Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Apr 15;100(9):5199–5204. doi: 10.1073/pnas.0930813100

Chaperone action in the posttranslational topological reorientation of the hepatitis B virus large envelope protein: Implications for translocational regulation

Carsten Lambert 1, Reinhild Prange 1,*
PMCID: PMC154322  PMID: 12697898

Abstract

The large L envelope protein of the hepatitis B virus utilizes a new folding pathway to acquire a dual transmembrane topology in the endoplasmic reticulum (ER). The process involves cotranslational membrane integration and subsequent posttranslational translocation of its preS subdomain into the ER. Here, we demonstrate that the conformational and functional heterogeneity of L depends on the action of molecular chaperones. Using coimmunoprecipitation, we observed specific interactions between L and the cytosolic Hsc70, in conjunction with Hsp40, and between L and the ER-resident BiP in mammalian cells. Complex formation between L and Hsc70 was abolished when preS translocation was artificially switched to a cotranslational mode, implicating Hsc70 to act as a preS holding and folding catalyst that controls partial preS posttranslocation. The functional role of Hsc70 in L topogenesis was confirmed through modulation of its in vivo activity by overexpressing its co-chaperones Hip and Bag-1. Overexpression of the Hsc70-stimulating molecule Hip led to increased entrapping of preS on the cytosolic ER face and hence to a decrease in preS posttranslocation, whereas the negative regulator Bag-1 had the opposite effects. Furthermore, Hip-mediated Hsc70 activation impaired virus production in hepatitis B virus-replicating hepatoma cells, likely due to the improper topological reorientation of L. Together, these results indicate that translocational regulation of protein topology by chaperones provides a means of generating structural and functional diversity. They also hint to the dynamic nature of the mammalian ER translocation machinery in handling co- and posttranslational substrates.

It is generally thought that all copies of a given membrane protein exist in a single orientation with respect to the membrane. However, certain proteins have been found to be expressed in two or more topological isoforms, with the heterogeneity apparently generated at the time of translocation at the endoplasmic reticulum (ER) membrane (1, 2). In most of these cases, the topological diversity results in protein multifunctionality, suggesting regulatory mechanisms controlling translocational variations at the ER (3–5). One example of such a protein is the large L envelope protein of the hepatitis B virus (HBV), a polytopic membrane protein existing in a mixed topology (6–8).

On biogenesis, the HBV L protein, together with the structurally closely related middle M and small S envelope proteins, is expressed from a single ORF of the viral genome by differential translation initiation. As a consequence, the entire sequence of S is repeated at the C termini of M and L, which contain the additional preS2 domain or preS2 and preS1 domains, respectively (9). All three proteins are cotranslationally integrated into the ER membrane by the topogenic signals of the S region that also direct cotranslational translocation of the upstream preS2 region of M into the ER lumen (10, 11). In contrast, the preS2 plus preS1 (preS) domain of L fails to be translocated and initially remains cytosolic. During maturation, about half of the L molecules posttranslationally translocate their preS region into the ER, thereby generating a dual topology that is maintained in the secreted virion envelope (refs. 6–8 and 12; Fig. 1). By orientating the preS domain both at the cytosolic (i.e., inside the virus) or luminal (i.e., outside the virus) location, L seems to serve its topological conflicting functions in the viral life cycle, like performing a matrix-like function in nucleocapsid envelopment and mediating receptor binding during host cell attachment, respectively (13, 14).

Figure 1.

Figure 1

Open in a new tab

Domain structure and transmembrane topology of the HBV L envelope protein. (Upper) Schematic representation of L consisting of the preS1, preS2, and S domains. Numbers above the domains refer to the corresponding amino acids positions. The usage of the second and third start codons located at positions 109 and 164 of L leads to synthesis of the M and S proteins, respectively. Partial N-glycosylation occurring within the S domain is indicated by ¥. (Lower) Mixed topology of L at the ER membrane. On cotranslational membrane integration, the preS domain of L is initially located on the cytosolic surface of the ER (left). During maturation, ≈50% of the L molecules posttranslationally translocate preS into the ER (right).

