INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

The Hepatitis B virus (HBV)¹ e antigen (HBe) is found into the serum of patients suffering from acute hepatitis (1). Even though a role for infectivity or viral multiplication is excluded (2-4), conservation of this antigen during evolution (all viruses from the family encode a similar e antigen) suggests that it has an important role in the life cycle of hepadnaviruses.

HBe derives from a precursor (the precore protein) encoded by the entire HBV C open reading frame, which contains two in-frame initiation codons delimiting the pre-C sequence (87 nucleotides) and the C gene. The precore protein is translated from the pre-C AUG on the pre-C RNA, whereas the core protein (the subunit of the capsid) is translated from the C AUG on the pregenomic RNA, a slightly shorter transcript that does not include the pre-C AUG. The precore protein, a 25-kDa unglycosylated protein (P25), is directed to the secretory pathway by a 19-amino acid-long signal sequence that is cleaved during translocation into the lumen of the endoplasmic reticulum (ER) (5, 6), producing a 22-kDa protein (P22) (Fig. 1A). P22 is further processed in a post-ER compartment by removal of its C-terminal extremity (7, 8), most likely through a multiple-steps process.² The resulting mature HBe (17 kDa) is then secreted into the blood in a monomeric form (9). The cleaved C-terminal domain of P22 is 34 amino acids long and contains 16 arginine residues arranged in clusters (arginine-rich domain) (Fig. 1B). Maturation of e antigen from precore protein is similar for two other hepadnaviruses (10, 11).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Schematic representation of HBe synthesis. A, the precore protein P25 (212 amino acids) is directed to the ER by a 19-amino acid-long signal sequence located at its N terminus. This signal sequence is removed by a signal peptidase during translocation into the ER, leading to the P22 luminal protein. P22 is then transported through the secretory pathway and further processed in a post-ER compartment into mature HBe (HBeAg) by a cellular protease that eliminates the 34 C-terminal amino acids (arginine-rich domain). Mature HBeAg is then secreted (see text for details). B, P22 C-terminal domain primary sequence. Amino acid sequence of the P22 C-terminal domain (subtype ayw) is shown in the single-letter code. Numbers above indicate the positions of residues relative to the P22 C terminus. The arrow indicates the HBe C terminus.

It has been shown that the cleaved C-terminal domain of P22 is crucial for the efficiency of the HBe secretion process (12, 13). Here, we provide evidence that a cellular protein contributes to the efficiency of this process by a direct or indirect interaction with the P22 C-terminal domain.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Plasmids-- Schematic representation of proteins referred to in this study is shown in Fig. 2. Proteins were expressed under the control of the adenovirus major late promoter. Plasmids pHPC, pHPC²⁵ and pHPC³⁹ have been described previously (13). For plasmids pHPC¹⁸ and pHPC¹⁴, stop codons were introduced, respectively, at codons 195 or 199 of the entire C open reading frame by site-directed mutagenesis of pHPC using polymerase chain reaction (14). Plasmid pPC-LZ was assembled from pHPC and pGH101 (15). Plasmid pPCLZE derives from pPC-LZ and contains a stop codon at codon 1003 of the Escherichia coli lacZ gene. Plasmids pPC-LZE-C58 and pPC-LZE-C39 were assembled from pPC-LZ and pHPC.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Schematic representation of studied proteins. A, plasmid pHPC encodes the HBV precore protein, P25. The 19-amino acid-long signal sequence at the N terminus and the 34-amino acid-long arginine-rich domain at the C terminus are indicated. Plasmids pHPC14, pHPC18, pHPC25, and pHPC39 encode C-terminal-truncated precore proteins lacking 14, 18, 25, or 39 amino acids, respectively. B, plasmid pPC-LZ encodes E. coli -galactosidase (dark rectangle), fused at its N terminus to the 25 first residues of P25. PCLZE corresponds to PCLZ truncated for the last 17 residues. PCLZEC58 and PCLZEC39 correspond to PCLZE fused to, respectively, the 58 or 39 last residues of P25.

Labeling of Transfected Cells and Immunoprecipitation-- Adenovirus-transformed human embryo cells (line 293-31) (16) were grown as described (13). Cells at 80% confluency were transfected by the calcium phosphate method (17) with 30 µg of DNA/100-mm dish. Forty-eight h post-transfected cells were grown for 1 h in 10 ml of methionine-free cysteine-free Eagle's minimal essential medium (ICN), then for 3 h in 6 ml of methionine-free cysteine-free Eagle's minimal essential medium containing 500 µCi of Pro-Mix protein labeling mix (Amersham Pharmacia Biotech, specific activity >1,000 Ci/mmol). After labeling, media and cell extracts were prepared, and proteins were immunoprecipitated and analyzed as described previously (13, 18).

Time Course Experiments-- Forty-eight h post-transfected cells were grown for 1 h in 10 ml of Eagle's minimal essential medium then for 2, 3, or 5 h in 6 ml of Eagle's minimal essential medium containing 500 µCi of Pro-Mix protein labeling mix. After labeling, media and cell extracts were prepared, and proteins were immunoprecipitated and analyzed as described above.

