF-23), the smallest fragment, is not detected in the Western blot. (C) HBV infection competition assay in HepaRG cells, using 0 nmol/L, 10 nmol/L, 40 nmol/L, 160 nmol/L, and 640 nmol/L HBVpreS/2-48XaGSTmyr, HBVpreS/2-48XaGSTmyrΔ29-33, HBVpreS/2-48XaGSTmyrΔ35-39, and HBVpreS/2-48XaGSTmyrΔ41-45. HBsAg secretion between days 6 and 12 post-infection was determined by ELISA. Values are given as percentage of a control infection in the absence of an inhibitor. The sensitivity of the ELISA (cutoff) is indicated. Note that the given concentrations include NMT and the two N-terminally truncated (and therefore inactive) fragments and should be corrected by the factors given in Table 2 to obtain the effective inhibitory concentration. (D) Coomassie-stained SDS-PAA gel of HBVpreS/2-48XaGSTmyr (WT), HBVpreS/2-48XaGSTmyrΔ5-9, HBVpreS/2-48XaGSTmyrΔ11-15, HBVpreS/2-48XaGSTmyrΔ17-21, and HBVpreS/2-48XaGSTmyrΔ23-27 (left) and two corresponding Western blots using the polyclonal antiserum H863 (top right) and MAb Ma18/7 (lower right) of the GSH-agarose purified fusion proteins. Molecular weights are indicated at the left. The full-length fusion protein is indicated with stars. Note that antibody H863 recognizes all mutants, whereas Ab MA18/7, due to its specificity, shows no reaction with Δ17-21 and Δ23-27. (E) HBV infection competition assay using 0 nmol/L, 25 nmol/L, 50 nmol/L, 100 nmol/L, and 500 nmol/L HBVpreS/2- 48XaGSTmyr (WT), HBVpreS/2-48XaGSTmyrΔ5-9, HBVpreS/2-48XaGSTmyrΔ11-15, HBVpreS/2-48XaGSTmyrΔ17-21, and HBVpreS/2-48XaGSTmyrΔ23-27. HBsAg secretion between days 6 and 12 post-infection was determined by ELISA. Values are given as percentage of an uncompeted control infection. The detection limit (cutoff) is indicated by the dotted line. Note that the given concentrations include NMT and the two N-terminally truncated inactive fragments and should be corrected by the factors given in Table 2 to obtain the effective inhibitory concentration.

To examine whether the HBVpreS/2-48XaGST fusion protein becomes properly modified by recombinant NMT, we analyzed tryptic fragments of it by mass spectrometry. As indicated by bold letters in Fig. 1A, Arg-24 of the preS1 sequence and Arg-52, which constitutes the C-terminus of the Xa recognition site IEGR, serve as cleavage sites for trypsin. If myristoylation had taken place, the calculated average molecular weights of the expected fragments HBVpreS/2-24myr and HBVpreS/25-48Xa should be 2768.6 Da and 3142.4 Da, respectively. This is in excellent accordance with the fragment sizes found experimentally (Fig. 1C), indicating that accurate N-terminal removal of methionine and subsequent myristoylation of glycin-2 of the HBVpreS/2-48XaGST fusion protein occurred.

To obtain the free HBVpreS/2-48Xamyr peptide, we digested HBVpreS/2-48XaGSTmyr with endopeptidase Xa and analyzed the products by SDS-PAGE, Western blot, gel filtration, and mass spectrometry. We found that correct cleavage occurs, resulting in the expected product (Table 1). However, shorter fragments such as HBVpreS/2-24myr were also detectable.

