HBV secretion is regulated through the activation of endocytic and autophagic compartments mediated by Rab7 stimulation

Hepatitis B virus (HBV) infection is a global health problem as infected individuals are at risk of chronic hepatitis, liver cirrhosis and hepatocellular carcinoma. It is estimated that 350–400 million people are chronically infected with HBV (McMahon, 2005), despite the availability of effective vaccines. Although nucleoside and nucleotide analogues and peginterferon are used to prevent the progression as standard therapies (Rijckborst et al., 2011), these drugs cannot eradicate HBV completely and their efficacies are limited. Thus, a better understanding of the HBV life cycle and the mechanisms by which it usurps established host cell pathways is essential to enable the identification of new antiviral targets.

The infectious HBV virion (or Dane particle) is a spherical particle that is 42 nm in diameter. The virion consists of an inner icosahedral nucleocapsid, which contains circular partially double‐stranded genomic DNA, and an outer envelope composed of cellular lipids and three HBV surface proteins [small (SHBs), medium (MHBs) and large (LHBs)]. These surface proteins form also an excess of empty subviral particles (SVPs). Although the mature nucleocapsid is known to originate in the cytosol and contains the HBV core protein (HBc) (Bruss, 2007), the mechanisms by which the nucleocapsid is assembled, enveloped within the endoplasmic reticulum (ER) membrane (Eble et al., 1987) and trafficked to the cell surface for release are poorly understood. Recently, a compartment of the late endocytic pathway, the multivesicular body (MVB), has been shown to participate in the final stages of HBV maturation and release. The HBV appears to reside transiently in the MVB and disruption of this compartment results in a decrease of released HBV (Kian Chua et al., 2006; Lambert et al., 2007; Stieler and Prange, 2014; Watanabe et al., 2007). Although the structural interactions between the HBV and this organelle remain unclear, it appears that virus buds inwards into invaginations of the MVB limiting membrane, forming intraluminal vesicles (ILVs), a process that is topologically equivalent to enveloped viral budding from the cell surface (Prange, 2012). Subsequently, the MVBs are believed to dock at the hepatocyte surface and release their intraluminal viral cargo in a process that might resemble exosomal release (Hanson and Cashikar, 2012). More recently, the autophagic machinery has been implicated in HBV replication both morphologically and functionally, as targeted knockdowns of the autophagic proteins Atg5, Atg7 or Beclin1 reduce HBV release (Li et al., 2011; Sir et al., 2010; Tian et al., 2011). A physical and dynamic relationship is known to exist between MVBs and autophagosomes as they appear to interact (Berg et al., 1998; Fader et al., 2008) prior to fusion with a terminal lysosome and exchange a variety of different components (Fader and Colombo, 2009). The mechanisms that support the interactions between these organelles are a topic of intense study, and the processes by which the HBV utilizes these organelles to its own end are almost completely undefined.

The Rab proteins are small GTPases that act as molecular switches, orchestrating the formation, transport, tethering and fusion of vesicles in the secretory and endocytic pathways. As for most switch enzymes, the GTP‐bound state of a Rab regulates its binding to a variety of distinct effector proteins that facilitate vesicular traffic (Stenmark, 2009). In humans, there are more than 60 Rabs, many with unknown functions, although Rab7 (for which there are two isoforms, Rab7a and Rab7b) has been demonstrated to play a central role in regulating endo‐lysosomal membrane traffic (Bucci et al., 2000; Cantalupo et al., 2001). Rab7 is required for the maturation of late endosomes/MVBs and also autophagosomes, directing the trafficking of cargos along microtubules and participating in the fusion step with lysosomes (Hyttinen et al., 2013; Vanlandingham and Ceresa, 2009). We hypothesized that Rab7 could play an important role in the regulation of HBV assembly through the late endocytic pathway of hepatocytes. In this study, we show that the HBV itself is capable of activating Rab7 leading to the formation of pronounced tubular networks extending from MVBs and autophagosomes. These tubules appear to promote the fusion of MVBs, autophagosomes and lysosomes and facilitate the degradation of HBV. These findings indicate that the HBV alters the endo‐lysosomal and autophagic pathways through the regulation of a specific Rab protein and provides novel insights into host–pathogen interactions.