Tunicamycin Treatment-- Cells were exposed to 6 µM tunicamycin (Boehringer Mannheim) for 6 h before methionine/cysteine depletion in 10 ml of fresh Dulbecco's modified Eagle medium. Tunicamycin was also present during depletion and protein labeling.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

HBe Secretion Decreases in Cells in Which N-Glycosylation Is Abolished-- To obtain new insights into the mechanism of P22 intracellular transport, we first examined the effect of the inhibition of N-glycosylation upon HBe secretion. Cells expressing the precore protein were grown in the presence of tunicamycin, an inhibitor of N-glycosylation. As shown on Fig. 3, the addition of tunicamycin 7 h before labeling provoked a significant and reproducible decrease in the amount of secreted HBe, whereas the amount of neosynthesized P22 was not reduced. This diminution of HBe secretion cannot be explained by a direct effect of tunicamycin on the biosynthesis of HBe inasmuch as there is no N-glycosylation consensus site in the sequence of the precore protein and must be interpreted as an indirect consequence of the inhibition of N-glycosylation.

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Effect of tunicamycin on HBe secretion. Cells were transfected with plasmid pHPC and 48 h later were metabolically labeled for 3 h. Tunicamycin (6 µM) was added 7 h before labeling and was present up to the end of labeling (lanes 2 and 4). Proteins from cell extracts (lanes 1 and 2) or media (lanes 3 and 4) were immunoprecipitated with anti-HBV core antigen antiserum and analyzed on a 12.5% SDS-PAGE (see "Experimental Procedures"). On the right are indicated migrations of P22 and mature HBe and on the left migrations of 14C-labeled molecular mass standards (Amersham Pharmacia Biotech) in kDa.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Effect of tunicamycin on the secretion of a chimeric -galactosidase. pPC-LZ-transfected cells were grown and metabolically labeled in the presence or in the absence of tunicamycin as in the legend of Fig. 3. After labeling, proteins from cell extracts or media were immunoprecipitated with anti--galactosidase antiserum and separated on 7.5% SDS-PAGE. On the right are indicated migrations of PCLZ and its glycosylated form, PCLZ*. On the left are indicated migrations of molecular mass standards in kDa.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Effect of tunicamycin on the time course of HBe secretion. pHPC-transfected cells were metabolically labeled for 2, 3, or 5 h as indicated. Seven h before labeling, 6 µM tunicamycin was added (+) or not () to the culture medium and was present up to the end of labeling. Proteins from culture media were immunoprecipitated with anti-HBV core antigen antiserum and analyzed on 12.5% SDS-PAGE. On the left is indicated the migration of mature HBe and on the right, the migrations of molecular mass standards in kDa.

The Action of the Cellular Tunicamycin-sensitive Protein Requires the P22 C-terminal Region-- As the C-terminal domain of P22 is important for HBe secretion (12, 13), it was tempting to speculate that TSP would interact with this region. To test this hypothesis, we first determined if the inhibition of N-glycosylation would still adversely affect the secretion of HBe derived from precursors truncated at different positions³ within the C-terminal part: P25¹⁴, P25¹⁸, P25²⁵, and P25³⁹ (Fig. 2A). As shown in Fig. 6, tunicamycin treatment reduced the level of secretion of HBe, HBe¹⁴ and HBe¹⁸ but not that of HBe²⁵ or HBe³⁹. In this particular experiment, HBe¹⁸ appears to migrate slightly slower when synthesized in the absence of tunicamycin, and HBe²⁵ appeared as two fuzzy bands when tunicamycin was present. These slowest migrating bands most likely correspond to non-fully mature HBe²⁵ molecules, which are sometimes observed in pHPC²⁵-transfected cells, even in the absence of tunicamycin. However, these results show that a region located upstream of position 18 (relative to the P22 C terminus) is crucial for tunicamycin sensitivity of HBe secretion. This strongly suggests that the C terminus of P22 facilitates an interaction with TSP. Furthermore, as HBe²⁵ and HBe³⁹ were still secreted in the presence of tunicamycin, the action of TSP, even if it is required for optimal HBe secretion, is dispensable.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Effect of tunicamycin on the secretion of HBe derived from C-terminal-truncated precore proteins. Cells were transfected with plasmids pHPC, pHPC14, pHPC18, pHPC25, or pHPC39 and grown in the presence or in the absence of tunicamycin as described in the legend of Fig. 3. After labeling, proteins from culture media were immunoprecipitated with anti-HBV core antigen antiserum and analyzed on 12.5% SDS-PAGE. Migration of mature HBe derived from P22, P2214, P2218, or P2225 (HBe) and migration of HBe39 are indicated on the right. On the left are indicated migrations of molecular mass standards in kDa.