To examine whether the myristoylated GST-preS fusion protein interferes with HBV infection, we performed competition assays in HepaRG cells. Cells were infected in the presence of the indicated concentrations of myristoylated HBVpreS/2-48XaGSTmyr and, as a control, the non-myristoylated HBVpreS/2-48XaGST. As depicted in Fig. 1D, HBVpreS/2-48XaGSTmyr abrogated infection as determined by HBsAg secretion in a concentration-dependent manner, with an almost complete reduction to background levels at concentrations of approximately 40 nmol/L. At 160 and 640 nmol/L, no signals exceeding the sensitivity of the assay were detectable. Because the protein preparation contained free NMT as well as the two unmyristoylated, inactive fragments HBVpreS/15-48XaGST and HBVpreS/25-48XaGST, the indicated concentrations have to be corrected by a factor of 0.27 as calculated by quantification of the Coomassie-stained bands (Table 2). Thus, nearly complete inhibition has to be assumed at 11 nmol/L of the pure fusion protein, a concentration that compares to the specific activity of the unfused synthetic peptide HBVpreS/2-48myr. Like the non-acylated synthetic HBV preS-peptides, HBVpreS/2-48XaGST showed no interference with infection at the same concentrations.13 This emphasizes the specificity of the competition assay and excludes the possibility that contaminants from E. coli are responsible for the effect.

Infection Competition Activity of Internally Deleted HBVpreS-GST Fusion Proteins.

Using C-terminally shortened synthetic peptides, we found that the sequence located between amino acids 19 and 48 gradually contribute to the infection competition activity. Amino acids 2 through 19 were essentially required, because an artificially myristoylated HBVpreS/19-48myr peptide was completely inactive.13 To precisely define preS-sequence elements crucial for the competition activity, we examined a set of myristoylated HBVpreS/2-48XaGST fusion proteins comprising internal 5 amino acids deletions throughout the 48 N-terminal preS1 amino acids (Fig. 2A). Figure 2B and D depict the Coomassie-stained SDS-PAA gels (left) and the corresponding Western blots (right) of the purified fractions from the GSH-agarose column. As shown in Fig. 2B, the deletion mutants Δ29-33, Δ35-39, and Δ41-45 were obtained in amounts comparable to the full-length protein (WT). As expected, the deletion mutants migrate at slightly lower molecular weights in SDS-PAGE. In all preparations, NMT and the two N-terminally shortened degradation products were detectable, the latter with the expected reduced molecular weights, due to cleavage sites N-terminal to the deletion. All proteins were analyzed by mass spectroscopy after digestion with endopeptidase Xa (Table 1). In addition to the verification of myristoylation, this allowed the confirmation of the mutations at the protein levels.

Fig. 2C depicts the activities of the deletion mutants Δ29-33, Δ35-39, and Δ41-45 through determination of newly synthesized HBsAg at days 6 through 12 post-infection. All proteins interfere with HBV infection in a concentration-dependent manner. However, compared with the full-length protein that inhibits infection almost completely at an apparent concentration of 40 nmol/L, Δ35-39 and Δ41-45 displayed an approximately fivefold reduction in their specific activities. HBVpreS/2-48XaGSTmyrΔ29-33 showed only approximately 50% inhibition at 640 nmol/L. We extended this type of investigation using the more N-terminal preS-deletions affecting amino acids 23-27, 17-21, 11-15, and 5-9. Verification of the purity and integrity of the corresponding proteins is shown in Fig. 2D and Table 1. With the exception of HBVpreS/2-48XaGSTmyrΔ5-9, all mutants became myristoylated. For HBVpreS/2-48XaGSTmyrΔ5-9, we were not able to detect a myristoylated N-terminal fragment. We have therefore only the indirect evidence of NMT co-elution that myristoylation occurred. Concerning the infection competition activities (Fig. 2E), HBVpreS/2-48XaGSTmyrΔ23-27 exhibited a significant but weakened inhibitory effect, similar to HBVpreS/2-48XaGSTmyrΔ29-33 with 50% inhibition at 640 nmol/L. In contrast, all the mutants with further N-terminal deletions HBVpreS/2-48XaGSTmyr Δ17-21, Δ11-15, and Δ5-9 did not show significant interference with HBV infection, indicating a pivotal role of this preS part for infection competition.

A Duck Hepatitis B Virus/HBV Interspecies Domain Swap Abrogates the Infection Inhibition Activity of HBVpreS/2-48XaGSTmyr.