Based on the findings of others, HBV appears to utilize the MVB (Kian Chua et al., 2006; Lambert et al., 2007; Watanabe et al., 2007) and autophagic compartments (Li et al., 2011; Sir et al., 2010; Tian et al., 2011) during assembly, maturation and release. As Rab7 is known to regulate the dynamics and traffic between these late endocytic compartments and the lysosome (Cantalupo et al., 2001; Jäger et al., 2004; Stenmark, 2009), we tested whether alterations in Rab7 expression and/or function might interfere with the viral life cycle. First, HepG2.2.15 cells, which stably express HBV, but are not re‐infectable with released virus, were transfected with GFP‐tagged wild‐type Rab7 (GFP–Rab7wt) and the relationship between the virus and the Rab7‐positive compartment was assessed by confocal microscopy. Rab7 localized along numerous large punctate late endosomes and MVBs that also contained large quantities of HBV proteins (LHBs and HBc; Fig. 1A,B). As expected, mCherry‐tagged Rab‐interacting lysosomal protein (RILP), which is a classic effector protein of active Rab7 and has been used to visualize active Rab7 in the cell (Cantalupo et al., 2001), also colocalized with these HBV proteins (supplementary material Fig. S1A). We further confirmed that HBV proteins reside within MVBs and other late endosomal structures by colocalization studies of these viral antigens in compartments positive for the GFP‐tagged autophagosome marker LC3 (also known as MAP1LC3) and GFP‐tagged LAMP1 as a lysosomal marker (Fig. 1A,B). In addition, cells were stained with antibodies against the MVB marker Hrs, the tetraspanin MVB marker CD63 or the autophagosomal protein LC3. Cells were then co‐stained with for the viral LHBs (supplementary material Fig. S1B–D). In addition, transmission electron microscopy (TEM) images of HepG2.2.15 showed numerous electron‐dense virus‐like particles of 40 nm in diameter residing within both MVBs and autophagosome compartments (Fig. 1C,D). However, these virus‐like structures were not observed in the untransfected HepG2 cells (data not shown). These particles do not look like the typical Dane particles, which are observed with negative staining. It was considered that this is due to the difference in staining technique.

Surprisingly, although some HBV‐expressing cells displayed normal spherically shaped MVBs and autophagosomes containing viral proteins (Fig. 1A,B), many possessed large and very long Rab7‐ or LC3‐positive tubules extending from these compartments, which were not found in control cells (Fig. 2A,B; supplementary material Fig. S1E,F). The tubules were also observed in HepG2 cells transfected with 1.3‐fold HBV full genome (Fig. 2C). In many instances, these tubules interconnected to form reticular networks that could also be resolved by TEM (Fig. 2D–F; supplementary material Fig. S1G–J). The viral protein (LHBs) was observed in these tubules (supplementary material Fig. S2A). Importantly, HBV infection also induced the formation of Rab7 tubules in primary human hepatocytes (supplementary material Fig. S2B–E), indicating that this dramatic response was not an aberration that occurs only in a neoplastic cell line.

To test whether other Rab‐centric endocytic pathways are altered by HBV infection, HepG2.2.15 and HepG2 cells were transfected to express GFP‐tagged Rab5 or Rab11. We speculated that changes in these compartments within the infected cells that would represent activation of these Rab proteins similar to that observed for Rab7. No changes in these compartments were observed in infected versus control cells. Because it is thought that MVB functions are associated with the HBV life cycle, we investigated whether there were changes in the distribution of endosomal sorting complexes required for transport (ESCRT) proteins. As well as staining for Hrs, GFP–Tsg101 (ESCRT1), GFP–Vps25 (ESCRT2) or RFP–CHMP6 (ESCRT3) was expressed in these cells but no differences in their distribution was observed (data not shown).