We recently demonstrated that posttranslational preS translocation of L is not established by an HBV-specific channel generated during virion envelope assembly, but that it is physically linked to the ER membrane (12). Posttranslational protein translocation into the ER is found in all eukaryotic cells (15, 16), but so far this pathway has been best studied in yeast where it involves the heptameric Sec complex plus cytosolic and luminal chaperones of the heat shock protein (Hsp) Hsp70 ATPase family (16–19). Cytosolic Hsp70s, such as the cognate member Hsc70, in cooperation with Hsp40 class molecules, stimulate the import of proteins into the ER (17, 18, 20) whereas the luminal Hsp70, the BiP protein (in yeast: Kar2p) is required to provide the driving force for the vectorial movement of polypeptides into the ER lumen (16, 19, 21). During our search for cellular factors potentially involved in regulating L topogenesis, we previously identified Hsc70 as a specific binding partner of the preS1 domain in vitro (22). Moreover, deletion of the Hsc70-binding site led to cotranslational preS translocation and a uniform luminal topology of L (22), suggesting that L may opt for Hsc70 as a holding and folding device to achieve more than one topological form. In this study, we examined the putative role of Hsc70 and those of the Hsp40 and BiP chaperones in guiding posttranslational preS translocation. To provide evidence for a functional interaction between L and Hsc70 in living mammalian cells, we used a recently reported approach by which the in vivo Hsp/Hsc70 activity is manipulated by overexpression of its positive or negative modulators, the Hip or Bag-1 co-chaperones, respectively (23).

Materials and Methods

Plasmid Construction.

The mammalian expression vector carrying the HBV L gene (pNI2.L) with a C-terminally tagged influenza virus hemagglutinin (HA) epitope under the control of the human metallothionein IIA promoter has been described (8, 12). To abolish concomitant expression of the HBV M and S proteins from the L ORF, their translational start codons had been inactivated (pNI2.Lo) as described (24). For induction of a uniform luminal preS topology, mutant interleukin 9 (Ile-9)∷L was used, which carries the signal sequence from Ile-9 fused to the N terminus of L (25).

Vectors for Hip (pCDNA/HA.Hip) or Bag-1 (pCDNA/HA.Bag-1), which contain, respectively, either the Hip or Bag-1 gene with an N-terminal HA tag preceded by the cytomegalovirus promoter, were kindly provided H. H. Kampinga (University of Groningen; ref. 23). To construct bicistronic plasmids encoding L plus Hip (pCDNA/HA.Hip-L) or L plus Bag-1 (pCDNA/HA.Bag-1-L), the Lo-specific transcriptional cassette was cut from pNI2.Lo by using ScaI and HpaI and inserted into the SmaI sites of the pCDNA/HA.Hip or pCDNA/HA.Bag-1 vectors, respectively.

Cell Culture, Transfection, and Coimmunoprecipitation.

For expression of the HBV proteins, transient transfection of COS-7 cells by electroporation was used. Cells were lysed under nondenaturing conditions with lysis buffer [PBS, pH 7.5/0.1% Triton X-100/1× protease inhibitor mixture (Roche Molecular Biochemicals)] for 15 min on ice. Lysates were homogenized by passing twice through a 26-gauge needle and centrifuged for 5 min at 16,000 × g and 4°C. The supernatants were then subjected to immunoprecipitation by using tosyl-activated, superparamagnetic polystyrene beads (Dynabeads M-450; Dynal, Great Neck, NY) that had been coated with an anti-Hsc70 monoclonal rat antibody (SPA-815; StressGen Biotechnologies, Victoria, Canada), an anti-BiP monoclonal mouse antibody (SPA-827, StressGen Biotechnologies), or an L-specific rabbit polyclonal antiserum (12). Therefore, 107 beads were washed with buffer A (20 mM NaH2PO4/80 mM Na2HPO4), collected by exposure to a magnetic field, and incubated with 5 μg of specific or nonspecific antibodies for 20 h at 37°C with slow rotation. Beads were washed twice with buffer B (PBS, pH 7.5/0.1% BSA) for 5 min at 4°C, once with buffer C (0.2 M Tris, pH 8.5/0.1% BSA) for 20 h at 20°C, and finally once with buffer B for 5 min at 4°C, before incubation with one half of the lysates for 2 h at 4°C. After washing of the beads twice with lysis buffer, they were suspended in sample buffer and analyzed by SDS/PAGE. Subsequent Western blotting with an HA-specific mouse monoclonal antibody was performed as described (12); the Hsc70- or BiP-specific antibodies were used in 1:5,000 or 1:1,000 dilutions, respectively, whereas the Hsp40-specific antiserum (SPA-400, StressGen Biotechnologies) was diluted 1:5,000. The antibody specific for protein disulfide isomerase (SPA-891, StressGen Biotechnologies) was applied in 1:2,000 dilutions. After incubation with peroxidase-labeled secondary antibodies, blots were developed with enhanced chemiluminescence detection reagents (Amersham Pharmacia Biosciences) and, if indicated, stripped with 0.5 M sodium hydroxide before a second immune reaction.