The requirement for the N-terminal 18 preS1-amino acids and the dependence of an acyl residue linked to glycine-2 for efficient infection inhibition might be due to a linker-function to correctly position the more C-terminal amino acids 19 through 48. To resolve this question, we produced the chimeric duck hepatitis B virus (DHBV) preS/2-15-HBVpreS/16-48XaGSTmyr fusion protein (Fig. 3A) and tested its ability to block HBV infection. This protein encodes the first 15 preS amino acids of the DHBV L-protein, followed by amino acids 16 through 48 of HBVpreS1, the Xa cleavage site, and GST. As a control we produced the DHBV construct DHBVpreS/1-44XaGST. Western blot analyses using MAb Ma18/7 (recognizing the epitope 20-DPAF-23) and a DHBV preS-specific antiserum confirmed the presence of both epitopes in the chimera (Fig. 3B). DHBVpreS/1-44XaGST is not detectable by Ma18/7, and HBVpreS/2-48XaGSTmyr does not react with αDHBVpreS. Mass spectrometric analysis of the chimera showed that proper myristoylation occurred (Table 1). Whereas HBVpreS/2-48XaGSTmyr blocked HBV infection at the expected concentrations, DHBVpreS/1-44XaGST and DHBVpreS/2-15-HBVpreS/16-48XaGSTmyr were inactive. The latter led to an approximately 1.8-fold enhancement of infection for unknown reasons. This result indicates that specific sequence elements located in the N-terminal 15 preS amino acids are essential for infection interference.

Figure 3.

Expression and characterization of a chimeric myristoylated DHBV/HBV preS-GST fusion protein. (A) The chimeric duck hepatitis B virus (DHBV)/HBVpreS-GST fusion protein DHBVpreS/2-15-HBVpreS/16-48XaGSTmyr used for the infection competition studies depicted in C. Numbers above the bars indicate amino acids that constitute the boundaries between DHBV (black), HBV (dark gray), Xa (light gray), and GST sequences. As a control, the DHBVpreS-GST fusion protein DHBVpreS/1-44XaGST (shown below) was constructed. (B) Western blot analysis of DHBVpreS/2-15-HBVpreS/16-48XaGSTmyr and DHBVpreS/1-44XaGST using the HBV preS1 specific antibody MA18/7 (left) and anti-DpreS2-41 recognizing the N-terminus of the DHBV L-protein (right). As controls, HBVpreS/2-48XaGSTmyr and a recombinant DHBVpreS/1-165 protein21 were included. Myristoylation was verified by mass spectrometry (Table 1). (C) HBV infection competition assay in HepaRG cells, using 0 nmol/L, 10 nmol/L, 40 nmol/L, 160 nmol/L, and 640 nmol/L HBVpreS/2-48XaGSTmyr (left), the DHBV/HBV chimera (middle), and the DHBV preS-GST fusion protein (right). Hepatitis B surface antigen (HBsAg) secretion between days 6 and 12 post-infection was determined by ELISA. Values are given as percentage of an uncompeted control infection. The sensitivity of the ELISA (cutoff) is indicated by the dotted line.

Single Amino Acid Exchanges in a Highly Conserved N-Terminal preS-Region Abrogate Infection Inhibition Activity.

PreS-sequence alignments of different HBV genotypes as well as woolly monkey hepatitis B virus (WMHBV) revealed two highly conserved regions (amino acids 9-13 and 17-21) that are part of the essential inhibitory domain defined above (Fig. 4A). To investigate the contribution of particular amino acids within this region, we introduced single (L11R, G12E, F13S) and one double (FF13/14SS) amino acid exchanges (Fig. 4B). The fusion proteins were characterized (Fig. 4C) and analyzed by mass spectrometry (Table 1). Equal concentrations of the mutant proteins were compared with HBVpreS/2-48XaGSTmyr in their ability to block HBV infection [Fig. 4D (HBsAg) and 4E (HBeAg)]. Whereas HBVpreS/2-48XaGSTmyr competed with infection in the expected manner, the point mutants showed no (G12E, F13S) or only negligible effect on infection, even at concentrations of 500 nmol/L. This underlines the specificity of the inhibitory effect and demonstrates the importance of this highly conserved sequence.

Figure 4.