To test whether the dramatic morphological changes observed in the late endocytic compartments of HBV‐producing cells (Fig. 2) required the participation of Rab7, both control HepG2 and virus‐expressing HepG2.2.15 cells were treated with Rab7‐specific small interfering RNA (siRNA) to reduce the levels of this small GTPase. The efficiency of Rab7 depletion was good (see Fig. 5A). Indeed, we observed in TEM images very little MVB tubulation following Rab7 depletion (Fig. 3A,B). As assessed by confocal microscopy, Rab7‐depleted cells had significantly less LC3‐positive tubules extending from LC3‐positive organelles than control cells, suggesting that these membrane dynamics required functional Rab7 (Fig. 3C,D).

To test whether the observed Rab7‐dependent tubule formation was a result of a specific interaction with its effector protein RILP, which is known to bind active Rab7 and mediate an interaction with microtubule‐based motors (Jordens et al., 2001; van der Kant et al., 2013), we reduced RILP levels in HepG2.2.15 cells by siRNA‐mediated knockdown (supplementary material Fig. S2F,G). This manipulation resulted in a significant reduction in Rab7‐dependent tubule formation and induced a vesiculation of the Rab7 compartment compared to infected control cells (supplementary material Fig. S2H–J). Additionally, expression of exogenous mCherry‐tagged RILP also reduced tubule formation substantially (supplementary material Fig. S2K,L). This response was seen both upon expression of the wild‐type protein, an effect we observe often with an overexpression of tagged proteins, and was accentuated by expression of a tagged Rab7‐binding domain (RBD, amino acids 241–320). Interestingly, expression of an RILP C‐terminal deletion protein (amino acids 1–220, supplementary material Fig. S2K) restored Rab7‐based MVB tubulation in HepG2.2.15 cells substantially to levels seen in cells that had not been transfected with exogenous RILP (supplementary material Fig. S2L), suggesting that expression of a 79 amino acid fragment (amino acids 241–320) makes the protein inert and unrecognizable by the cellular machinery. Taken together, these findings are consistent with the premise that a Rab7–RILP complex mediates HBV‐activated MVB tubulation.

As the tubulation of late endocytic compartments is between three and ten times more prevalent in the virus‐expressing versus control cells (Fig. 2) and is dependent upon Rab7 expression (Fig. 3), we tested whether the HBV is responsible for its activation. To this end, Rab7 activity in HepG2.2.15 cells was compared with that in HepG2 cells using a GST–RILP pulldown assay. Surprisingly, HepG2.2.15 cells had five to eight times higher Rab7 activity than HepG2 control cells (Fig. 4A,B) although there was no difference in total Rab7 levels. Furthermore, HBV from HepG2.2.15 supernatant was added to primary human hepatocytes and also increased Rab7 activity as well as did the secreted viral protein HBsAg, which was detected by an enzyme‐linked immunosorbent assay (ELISA; Fig. 4C,D). To define the HBV component responsible for this dramatic activation, HuH7 cells, which have a higher transfection efficiency than HepG2 cells, were transfected individually with FLAG‐tagged variants of the five distinct proteins encoded by the HBV genome [polymerase, LHBs, HBc, precore, and HBVx protein (HBx)]. Interestingly, the precore protein alone was sufficient to increase Rab7 activity (by fourfold; Fig. 4E,F), although the expression level was significantly lower than that of HBc and HBx and similar to LHBs (supplementary material Fig. S2M). The precore protein is a precursor of HBeAg, and the processed form of HBeAg, in which both N‐terminus and C‐terminus are cleaved (supplementary material Fig. S2N). To further confirm the Rab7‐activating capacity of this viral antigen, the FLAG‐tagged precore protein was expressed in two additional epithelial cell models (Hep3B and HeLa). This protein significantly increased Rab7 activity in both cell lines (Fig. 4G,H), and also the processed form of HBeAg increased Rab7 activity in HuH7 cells (supplementary material Fig. S2N–P). Thus, our data suggests that precore‐ and HBe‐induced Rab7 activation promotes tubulation of late endosomes, MVBs and autophagosomes, which might help regulate virus maturation and subsequent release. The specific knockdown of the precore protein is technically difficult as an siRNA targeting the precore mRNA can also affect the expression of core protein and polymerase.