Trypsin Protection Assay.

Two days after transfection, microsomes were prepared from homogenized COS-7 cells and subjected to a trypsin protection assay essentially as described (8, 12). Briefly, microsomes were proteolyzed with trypsin in the presence or absence of 0.5% Nonidet P-40 (NP-40) for 60 min on ice. After inactivation of trypsin, solubilized samples were precipitated with trichloroacetic acid and subjected to SDS/PAGE and HA-specific immunoblotting. For quantification, scanned Western blots were evaluated by using Kodak 1d image analysis software.

Metabolic Labeling and Immunoprecipitation.

Transfected cells were starved in methionine-free MEM without FCS before pulse-labeling with 300 μCi (1 Ci = 37 GBq) of [35S]methionine/cysteine (Perkin–Elmer) for 10 min. Cells were harvested either directly or after a chase of 15 min, performed with DMEM containing 10% FCS and 1.5 mg/ml unlabeled methionine and cysteine (8, 12). Microsomes were prepared and subjected to trypsin digestion as above, except that proteins were immunoprecipitated with an L-specific polyclonal antiserum (12). Precipitates were separated by SDS/PAGE, visualized in a Molecular Dynamics PhosphorImager and quantitated with imagequant software.

Detection of Intracellular HBV Nucleocapsids and Extracellular Virions.

For replication of HBV in the HuH-7 liver cell line, plasmid pHBV was used, which carries a 1.1-mer of the HBV DNA genome (26). To ensure Hip overexpression in pHBV-transfected cells, the pHBV.L− mutant was used, which is defective in expression of L (24) and hence can be complemented in trans with the pCDNA/HA.Hip-L bicistronic vector. Four days after calcium phosphate-mediated transfection, intracellular nucleocapsids and extracellular virions were isolated by capsid- or envelope-specific immunoprecipitations, respectively, before detection of the encapsidated viral progeny DNA by radioactive labeling of the partially double-stranded genome with 10 μCi of [α-32P]dATP (Amersham Pharmacia Biosciences) by the endogeneous polymerase as described previously (24). After extraction of the labeled DNA genomes from the immunoprecipitated samples, they were resolved on agarose gels and visualized by PhosphorImaging.

Results

L Interacts with Hsc70 and BiP in Vivo.

We previously identified Hsc70 as a specific binding partner of the preS1 domain of L in vitro (22) and sought to now determine whether Hsc70 is also associated with the full-length L protein in vivo. COS-7 cells were transiently transfected with an HA-tagged version of the L gene. Lysates, prepared under nondenaturing conditions, were subjected to (co)immunoprecipitation with magnetic beads covalently coated with anti-Hsc70 antibodies, and the immune complexes were analyzed by HA-specific immunoblotting. Untreated lysates contained L in its characteristic doublet of a 39-kDa nonglycosylated (p39) and a 42-kDa single-glycosylated (gp42) form as a consequence of partial N-glycosylation in its S domain (Fig. 2A, lane 1). In addition, nonglycosylated and glycosylated species of both, the S (p24/gp27) and M (p30/gp33) envelope proteins were obtained due to the internal initiation of translation (Fig. 2A, lane 1). N-linked glycosylation of the slower migrating forms had been confirmed previously by treatment with PNGase F (22). Importantly, both forms of L were efficiently coimmunoprecipitated with Hsc70 (Fig. 2A, lane 2), indicating a specific interaction between L and Hsc70 in living cells. The specificity of this interaction was ascertained by the inability of the related S and M proteins to coprecipitate with Hsc70. Only after heavy overexposure of the blot, small traces of M chains but no S chains could be detected (data not shown).

Figure 2.