Infection inhibition activity of myristoylated point mutants of HBVpreS/2-48XaGST. (A) Sequence alignments of different HBV preS sequences: (1) genotype D used in this study, (2) genotype A, (3) genotype B, (4) genotype E, (5) genotype F, (6) genotype G, and (7) woolly monkey hepatitis B virus (WMHBV) preS sequence. The numbers at the top represent the amino acid position of genotype D. Identical amino acids in the different sequences are highlighted by boxes in dark gray; conservative exchanges are in light grey. Not boxed are regions with high sequence variability. (B) Schematic illustration of myristoylated preS-GST fusion proteins carrying point mutations (indicated by bold letters) at positions 11, 12, and 13 in comparison with the WT sequence at the top. Note that the mutations have been chosen to preserve the integrity of the overlapping reading frame of the HBV polymerase. (C) Coomassie-stained SDS-PAA gel of HBVpreS/2-48XaGSTmyrL11R, HBVpreS/2-48XaGSTmyrG12E, HBVpreS/2-48XaGSTmyrF13S, and the double point mutant HBVpreS/2-48XaGSTmyrFF13/14SS (left) and a corresponding Western blot using the antiserum H863 (right). Molecular weights are shown at the right; NMT indicates the co-purified N-myristoyltransferase. Note that exchange of F14 to S in the double mutant FF13/14SS leads to disappearance of the band at 30 kDa, due to the destruction of a cleavage site. (D) HBV infection competition assay using 0 nmol/L, 25 nmol/L, 50 nmol/L, 100 nmol/L, and 500 nmol/L HBVpreS/2-48XaGSTmyr (WT), HBVpreS/2-48XaGSTmyrL11R, HBVpreS/2- 48XaGSTmyrG12E, HBVpreS/2-48XaGSTmyrF13S, and HBVpreS/2- 48XaGSTmyrFF13/14SS. Hepatitis B surface antigen (HBsAg) secretion between days 6 and 12 post-infection was determined by ELISA. Values are given as percentage of an uncompeted control infection. The detection limit (cutoff) is indicated.

Introduction of Point Mutations in the Highly Conserved N-Terminal PreS Sequence of HBV Abrogates Infectivity.

Infection inhibition by HBVpreS-derived peptides is caused by interference with an early infection event that presumably also comes to pass for the complete virion. We therefore introduced the point mutations L11R, G12E, F13S, and FF13/14SS into the L-protein sequence of the HBV expression vector pCH/9-3091. None of the introduced preS1-mutations change the overlapping polymerase open reading frame. After transfection of HuH7 cells, virus-containing cell culture supernatants were subjected to analytical CsCl density gradients. DNA-containing virion fractions as well as free nucleocapsids were quantified by DNA Dot-Blot (Fig. 5A). None of the mutants was deficient in assembly. HepaRG cells were infected with the concentrated supernatants of pCH/9-3091-transfected HuH7 cells (WT) in the absence or presence of synthetic HBVpreS/2-48myr or the mutants L11R, G12E, F13S and FF12/14SS (Fig. 5B). Whereas WT inoculation of HepaRG cells results in productive infection, all point mutants showed a complete loss in infectivity. This indicates that single point mutations in this highly conserved preS-region are deleterious for productive virus entry into hepatocytes.

Figure 5.

Infectivity of HBV mutants carrying preS point mutations. (A) HBV-DNA specific dot-blot analysis of fractions of an analytical CsCl density-gradient obtained from supernatants of HuH7 cells transfected with the HBV expression plasmids encoding either WT or the mutated L-proteins L11R, G12E, F13S, and FF13/14SS. The density in the gradient is indicated by the arrow below. Free nucleocapsids peak in fractions 3 and 4, virions in fractions 6 and 7 as indicated at the top. (B) Infection of HepaRG cells using equal multiplicities of genome equivalents of approximately 100-fold concentrated virus preparations obtained from the supernatants of the transfected HuH7 cells characterized above. The control infection using WT virus from supernatants of transfected HuH7 cells is shown on the left. The specificity of the infection was confirmed by its inhibition with 100 nmol/L synthetic peptide HBVpreS/2-48myr.

HDV Infection of HepaRG Cells Is Inhibited by Myristoylated HBVpreS Fusion Proteins.