As HBV infection results in a profound remodeling of the Rab7‐associated MVB and autophagosome compartments in cells it seemed likely that these changes could affect HBV assembly, intracellular trafficking, and secretion from cells. To test this, Rab7 was depleted in HepG2.2.15 cells by siRNA and the levels of internal viral proteins were assayed by western blot analysis and the level of secreted virus was assayed by real‐time PCR. Surprisingly, we observed that Rab7 depletion induced a threefold increase in the intracellular accumulation of LHBs (Fig. 5A,B) and also significantly increased the release of HBV DNA into the supernatant (Fig. 5C). To confirm that the increase in HBV DNA released from Rab7‐depleted cells represents intact enveloped virus, HepG2.2.15 cells were transfected with FLAG–HBs and the supernatant was subjected to an immunoprecipitation with anti‐FLAG antibody and then subjected to PCR for the HBV genome. The rationale for this approach was that the tagged HBs could associate with and subsequently precipitate the viral genome for detection only if packaged into a virion. As shown in Fig. 5D, the levels of HBV DNA post‐immunoprecipitation were also increased in the supernatant from Rab7‐depleted cells, an indication that Rab7 depletion increases the secretion of enveloped HBV particles. In control experiments, the expression of FLAG‐tagged mock protein or an addition of control antibody did not increase the HBV DNA levels after immunoprecipitation. There was no significant difference in the amount of HBsAg in the supernatant from cells with or without Rab7 depletion (data not shown), suggesting that the secretion pathways of HBV particles and HBsAg subviral particles (SVPs) are different, as reported previously (Garcia et al., 2009), and Rab7 does not affect the secretion pathway of SVPs. The expression of a GTPase‐defective Rab7 mutant (Rab7T22N) also increased the amount of intracellular LHBs and released viral genome, as with the Rab7 knockdown (Fig. 5E,F). Additionally, the expression of mCherry–RILP reduced the secretion of HBV DNA in the supernatant but the effect was not seen after Rab7 knockdown (supplementary material Fig. S3A,B).

To further define the role of Rab7 function in the HBV life cycle, ‘rescue’ experiments were performed in which GFP–Rab7wt or GFP–Rab7T22N were re‐expressed in the Rab7‐depleted cells. As expected, we observed that GFP–Rab7wt re‐expression, but not that of GFP–Rab7T22N, restored the LHBs level (supplementary material Fig. S3C,D) as well as the colocalization of LHBs and LAMP1 (supplementary material Fig. S3E,F). In addition, the reduced number of LC3‐positive tubules seen after Rab7 depletion were restored by the expression of FLAG–Rab7wt (supplementary material Fig. S3G–I). These data provide additional support for the concept that the effects of the Rab7 depletion on HBV production and localization were not due to off‐target effects. More importantly, the viruses secreted from the Rab7‐depleted cells appear to represent infectious particles because HuS‐E/2 cells, which are susceptible to HBV infection (Huang et al., 2012), show equal infectivity in response to virus collected from either knockdown or control cells (supplementary material Fig. S4A–C). A successful infection was confirmed by the gradual increase of HBV DNA in the supernatant from HuS‐E/2 cells (supplementary material Fig. S4A).

From the findings described above, we hypothesized that the effect of decreased Rab7 levels leading to an increase in both internal and secreted virus could be a result of altered trafficking to the degradative lysosome. To test whether this transport occurs in live cells, we analyzed HepG2.2.15 cells expressing GFP–LC3 and mCherry–LAMP1 by confocal microscopy. As depicted in Fig. 6A, time‐lapse images revealed that LAMP1 puncta were tethered to and then fused with LC3‐positive tubules or, alternatively, were pulled towards an autophagosome. Therefore, it was considered that the Rab7‐dependent tubules induced by HBV might facilitate the lysosomal degradation of virus in MVBs or autophagosomes. Accordingly, we tested for changes in the distribution of viral proteins and lysosomes in virus‐producing cells in the context of Rab7 depletion. Knockdown of Rab7 in these cells induced a significant (70%) decrease in the localization of LHBs with LAMP1 compartments (Fig. 6B,C), suggesting that Rab7 facilitates the transport to and subsequent degradation of HBV in lysosomes. In addition and consistent with previous observations (Vanlandingham and Ceresa, 2009), significantly more MVBs were found in Rab7‐depleted cells when assessed by TEM (Fig. 6D–F), suggesting an induced defect in MVB–lysosome fusion. In support of this premise, we observed enlarged MVBs in Rab7‐depleted cells that were not seen in control cells (supplementary material Fig. S4D,E).