Figure 2

Open in a new tab

L forms complexes with Hsc70 and BiP in vivo. (A) COS-7 cells were transfected with the HA-tagged L gene, and synthesis of the nonglycosylated (p39) and single-glycosylated (gp42) forms of L was verified by SDS/PAGE of the lysate and HA-specific immunoblotting (IB). Due to internal translational initiation, the M and S proteins were obtained in addition. For immunoprecipitation (IP) analysis, lysates were reacted with antibodies specific for Hsc70 or BiP before HA-specific IB of the precipitates. As a control, nonimmune mouse IgG was used for IP. (B) To evaluate the amounts of Hsc70 and BiP coprecipitating L, the blot was rehybridized with a mixture of Hsc70- and BiP-specific antibodies. Numbers to the right refer to positions of molecular mass standards in kDa.

Besides Hsc70, posttranslational protein translocation into the yeast ER involves the function of luminal Kar2p, the yeast homologue of the mammalian BiP chaperone (16, 19). Accordingly, we tested whether L might also interact with BiP. Indeed, complex formation between L and BiP was readily observed by (co)immunoprecipitation (Fig. 2A, lane 3). Again, the S and M proteins did not associate with this chaperone, indicating that the L-specific preS1 domain is the candidate binding domain for BiP.

As evident from Fig. 2A, the amount of L precipitated by BiP was higher than that of the Hsc70-driven coprecipitation, which might reflect differences in the quantities of BiP and Hsc70 precipitated by their antibodies and/or in the degree of complex formation between L and either chaperone. To address this issue, the same blot was reprobed with anti-Hsc70 and BiP antibodies. With this analysis, roughly equal amounts of the 73-kDa Hsc70 and the 78-kDa BiP proteins present in the crude lysate were detected, whereas the amount of immunoprecipitated BiP was lower than that of Hsc70 (Fig. 2B). In combining these data, BiP binding to L was more efficient than binding of Hsc70 to L. Unexpectedly, the immune capture of BiP in L-transfected cells coprecipitated not only L, but also Hsc70 (Fig. 2B, lane 3), indicating that a subset of L chains simultaneously interacted with both cytosolic Hsc70 and luminal BiP.

Hsp40 Participates in Hsc70 Binding to L.

To productively guide translocating proteins across membranes, Hsc70 must interact with co-chaperones, such as with the human DnaJ family protein Hsp40, that regulate its ability to bind and release substrate polypeptides (20, 27). Therefore, we inspected the L-Hsc70 complex for the presence of Hsp40 (Hdj-1). The Hsc70-driven coprecipitation of L was done as above and confirmed by L-specific Western blotting (Fig. 3). The immunoblot was then stripped and reprobed with Hsc70- and Hsp40-specific antibodies. This analysis led to clear identification of Hsp40 as part of the L-Hsc70 complex (Fig. 3).

Figure 3.

Figure 3

Open in a new tab

L forms a ternary complex with Hsc70 and Hsp40. (Upper) Lysates of L-transfected cells were subjected to SDS/PAGE and HA-specific IB either directly (lane 1) or after Hsc70-driven IP, as described for Fig. 2A. (Lower) After removal of the antibodies from the blot, it was reprobed with Hsc70- plus Hsp40-specific antibodies. Positions of L, Hsc70, and Hsp40 and migration of molecular mass standards (in kDa) are denoted.

Enforced Cotranslational preS Translocation Blocks Hsc70 Binding to L.

To probe the functional significance of the observed interactions, we examined the Hsc70 binding properties of an L mutant that lacked the ability to retain its cytosolic orientation. A uniform luminal preS topology of L can be achieved by fusion of foreign signal sequences to the N terminus of L, e.g., that of Ile-9 (13, 25). Concomitantly, this enforced cotranslational preS translocation leads to de novo N-glycosylation at the two glycan acceptor sites within preS (25), as shown by the expression pattern of the Ile-9∷L mutant, which appeared in non-, single-, double-, and triple-glycosylated forms (p, gp, ggp, and gggp, respectively; Fig. 4, lane 2). The uniform luminal preS orientation was evidenced further by protease protection experiments. Unlike wild-type L (see below), the microsome-associated Ile-9∷L forms were almost fully protected from exogenously added trypsin (data not shown). When lysates were subjected to the coimmunoprecipitation procedure, the Ile-9∷L mutant failed to interact with Hsc70 (Fig. 4, lane 4). By contrast, the artificially induced cotranslational preS translocation did not affect the binding of Ile-9∷L to BiP (Fig. 4, lane 6). Quantitative evaluation of these data revealed that 84% of Ile-9∷L molecules relative to L were coprecipitated by BiP compared with 12% of Ile-9∷L chains brought down by Hsc70. The slight Ile-9∷L-Hsc70 interaction was accounted to be likely due to nonspecific binding occurring during the cell lysis and immunoprecipitation reaction. Therefore, these results suggested that Hsc70 is involved in posttranslational preS translocation.