HDV contains an envelope that resembles the envelope of hepatitis B virions. It has therefore been suggested that HBV and HDV use the same cellular receptor molecule(s) and employ a comparable fusion mechanism.20 We therefore investigated the interference of HDV infection by myristoylated HBVpreS fusion proteins that have been characterized previously. We transfected the HBV-producing cell line HepG2.2.15 with the HDV-encoding plasmid JC126, concentrated the HBV- and HDV-containing supernatants by PEG-precipitation, and performed a mixed infection of HepaRG cells in the presence of 500 nmol/L HBVpreS/2-48XaGSTmyr, the two deletion mutants Δ11-15 and Δ17-21, and the preS point mutants L11R, G12E, F13S, and FF13/14SS. We compared the results with an uncompeted infection and an infection performed in the presence of the HHBVpreS/2-44myr control peptide. Figure 6A shows the cell-associated HDV genome equivalents determined by quantitative real time RT-PCR, 11 days post-infection. Figure 6B depicts the amounts of secreted HBsAg (days 7-11) from culture supernatant of the same cells. Whereas HDV replication was unaffected by any of the supplied mutant HBVpreS-peptides, an approximately 100-fold reduction was achieved in the presence of HBVpreS/2-48XaGSTmyr. This indicates that HepaRG cells support HDV infection, which can be inhibited by myristoylated HBVpreS peptides with the same specificity as HBV.

Figure 6.

Inhibition of hepatitis delta virus (HDV) infection by myristoylated HBVpreS/2-48XaGST fusion proteins. (A) HDV infection competition assay in HepaRG cells using 500 nmol/L HBVpreS/2-48XaGSTmyr, the two deletions Δ11-15 and Δ17-21 and the four point mutants L11R, G12E, F13S, and FF13/14SS in comparison with the uncompeted infection (left) and an infection in the presence of a peptide derived from the heron hepatitis B virus HHBVpreS/2-44myr.22 HDV genome equivalents were quantified using real-time polymerase chain reaction (PCR) after reverse transcription of HDV RNA extracted from HepaRG cells, 11 days post-infection. (B) Quantification of hepatitis B surface antigen (HBsAg) collected from the supernatants of the HDV/HBV-coinfected HepaRG cells used in A. Note that the secreted HBsAg resulting from the HBV infection entirely correlates to the intracellular HDV RNA.


Based on our previous observations describing acylated preS1-peptides as efficient entry inhibitors of the HBV infection in susceptible hepatocytes,13, 14 we performed a comprehensive analysis of the sequence requirement(s) for this effect. We established an E. coli expression system allowing the production of large quantities of myristoylated preS-GST fusion proteins with comparable activities as their synthetic counterparts. Thus, C-terminal fusion of a molecule with approximately fivefold the size of the peptide is tolerated without a loss of function. Introduction of an endo-protease Xa cleavage site facilitates the release of an unfused myristoylated peptide equivalent to synthetic HBVpreS/2-48myr. Given that yeast NMT is capable of transferring other activated fatty acids such as palmitoyl-CoA instead of myristoyl-CoA to appropriate substrates, this system also may be used to generate other acylated preS-peptides with higher specific activities.13 Our system therefore provides an inexpensive alternative for the production of acylated peptide inhibitors of HBV infection for analytical and preparative purposes.

The major objective of our study was the application of this peptide expression system for the mapping of amino acids that play crucial roles in HBV entry and its inhibition. Using several HBVpreS/2-48XaGST fusion proteins bearing internal preS-deletions, we confirmed that amino acids 22 through 48 significantly contribute to peptide activity. However, none of the deletions led to a complete inactivation. In contrast, removal of amino acids in the N-terminal part of preS1 (Δ11-15 and Δ17-21) resulted in a complete abrogation of peptide activity at a concentration of 0.5 μmol/L. Assuming that the peptides address a hepatocyte surface receptor (see below), we have previously hypothesized whether the N-terminal acyl residue of the peptide serves as a membrane anchor to correctly position an active site consisting of amino acids 19 to 48 an exact distance from the cell surface.13 In that case, amino acids 2 through 18 would serve a spacer function and may be replaced by another sequence, such as from the DHBV preS-domain. The fact that the chimeric protein DHBVpreS/2-15-HBVpreS/16-48XaGSTmyr is inactive (Fig. 3) disproved this hypothesis and emphasizes a sequence-specific contribution of the N-terminal HBV preS region to infection inhibition. This is in accordance with the high degree of sequence homology within this particular preS-part of all HBV subtypes and even WMHBV (Fig. 4A). The preS domains of hepadnaviruses are extremely variable, with in some cases more than 50% variation within one genus and barely any recognizable similarities between HBV, WHV, and DHBV.