To investigate the effect of HBV on endocytic uptake and trafficking of non‐viral endogenous receptors, we compared the kinetics of the epidermal growth factor receptor (EGFR) following stimulation with EGF ligand in HepG2 versus the infected HepG2.2.15 cells as described previously (Schroeder et al., 2010). Although HepG2.2.15 cells exhibited markedly higher levels of EGFR to begin with, the kinetics of degradation up to 6 h appeared very similar. Furthermore, we observed little difference in the uptake and trafficking of rhodamine‐tagged transferrin (data not shown).

Because Rab7 appears to be mediating the transfer of nascent HBV from MVBs to the lysosome for potential degradation, we tested whether treating cells with CQ, which disrupts lysosome function by preventing acidification, might also increase viral secretion and therefore mimic the effects of the Rab7 knockdown. HepG2.2.15 cells were treated with 0–50 µM CQ for 24 h and the levels of LHBs in the cells and HBV DNA in the supernatant were assayed. As predicted, both were increased significantly after CQ treatment in a dose‐dependent manner (Fig. 7A–C). In addition, another lysosome inhibitor NH₄Cl produced similar results in two independent experiments (supplementary material Fig. S4F,G). This indicates that a substantial percentage of HBV is degraded in lysosomes under normal conditions. When the virus‐containing cell supernatant was collected from the drug‐treated cells and added to HuS‐E/2 cells (the CQ concentration was diluted 300 times beyond its working concentration), we did not see a significant difference in the infectivity compared to control cells (data not shown), similar to our observations using virus‐containing supernatant from Rab7‐depleted cells (supplementary material Fig. S4A–C). Most importantly, the CQ‐treated cells no longer exhibited a Rab7‐depletion effect on the intracellular levels of LHBs (Fig. 7D,E), supporting the concept that Rab7 mediates the transport of HBV to the lysosome to reduce viral production and secretion.

In this study, we provide novel findings implicating the small GTPase Rab7 as a central regulator of HBV transport and secretion. We find that assembled virus accumulates in Rab7‐positive MVB and autophagosome compartments (Fig. 1), which then become markedly altered and tubulated upon infection of hepatocytes (Fig. 2). Equally surprising was the finding that HBV significantly enhances Rab7 activity through the expression of the precore or HBe protein encoded by its own invasive genome (Fig. 4). In this context, Rab7‐dependent activation of the late endosomal pathway (Fig. 3) appears to facilitate the interaction with, and potentially exchange of contents between (Fig. 6), the MVB, autophagosome and lysosome (Figs 6,7), leading to an increase in HBV degradation and a subsequent attenuation of viral shedding (Fig. 5). Fig. 8 illustrates our main concept, demonstrating that the HBV infection promotes the formation of an interconnecting reticular network between MVBs, autophagosomes and lysosomes driven by the activation of Rab7. Although tubules extending from MVBs and autophagosomes have been reported previously (Cooney et al., 2002; Cortese et al., 2013; Gao et al., 2010), those induced by HBV in the current study are much more prominent.

Rab7 is a key trafficking regulator involved in multiple post‐endocytic processes including transport between late endosomes/MVBs, autophagosomes, lysosomes and other lysosome‐related organelles (Wang et al., 2011). Activated Rab7 on the endosomal membrane binds to RILP, a well‐characterized effector of Rab7, together forming a complex with p150^Glued (also known as DCTN1) and dynein–dynactin to drive the transport of endosomes to the minus‐end of microtubules (Jordens et al., 2001). Additionally, Rab7 is known to associate with the FYVE coiled coil domain protein (FYCO1) and the molecular motor kinesin to participate in the microtubule plus‐end‐directed transport of endosomes (Pankiv et al., 2010). The activation of this complex might promote the formation of tubular extensions bidirectionally along microtubules, as discussed previously (Harrison et al., 2003), and, as observed by our electron micrographs (Fig. 3A; supplementary material Fig. S1J). Although many tubules were found to extend from MVBs and autophagosomes in the HBV‐producing cells (Fig. 2), we observed no alterations in the lysosomal structure (data not shown), suggesting that Rab7 specifically acted in the MVB and autophagosome compartments.