Figure 4.

Figure 4

Open in a new tab

Enforced cotranslational preS translocation of L affects its association with Hsc70 but not with BiP. Synthesis of the HA-tagged L and Ile-9∷L mutant proteins in transfected COS-7 cells is shown by HA-specific IB. Unlike wild-type L, the Ile-9∷L mutant appeared in non-, single,- double-, and triple-glycosylated forms (p, gp, ggp, and gggp) due to cotranslational N-glycosylation of its preS domain. To probe for an association with Hsc70 or BiP, coimmunoprecipitation (IP) analysis of lysates was done as for Fig. 2A.

Modulation of in Vivo Hsc70 Activity by Hip and Bag-1 Affects Posttranslational preS Translocation of L.

To corroborate these findings, a gain-of-function analysis was performed in which the in vivo activity of Hsc70 was manipulated through overexpression of its co-chaperones Hip and Bag-1 (23, 28). Hip is a positive modulator that stabilizes the ADP-state of Hsc70 and thus the Hsc70-substrate complex (29), whereas Bag-1 negatively modulates the Hsc70 function by stimulating nucleotide exchange and substrate release (28). As a prerequisite to such an approach, L must be synthesized together with the co-chaperones within one cell. Accordingly, we constructed two bicistronic vectors containing the L gene and either the Hip- or Bag-1-specific expression units. On transient expression, the partial preS posttranslocation across membranes can be scored by trypsin protection assays of microsomes. Whereas the preS domain of newly synthesized L chains is almost fully sensitive to cleavage with trypsin, over time it becomes increasingly protected due to the preS posttranslocation into the ER, yielding up to 50–60% resistant chains at steady state (6, 8, 12). The results in steady state are shown in Fig. 5A (lanes 1–3). Treatment with trypsin in the absence of detergents (NP-40) yielded two fractions of L: trypsin-resistant full-length molecules with preS domains posttranslationally translocated into the ER, and trypsin-sensitive chains with preS domains orientated to the cytosol where they are cleaved to a 25-kDa nonglycosylated (T) and a 28-kDa single-glycosylated (gT) fragment. On disruption of microsomes with NP-40, trypsin completely converted L to these fragments. To facilitate the interpretation of results, this assay was performed within the so-called Lo-(only)-background in which synthesis of the S and M proteins is prevented by missense mutations of their initiator codons (24). As we have shown previously, the topological features of Lo are virtually identical to wild-type L (12). To prove the polarity and integrity of the microsomal membranes, the same blot was reprobed with an antibody specific to the ER-resident protein disulfide isomerase that was fully protected from proteolysis in the absence of NP-40 (Fig. 5B, lanes 1–3).

Figure 5.

Figure 5

Open in a new tab

Overexpression of Hip and Bag-1 affects preS posttranslocation. (A) COS-7 cells were transfected with vectors containing the HA-tagged Lo gene either alone (Left) or together with the HA-tagged Hip (Center) or Bag-1 (Right) genes. Two days after transfection, microsomes were prepared and either left untreated or digested with trypsin in the absence (−) or presence (+) of NP-40, and samples were analyzed by HA-specific (lanes 1–9) or L-specific (lanes 10–12) IB. The two L forms and its nonglycosylated (T) and single-glycosylated (gT) tryptic fragments are indicated on the left. Hip (H) and Bag-1 (B) are marked on the left of the corresponding panels, whereas numbers to the right refer to molecular mass standards in kDa. (B) The integrity of microsomes was analyzed by staining the blot shown in A Left with antibodies specific for protein disulfide isomerase. (C) Analysis of the Hip-effect on preS posttranslocation by using trypsin and pulse–chase labeling. Cells transfected with Lo or Lo + Hip constructs were pulse-labeled for 10 min and lysed immediately or chased for 15 min. Microsomes were mock-treated or digested with trypsin in the absence or presence of NP-40. As a control for unspecific IP, microsomes of nontransfected (NT) pulse–chase labeled cells are shown in lane 13.