View larger version (23K): [in this window] [in a new window]   Fig. 7.   Effect of tunicamycin on the secretion of -galactosidase fused to the C-terminal region of HBe precursor. pPC-LZE-C58 (A) or pPC-LZE-C39 (B) -transfected cells were metabolically labeled for 3 h. Tunicamycin was present or not as in the legend of Fig. 3. Proteins from cell extracts or culture media were immunoprecipitated with anti--galactosidase antiserum and separated on 7.5% SDS-PAGE. On the left are indicated migrations of expressed proteins; the stars indicating the glycosylated form. On the right are indicated migrations of the molecular mass standards in kDa.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Our study shows that the inhibition of N-glycosylation in cells expressing the HBe precursor led to a significant decrease in HBe secretion, an unexpected finding since the HBe precursor contains no glycosylation consensus sites. A general effect of tunicamycin on the cellular secretion machinery is ruled out as the secretion of HBe²⁵, HBe³⁹, and PCLZ was not adversely affected by the inhibitor. Thus, our findings are consistent with the idea that a cellular protein increases specifically HBe secretion and that the activity of this protein (TSP) depends upon the N-glycosylation status of the cell.

The molecular mechanism by which TSP increases HBe secretion remains to be determined, although we demonstrate that the tunicamycin sensitivity of secretion absolutely requires a part of the P22 C-terminal domain, a region absent in mature HBe. First, P22 truncations larger than 18 amino acids abolished tunicamycin sensitivity, suggesting that the sequence upstream of position 18 is important for the interaction with TSP. Second, the 39 C-terminal amino acids of P22 conferred tunicamycin sensitivity on PCLZEC39. Taken together, these results demonstrate that a sequence located between amino acids 19 and 39 is involved in the interaction with TSP. However, since in PCLZEC39 both the N- and C-terminal domains of P22 are present, we cannot exclude the possibility that the P22 N terminus may also be involved in this interaction, a possibility not supported by reports on the native structure of HBe (7, 21) and -galactosidase (22).

How might tunicamycin affect the activity of TSP? The simplest explanation is that TSP is only active when N-glycosylated. Another possibility is that its activity is indirectly affected by tunicamycin. TSP would require the involvement of a N-glycoprotein for its correct folding, processing, subcellular localization, or activity. The next question is how does such a protein function. TSP could be a chaperone that favors the folding of P22, increasing the amount of P22 that exits from the ER and is further transported to the cell surface. The first possibility is that the sequence 19 to 39 promotes the binding of TSP in a direct manner. This seems unlikely since it is generally assumed that, to recognize unfolded proteins, chaperones interact with exposed hydrophobic residues, free exposed sulfhydryl groups, or partially glucose-trimmed oligosaccharides (23, 24), features that are not present in the sequence 19 to 39 (see Fig. 1B). In particular, we can exclude a direct binding of BiP to this sequence as it has been shown that BiP preferentially binds a heptameric consensus motif containing a subset of aromatic and hydrophobic residues in alternating positions (25, 26), a motif not present in the sequence 19 to 39. Similarly, binding of calnexin can also be ruled out, as this chaperone is a lectin that recognizes specifically partially trimmed, monoglucosylated N-linked oligosaccharides (27-29). Involvement of the protein disulfide isomerase can also be excluded, as this chaperone requires hydrophobic interactions and formation of disulfide bonds (30). Alternatively, the sequence 19 to 39 may promote the binding of TSP to P22 in an indirect manner. This sequence would slow the folding of P22 and thus would indirectly promote the binding of TSP to a folding intermediate, allowing proper folding to proceed.

Whatever TSP binds to the sequence 19 to 39 or to another region of P22, abolishment of the N-glycosylation could reduce its chaperone activity as envisaged above. Another possibility is that treatment of cells with tunicamycin for 7 h, which is known to increase significantly the level of ER chaperones, leads to an increased level of TSP, which could consequently slow the export of P22 from the ER as demonstrated with GRP78/BiP for the secretion of different proteins (31-33).

Alternatively, TSP could be a "cargo receptor." Transported (or cargo) proteins are carried from one compartment to the next by small coated vesicles (34) and packaged into these vesicles by nonselective diffusion, a default pathway termed "bulk-flow" (35). Recent data have shown that, in contrast, some cargo proteins are specifically concentrated into vesicles, in particular at the exit from the ER (36-38). In this selective-transport model, sorting of soluble proteins would be mediated by specialized transmembrane cargo receptors. It is tempting to speculate that TSP would play such a role, mediating the packaging of P22 at one step of its vesicular transport, therefore increasing HBe secretion. So far, none of our data either support or argue against this hypothesis.

The role of HBe in the viral life cycle is still unknown (39 for review). However, recent data have shown that P22 down-regulates the level of HBV replication (40-43) by means of cytosolic P22 molecules (6, 42, 43) that most likely associate with core proteins in unstable capsids (43). Thus, one might speculate that the relative levels of P22 returned to the cytosol and P22 transported to the cell surface should be controlled and that TSP could be an actor in this repartition. Whether TSP intervenes in vivo during natural HBV infection remains to be determined.