Currently, the mechanism by which the precore or HBe protein activates Rab7 is unclear as we have been unable to detect any interactions between the viral protein and Rab7 biochemically (data not shown). It is possible, if not likely, that the HBe protein could interact with either guanine‐nucleotide‐exchange factors (GEFs) or GTPase‐activating proteins (GAPs) specific for Rab7 (Wang et al., 2011). Interaction with these regulatory proteins could activate Rab7 by enhancing the function of a GEF, or in contrast, inhibiting the function of a GAP. Studies to test for interactions between HBe and these accessory factors are ongoing.

A central finding of this study indicates that the HBV itself activates Rab7 to promote exchange between MVBs, autophagosomes and the lysosome (Fig. 6) resulting in reduced viral production. The marked intracellular accumulation of viral proteins accompanied by the increase in viral secretion following Rab7 knockdown (Fig. 5) or inhibition of lysosome function (Fig. 7) supports this concept. Some possibilities exist to explain how the virus would benefit from activating the pathway that would direct it into the lysosome for eventual destruction. First, processing of surface proteins by lysosomal enzymes might aid in viral ‘maturation’ and increase infectivity, or, alternatively, decrease viral recognition by neutralizing antibodies in the blood. To test these concepts, we compared the infectivity of virus released from cells with or without Rab7 depletion, as well as with or without CQ treatment; however, no differences in infectivity were observed (supplementary material Fig. S4A–C). Second, and perhaps more relevant, is the fact that HBV is considered a ‘stealth virus’ that is able to escape from the front line of host defenses (Wieland and Chisari, 2005). HBe, which is not required for viral replication, is considered to have a role in the establishment of chronic infection (Milich and Liang, 2003). It has been shown that HBe modulates innate immune signal transduction pathways through interaction with and targeting of the signaling pathways mediated by Toll‐like receptors (Lang et al., 2011). As a result of attenuated immune responses, the viral load increases greatly in the early phase of chronic infection (Liang, 2009). We hypothesize that HBV might put a brake on its own secretion through HBe‐mediated Rab7 activation to reduce the immune response. In support of this hypothesis, it has been shown that HBe‐negative strains have high replication capacity in vitro (Inoue et al., 2011; Ozasa et al., 2006) and that they can cause vigorous immune responses resulting in fulminant hepatitis (Milich and Liang, 2003).

An alternative explanation as to why we observed this Rab7 activation is that the activation of a Rab7‐mediated viral degradation pathway rather than representing a host defense mechanism – that is, hepatocytes respond to the expression of the HBe antigen by grossly activating the tubulation and fusion of MVBs and autophagosomes with the lysosome. Such membrane remodeling events could be part of an autophagy‐mediated clearance of invading pathogens (xenophagy), a well‐established cellular defense mechanism (Levine, 2005).

Finally, it is important to note that the specific role of Rab7 described here might represent just one of several functions in the HBV life cycle. A recent paper has shown that the early entry stages of HBV infection in HepaRG cells depend on both Rab5 and Rab7 (Macovei et al., 2013). The HepG2.2.15 cell model used in our current study stably expresses HBV and is not susceptible to further infection because it expresses very low levels of the putative HBV receptor, the sodium taurocholate cotransporting polypeptide (NTCP) (Yan et al., 2012). Therefore, HepG2.2.15 cells provide a useful model to study the production and release of the virus rather than infection. Thus, Rab7 activation by the HBe protein might also increase the efficiency of the early stages of infection.

It is clear from this and other studies implicating the endosomal pathways in HBV infection that a more complete understanding of how this virus usurps the vesicle trafficking machinery from the hepatocyte to suit its own ends will be a complex but rewarding challenge. Additional regulatory Rab GTPases, vesicle coat and adaptor proteins, as well as fission enzymes, are likely to participate in the HBV life cycle and thus will provide useful drug targets for future therapy.