To assess whether an in vivo manipulation of Hsc70 by overexpressing its co-chaperones had an effect on preS posttranslocation, trypsin protection assays were performed with microsomes from Lo + Hip or Lo + Bag-1 cotransfected cells. For simultaneous detection of the proteins, L, Hip, and Bag-1 were synthesized in HA-tagged forms. Of note, previous works had shown that neither HA-tag affected structure or function of the proteins (refs. 12, 23, and 28; and see Fig. 6). Hip and Bag-1 were obtained in single bands with molecular masses of about 50 or 39 kDa, respectively (Fig. 5A, lanes 4 and 7). On overexpression of Hip, posttranslational preS translocation was drastically inhibited, as evidenced by the decrease of trypsin protection of L in intact microsomes (Fig. 5A, lanes 4–6). Conversely, overexpressing Bag-1 that comigrated with the p39 form of L increased the amount of trypsin-resistant L chains and, by inference, the degree of preS posttranslocation (Fig. 5A, lanes 7–9). Assays were done in triplicate, and the amounts of the gp42-L forms protected from proteolysis in intact microsomes were calculated by comparing their intensities with those of the mock-treated samples. Quantification determinations revealed that 64.5 ± 5.6% of L molecules were protected in unmanipulated cells. By contrast, overexpression of Hip yielded only 28 ± 6% protection whereas in overexpressing Bag-1 cells 71.2 ± 10.4% of L molecules resisted protease cleavage. A complication to the interpretation of the results from Bag-1-overexpressing cells was the comigration of Bag-1 and p39-L. To avoid simultaneous detection of the two HA-tagged proteins, the stripped blot was rehybridized with an L-specific antibody (MA18/7) thereby providing that the amount of protected p39 was increased identical to that of gp42 (Fig. 5A, lanes 10–12).

Figure 6.

Figure 6

Open in a new tab

Overexpression of Hip reduces HBV production and increases Hsc70-L complex formation. (A) Schematic presentation of pHBV.L− containing a cloned L-negative HBV genome (black bar) preceded by the human metallothionein IIA promoter. The ORFs for the viral polymerase, core, envelope, and X proteins are shown below in open boxes, and the missense mutation of the L-specific start codon is indicated by L−. (B) For the complementation assay, HuH-7 cells were (co)transfected with pHBV.L− and vectors encoding HA-tagged L alone (L) or together with Hip (L + Hip). Secretion of enveloped virions in the culture medium was detected by IP and radioactive labeling of the viral genome by the endogeneous viral polymerase. The migration of the genome, as visualized by agarose electrophoresis and phosphorimaging, is indicated (Left). Nonenveloped cytosolic nucleocapsids were immunoprecipitated from cell lysates and processed as above (Right). (C) COS-7 cells were transfected with vectors carrying the HA-tagged L gene either alone (L) or together with Hip (L + Hip). For coimmunoprecipitation (IP), lysates were reacted with an L-specific antiserum before IB by using HA- (lanes 1–2) or Hsc70-specific (lanes 3–4) antibodies.

To unequivocally demonstrate that the observed inhibitory effect of Hip occurred during the process of posttranslational preS translocation, the kinetics of the preS reorientation were monitored by pulse–chase labeling (8, 12). As shown in Fig. 5C, after a 10-min pulse without chase, the majority of newly synthesized L chains were sensitive to trypsin independent of whether or not membranes were disrupted (lanes 1–3, 7–9; amount of protected chains: 21 ± 4.6% for Lo cells; 7 ± 1.5% for Lo + Hip cells). When pulse-labeled cells were chased for 15 min, the amount of protected L chains was estimated to be 42 ± 5.3% in the wild-type situation (Fig. 5C, lanes 4–6). By contrast, on overexpression of Hip, only 12 ± 1.7% of L molecules acquired protease resistance during the 15-min pulse (Fig. 5C, lanes 7–12). Together, the results of Fig. 5 indicated that Hip-mediated Hsc70 activation and hence the stabilization of the Hsc70-L complex substantially suppressed posttranslational preS translocation, whereas Bag-1-mediated destabilization of the Hsc70-L interaction had the opposite effect. From these data, we deduce that Hsc70 is a bona fide regulator of L topogenesis.