The observation that single amino acid exchanges at positions 11, 12, or 13 were sufficient to inactivate the peptide stresses the extraordinary specificity of the underlying inhibition mechanism. Furthermore, it indicates an absolute requirement of this conserved sequence and raises a question about the target that is addressed. Because preincubation of HepaRG cells with HBVpreS/2-48myr is sufficient to render them insusceptible to HBV, we assumed13 that a specific hepatocyte receptor is inactivated by occupation of the binding site. If this holds true, and binding depends on the integrity of the conserved preS-sequence, it might be impossible for the virus to escape this mode of inhibition by mutation. However, we cannot exclude a more complex mechanism, involving not only binding of the peptide to its postulated target but involvement in a receptor/fusion complex that depends on additional viral proteins (e.g., the S-protein). Considering possible future clinical applications of preS-derived inhibitors, elucidation of the mechanism of inhibition is of utmost importance because all approved nucleoside analogs used for the treatment of chronic hepatitis B induced the selection of resistant mutants. Our observation that the mutant virions HBVpreS/L11R, HBVpreS/G12E, HBVpreS/F13S, and HBVpreS/FF13/14SS entirely lost their ability to infect HepaRG cells strengthens the validity of peptide-mediated infection competition as an appropriate approach to investigate early HBV infection events. It also raises a question regarding the particular defect the mutants bear during the entry process. In the future, the establishment of highly sensitive binding assays (e.g., by real time PCR or visualization of fluorescently labeled particles) will allow us to decide whether binding or fusion is affected.

Due to similarities in their surface protein composition, HDV and HBV have been hypothesized follow related pathways to enter hepatocytes.20 Recently, Barrera et al.23 demonstrated that HDV infection of PHH can be blocked by myristoylated HBVpreS derived peptides. The data we present using HepaRG cells confirm their observation and further demonstrate that inhibition of HDV infection follows the same sequence specificity as demonstrated for HBV (Fig. 6). We therefore conclude that HDV utilizes—at least with respect to the step addressed by the myristoylated preS-peptides—the same entry pathway as HBV. HDV thus represents a valid surrogate system for the analysis of the HBV entry events and provides in addition the advantage of highly sensitive detection of infection markers.

The specific interference with HBV and HDV infection by N-terminally acylated fragments of the HBV-L-protein will allow the characterization of cellular receptor molecules that are involved in the entry of HBV and HDV into their target cells, the hepatocyte. Moreover, and similar to the gp41-derived T-20 peptide in HIV,24 the use of HBVpreS peptides represents a novel therapeutic antihepadnaviral strategy, namely combating the virus at the earliest point in its replication cycle. Medical indications for this type of inhibitor could be the prevention of HBV/HDV reinfection after liver transplantation, post-exposure prophylaxis, or horizontal transfer of HBV/HDV from an infected mother to her child. However, an interesting question remains as to whether efficient entry inhibition via preS peptides can improve the therapeutic outcome in chronically infected patients, either alone or in combination with established therapies that reduce the number of infected hepatocytes.


The authors thank Stephanie Held for technical assistance, Anja Kristina Meier and Silke Schmidt for help with the cloning and purification of HBV mutants. We thank Heinz Schaller for the plasmids pCH-9/3091 and pCD0, Jefry I. Gordon, Washington for the NMT expression vector pBB131 and the Cα expression vector pBB132, John Taylor for the HDV expression plasmid JC126, and Michael Roggendorf for WHV-encoding plasmids and antisera from HDV infected patients. We are indebted to Ralf Bartenschlager, who continuously contributes to our work.