To obtain FLAG‐tagged HBV individual protein constructs, individual DNA sequences specific for each protein were amplified from a total DNA extracted from the culture supernatant of HepG2.2.15 cells. Nucleotides [nt, the numbers are in accordance with a genotype D HBV sequence of 3182 nt from HepG2.2.15 (accession number U95551)] 2307–3182 and 1–1623, 2847–3182 and 1–835, 155–835, 1899–2453, 1814–2453, and 1374–1840 were amplified for FLAG–polymerase, FLAG–LHBs, FLAG–HBs, FLAG–HBc, FLAG–precore and FLAG–HBx, respectively. These PCR products were cloned into pcDNA3 (Invitrogen, Carlsbad, CA) modified to have a FLAG sequence upstream of the multiple‐cloning site. 1.3‐fold wild‐type HBV genome (nt 1051–3215 and 1–1953, which is 1.3-fold longer than a circular HBV genome) of genotype B, which was obtained from an acute hepatitis patient, was described previously (Inoue et al., 2011). GFP–Rab7wt was as described previously (Schroeder et al., 2012) and GFP–Rab7T22N was kindly provided by Dr Bruce Horazdovsky (Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN). FLAG–Rab7wt was made from a PCR product that was amplified from GFP–Rab7wt. GST–RILP was kindly provided by Dr Cecilia Bucci (Universita del Salento, Italy) and mCherry–RILP was provided by Dr Barbara Schroeder (Department of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, MN) and obtained by cloning the RILP sequence into the HindIII/XbaI sites of pmCherry‐C3 (Takara Bio, Shiga, Japan). GFP–LAMP1 and mCherry–LAMP1 were made as described previously (Schulze et al., 2013). siRNA targeting human Rab7, whose sequence was described previously (Jäger et al., 2004), siRNA targeting human RILP (5′‐GAUCAAGGCCAAGAUGUUAUU‐3′) and a non‐targeting (NT) control siRNA were purchased from Dharmacon (Thermo Fisher Scientific, Lafayette, CO).

Anti‐PreS1 and anti‐GST antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti‐HBc was purchased from Abcam (Cambridge, UK). Antibodies against Rab7, FLAG (monoclonal) and actin, as well as chloroquine (CQ), were purchased from Sigma‐Aldrich (St. Louis, MO). Anti‐FLAG (polyclonal) and anti‐Myc (monoclonal and polyclonal) antibodies were purchased from Cell Signaling (Danvers, MA). Anti‐Hrs antibody was purchased from Bethyl Laboratories (Montgomery, TX). Anti‐LC3 was purchased from Novus Biologicals (Littleton, CO).

The HBV‐expressing stable cell line HepG2.2.15, which was kindly provided by Dr Andrea Cuconati (Institute of Hepatitis and Virus Research, Doylestown, PA), and the parental human hepatoma cell line HepG2 were incubated in RPMI 1640 with L‐glutamine (Corning Cellgro, Manassas, VA) supplemented with 10% fetal bovine serum (FBS), 50 U/ml penicillin and 50 µg/ml streptomycin. HepG2.2.15 and HepG2 cells were seeded onto coverslips or plates coated with poly‐L‐lysine (Sigma‐Aldrich) as described previously (Abdulkarim et al., 2003). Cells were transiently transfected with Lipofectamine Plus (Invitrogen) or Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. HepG2.2.15 cells were transfected twice in a 24 h interval to increase the transfection efficiency. Transfection of HepG2.2.15 cells with siRNA was performed using RNAiMAX (Invitrogen) following the standard protocol. The re‐expression was performed as above 24 h after the siRNA transfection. HeLa and the human hepatoma cell lines HuH7 and Hep3B were incubated in MEM (Corning cellgro), supplemented with 10% FBS, 1.5 g/l sodium bicarbonate, 1× nonessential amino acids, 1 mM sodium pyrophosphate, 50 U/ml penicillin and 50 µg/ml streptomycin. An immortalized human primary hepatocyte cell line HuS‐E/2, which was kindly provided by Dr Makoto Hijikata (Kyoto University, Japan), was grown as described previously (Aly et al., 2007). HBV infection experiments with HuS‐E/2 cells were performed as described previously (Huang et al., 2012). Primary human hepatocytes (PHHs) were from Yecuris (Tualatin, OR) and maintained in DMEM with high glucose and L‐glutamine, supplemented with 10% FBS and 50 U/ml penicillin and 50 µg/ml streptomycin.