Overexpression of Hip Decreases HBV Production.

In a final experiment, we examined whether Hip-mediated stimulation of Hsc70 might affect the production of HBV. Virus formation is suggested to critically depend on the dual L topology because the cytosolically orientated preS domain is needed for nucleocapsid envelopment (13), whereas the preS domain disposed to the luminal (i.e., outside the virus) location is thought to mediate hepatocyte binding (14). To probe whether the decrease in preS posttranslocation by Hip-induced overactive Hsc70 had an effect on virus maturation, HuH-7 hepatoma cells were transfected with a cloned HBV genome that is defective in L synthesis (HBV.L−; Fig. 6A). For trans-complementation, either the expressing vector encoding HA-tagged L or the bicistronic L + Hip vector was used as cotransfecting plasmid. Virus production was monitored by immunoprecipitation of cellular supernatants, radioactive labeling of the partially double-stranded viral DNA genome by the viral polymerase, and detection of the genome by gel electrophoresis. As shown in Fig. 6B (Left), the HBV.L− construct was unable to support virion formation (lane 1) but could be rescued by cotransfecting the missing L expression vector (lane 2) as expected. Importantly, cotransfection of L together with Hip significantly reduced virion formation (lane 3). Because nucleocapsid assembly within the cells was not affected, regardless of the level of Hip (Fig. 6B Right, lanes 1–3), the Hip-induced block to HBV production was likely due to an increased holding of preS at the cytosolic ER face by overactive Hsc70. As a consequence, the improper topological reorientation of L presumably resulted in malfolded molecules defective in HBV assembly. Although not experimentally addressed, the Bag-1-induced Hsc70-L complex destabilization concomitant with the accelerated preS posttranslocation is expected to similarly inhibit HBV maturation, because nucleocapsid envelopment requires cytosolically exposed preS (13).

To assess whether the Hip-mediated decrease in preS posttranslocation and hence virus production coincided with an increased Hsc70 binding to L, lysates of transfected HuH-7 cells were subjected to an L-specific immunoprecipitation before Hsc70-specific immunoblotting. The level of protein expression, however, was found to be very low within these cells and as such prevented detection of Hsc70-L complex formation (data not shown). To circumvent these limitations, the coprecipitation studies were performed with COS-7 cells transfected with the L or L + Hip vectors. As evident from Fig. 6C, almost comparable amounts of L chains were precipitated from either cell (lanes 1–2). By contrast, immunostaining of the same blot with the anti-Hsc70 antibody revealed that the amount of Hsc70 captured by L was significantly increased in Hip-overexpressing cells (Fig. 6C, lanes 3–4).

Discussion

In this study, we provide evidence that the “special” biogenesis of the HBV L protein involves translocational regulation by molecular chaperones. We demonstrate that L interacts with cytosolic Hsc70 and ER-resident BiP in vivo. Consistent with our previous in vitro results (22), the preS1 domain of L was deemed to be the likely target for Hsc70 and also for BiP because both chaperones failed to recognize the HBV M and S proteins that share the primary sequence of L except for preS1. The lack of chaperone binding of M and S coincides with their classical cotranslational folding pathway and hence provides evidence for functional roles of Hsc70 and BiP in controlling the nonclassical translocation of L. In support, we observed that artificial enforcement of cotranslational preS translocation of L blocked its association with Hsc70, implicating this chaperone to be responsible for the suppression of the cotranslational preS import into the ER. Our finding that Hsp40 participates in Hsc70-L complex formation is further suggestive of a productive interaction between L and Hsc70, because members of the Hsp40 family are required for Hsc70 function (20, 30). In this respect, L shares features with the cystic fibrosis transmembrane conductance regulator, a polytopic membrane protein that also exposes large subdomains to the cytosolic side of the ER membrane where those are held and folded by the Hsp40/Hsc70 chaperone pair (31).