Cells were lysed in RIPA buffer [150 mM NaCl, 1% NP‐40, 0.5% deoxycholic acid, 0.1% SDS, 50 mM Tris‐HCl pH 8.0 and complete protease inhibitors (Roche Diagnostics, Indianapolis, IN)]. 30 µg of soluble proteins were resolved by SDS‐PAGE, and transferred onto a PVDF membrane. After blocking with 10% milk in PBS, the membrane was incubated with primary antibodies at room temperature for 2 h, and incubated with secondary antibodies for 1 h. The signals were detected with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific).

GST–RILP was transformed into Escherichia coli BL21 cells and 4 ml of an overnight culture was cultured further in 200 ml LB to an optical density (OD) at 600 nm of 0.6–0.8. After the addition of isopropyl β‐D‐1‐thiogalactopyranoside (IPTG, final concentration of 1 mM), it was incubated at room temperature for 3–4 h. The culture was spun down, and the collected bacteria were sonicated in PBS on ice. They were rotated at 4°C for 30 min with 1% Triton X‐100 and GST–RILP protein was purified using glutathione–Sepharose‐4B beads (Amersham‐Pharmacia Biotech, GE Healthcare Bio‐Sciences, Piscataway, NJ) according to the manufacturer's instructions. The beads and 0.25 mg of cell lysate to be analyzed were rocked in 600 µl of Bing buffer (25 mM Tris-HCI pH 7.4, 100 mM NaCI, 1 mM DTT, 1% NP-40, 2mM Na₃VO₄, 15 mM NaF, 0.1 mM EDTA and complete protease inhibitors) at 4°C for 1–2 h. The levels of active Rab7 binding to GST–RILP were quantified by western blot analysis.

Total DNA was extracted with a QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany) from 100 µl of the culture supernatant, and 2 µl of the extracted DNA solution was subjected to real‐time PCR using a LightCycler 480 system (Roche Diagnostics) with SYBR Green I Master (Roche Diagnostics), primers HBXF11 (5′‐ATGGCTGCTAGGCTGTGCTG‐3′) and HBXR4 (5′‐GTCCGCGTAAAGAGAGGTGC‐3′). The HBsAg levels in the culture supernatant were quantified with Hepatitis B Surface Antigen ELISA Kit (MyBioSource, San Diego, CA).

HepG2.2.15 cells were transfected with FLAG–HBs, and 2 days following transfection, enveloped HBV particles in the supernatant were precipitated using the anti‐FLAG antibody. The levels of HBV DNA from the precipitates were determined by real‐time quantitative PCR.

Fluorescence micrographs were acquired using a Zeiss Axio Observer.D1 microscope (Carl Zeiss, Oberkochen, Germany) with a 63× objective, an Orca II camera (Hamamatsu Photonics, Hamamatsu, Japan), and iVision software (BioVision Technologies, Exton, PA), or an LSM780 confocal microscope (Carl Zeiss) with a 63× objective and ZEN software (Carl Zeiss). Time‐lapse images of live cells were taken using the confocal microscope every 30 s on a heated chamber set at 37°C and under 5% CO2. Transmission electron microscopy (TEM) micrographs were acquired as previously described (Henley et al., 1998) with a JEOL 1200 electron microscope (JEOL Ltd, Tokyo, Japan). Images were adjusted with Adobe Photoshop software (Adobe, San Jose, CA).

Results of quantitative variables are expressed as mean±s.e.m. Statistical comparisons were made by using a Student's t‐test unless otherwise indicated and P<0.05 was considered significant.

The authors would like to thank members of the McNiven laboratory for their contributions, especially Barbara Schroeder and Ryan Schulze for critical readings of the manuscript.