To examine the role of Hsc70 in L topogenesis, we exploited the recently described strategy in which Hip and Bag-1 co-chaperones are overexpressed to stimulate or suppress the activity of cytosolic Hsp/Hsc70 in living mammalian cells (23, 28). Thereby, we observed that overexpressing Hip significantly reduced posttranslational preS translocation, whereas Bag-1 overexpression enhanced this process. By contrast, altering the levels of either Hip or Bag-1 did not affect the degree of cotranslational translocation of L into the ER membrane. Hip alone has been reported to be capable of binding directly to unfolded proteins in vitro (29, 32), so decrease of preS posttranslocation might have occurred independent of Hsc70. However, we were unable to detect a direct Hip–L interaction (data not shown), and we therefore consider the observed effect of Hip and also of Bag-1 to likely be mediated via Hsc70. Our results therefore not only proved the feasibility of this co-chaperone-based strategy to assess functions of Hsp/Hsc70-substrate interactions in live cells, but also demonstrated the requirement of Hsc70 in directing the formation of the dual L topology.

The collective data led us to propose the following sequence of events accompanying L biogenesis. First, on synthesis, Hsc70 recognizes and entraps the preS1 domain emerging from cytosolic ribosomes and prevents ER import during the signal recognition particle-dependent cotranslational translocation of growing L chains. Hsc70 in cooperation with Hsp40 simultaneously may keep preS competent for posttranslocation, similar to the established functions of these chaperones in posttranslational protein import into the yeast ER and eukaryotic mitochondria (33, 34). It is noteworthy, however, that in those cases Hsc70/Hsp40 are chaperoning completed protein substrates that are not (yet) membrane-embedded (17, 18, 33, 34), whereas during the route followed by L these chaperones act in context with the ER cotranslocational machinery. Next, after termination, preS is transported across the ER membrane via a currently ill-defined mechanism that might involve putative posttranslational translocation channels of the mammalian ER. The targeting of L to such translocation systems could be directed by the bound Hsc70/Hsp40 pair that subsequently may dissociate from L, comparable to the situation in yeast where posttranslational substrates are entirely stripped off cytosolic chaperones when bound to the Sec complex (35). During the actual process of preS posttranslocation, it is tempting to speculate that the incoming preS domains are entrapped by BiP that may regulate their vectorial ER import, in accord with the BiP (Kar2p) function in yeast (16, 19, 21, 33). Finally, preS posttranslocation would stop when the equilibrium of luminal and cytosolic preS exposure is established, thereby triggering the release of two different functional conformations of L into the membrane.

Furthermore, our data indicate that translocational regulation of protein topology provides a means of generating functional diversity. Intriguingly, the translocation of the prion protein that likewise has multiple folded states related to distinct functions with one particular form being responsible for neurodegeneration has been shown to be also a subject to regulation by as yet unidentified trans-acting cellular factors at the ER (4, 5). Similarly, the topological variability of the P-glycoprotein, responsible for multidrug resistance observed in cancer cells, seems to be cotranslationally controlled by ribosomes (1, 36). Together, these observations implicate a particular role for translocational regulation in pathogenesis of disease, and suggest that dysregulation of conformational control may affect the outcome of certain disorders. In support of this view is our finding that Hsc70 has the potential to regulate HBV production, as Hsc70 stimulation by Hip overexpression impaired virion maturation in HBV-replicating cell lines, possibly because of enhanced arrest of preS on the cytosolic face of the ER, resulting in an imbalance of partial preS posttranslocation and hence in aberrant folded L molecules defective in nucleocapsid envelopment. However, we cannot exclude the possibility that Hip overexpression may produce a more general intracellular transport defect that may effect HBV release in addition.

In summary, chaperone-regulated membrane protein biogenesis seems to be an intriguing means to provide for functional heterogeneity of a single polypeptide chain. Whether such a pathway is uniquely used by HBV for benefit or is a model for a broader set of proteins remains to be determined.

Acknowledgments

We are indebted to H. H. Kampinga for generously providing expression plasmids for Hip and Bag-1 and thank K. H. Heermann for the MA18/7 monoclonal antibody. We also thank S. Bhakdi for critical reading of the manuscript, and M. Sapp and R. E. Streeck for helpful discussions. This work was supported by grants to R.P. from the Deutsche Forschungsgemeinschaft (SFB 490-A1). C.L. was supported in part by Graduiertenkolleg 194.

Abbreviations

HBV

hepatitis B virus

ER

endoplasmic reticulum

Hsp

heat shock protein

HA

hemagglutinin

Ile-9

interleukin-9

NP-40

Nonidet P-40

IB

immunoblotting

IP

immunoprecipitation

Footnotes

This paper was submitted directly (Track II) to the PNAS office.

References

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES