HCV entry is a multistep process, and there is a limited understanding of the role individual receptors play in virus internalization. We demonstrate that CD81 endocytosis occurs via a clathrin- and dynamin-dependent process in association with claudin-1. Importantly, HCV promotes CD81 and claudin-1 endocytosis, in a manner specifically inhibited by neutralizing anti-HCV immunoglobulin, supporting a direct role for these receptors in virus internalization. Our observation that anti-CD81 MAbs neutralize HCV infection after virus internalization is consistent with an intracellular site of antibody neutralization and a role for CD81 in trafficking virus to the endosomes for subsequent fusion events.
Brazzoli et al. reported that anti-CD81 MAb JS81 induced a lateral movement to cell-cell contact areas in Huh-7 hepatoma cells in a Rho-dependent manner (11
). These authors interpreted cell-cell contacts to represent tight junctions and suggested that HCV engagement of CD81 would prime the movement of virus to tight-junction-associated entry factors claudin-1 and occludin in a manner similar to that of coxsackie B virus engagement of decay accelerating factor (5
). We did not observe any evidence for antibody or virus promoting CD81 trafficking to cell-cell contacts; however, we confirmed that CD81 endocytosis was Rho dependent. Our data showed an enrichment of CD81 at hepatoma cell-cell contacts in untreated cells, as reported for other tetraspanin proteins (72
). In agreement with our data, Coller et al. reported HCV endocytosis at diverse sites along the plasma membrane of Huh-7 hepatoma cells (15
). Recent reports using high-resolution single-particle tracking demonstrated the dynamic movement of tetraspanins in live cells (36
); using this technique we have demonstrated the mobile nature of CD81 in Huh-7 cells with no evidence of antibody-dependent directed movement to cell-cell contacts.
HCV entry is clathrin dependent (48
), and we demonstrate here that CD81 can endocytose via a clathrin- and dynamin-dependent process. Dynamin is intrinsically linked to clathrin- and caveolin-dependent and -independent routes of endocytosis (28
). The reported absence of Cav-1 in Huh-7 cells and the inhibitory effect of the Eps15 transdominant mutant (14
) suggest a role for dynamin in clathrin-dependent CD81 and HCV endocytosis. Unlike tetraspanins CD63, CD151, and CD82, which have been reported to internalize via an endocytic motif YXXØ in their C-terminal domains (10
), CD81 does not contain this motif, and a mutant protein lacking a C-terminal cytoplasmic tail can endocytose and support HCV entry (6
). These data suggest a role for CD81-associating proteins in regulating its endocytosis. EGF-R was recently reported to promote HCV entry, and we examined a role for EGF-R in CD81 endocytosis (43
). We failed to observe any effect of EGF-R activation or inhibition on CD81 endocytosis, suggesting an alternative mechanism for EGF enhancement of HCV infectivity. Many viruses have been shown to utilize integrins during cell entry (for a review, see reference 61
), integrins associate with tetraspanins, and CD81 has been reported to directly associate with α4β1 (60
). CD81 may utilize its association with integrins or other known endocytic tetraspanin partner proteins CD63, CD151, or CD82 to internalize. Further work is required to understand the mechanism of CD81 endocytosis in the absence of an endocytic motif.
Silencing claudin-1 expression had no detectable effect on antibody-primed CD81 endocytosis, demonstrating that claudin-1 expression alone does not regulate CD81 endocytosis. Huh-7.5 cells express low levels of receptor active claudin-6 or 9 molecules that associate with CD81 (25
) and may provide alternative partner proteins to regulate CD81 endocytosis. Attempts to visualize HCV in association with CD81 or claudin-1, either at the plasma membrane following binding or within intracellular vesicles using antibodies to the structural proteins, were unsuccessful, due in part to the relatively low titer of HCV particles available for these studies. However, it is interesting that both HCVpp and HCVcc preparations stimulated greater levels of CD81 and claudin-1 endocytosis than saturating levels of receptor specific MAbs, suggesting that the virus may be more effective at priming receptor endocytosis.
Antiviral antibodies can neutralize infection by a variety of mechanisms, including the blocking of virus binding to cellular receptors or the inhibition of postentry/prefusion events within the endosome (for reviews, see references 35
). In contrast, less is known about the mechanism and site of action of neutralizing anti-receptor antibodies. Anti-CD81 MAbs have been reported to neutralize cell surface-bound HCV, suggesting a postattachment role for CD81 in the virus internalization process. Our current data suggest that anti-CD81 MAbs can neutralize virus at late times, after escape from PK, proposing intracellular sites for antibody neutralization. HCV has been reported to fuse in early endosomes and Meertens et al. noted a time delay between HCV internalization and endosome fusion that may reflect the recruitment of other cell components or further CD81 molecules to intracellular sites (48
). Antibody ligation promotes CD81 internalization, and we propose that treating recently infected cells with anti-CD81 enables the subsequent fusion of internalized antibody and HCV-containing endocytic vesicles, allowing intracellular sites for antibody neutralization. Thus, the internalization of anti-receptor antibodies and their fusion with vesicular virus-receptor complexes is qualitatively different from previous reports on the intracellular targeting of infection by virus-specific antibodies to West Nile virus, tick-borne encephalitis virus, and influenza virus A (20
Evans et al. reported a role for claudin-1 at a late step in the HCV entry process, using a similar experimental design to that shown in with a FLAG epitope-tagged claudin-1 molecule and high-affinity anti-FLAG antibodies, demonstrating t1/2
times of 72 and 18 min for anti-FLAG and anti-CD81 antibodies, respectively (21
). In contrast, a recent report by Krieger and coworkers demonstrated comparable times for HCV to escape anti-claudin-1; these studies most likely reflect the epitope specificity and affinity of the anti-receptor antibodies under evaluation, the different viral strains, and a hepatoma cell density that alters the HCV internalization rate(s) (37
). Our study shows variation in the time for HCV to escape different MAbs targeting CD81, which may be explained by differences in antibody affinity or epitope recognition, although in our study the t1/2
for JS81 was comparable to that reported by Krieger et al. (37
). Importantly, we highlight here the complexities inherent in this experimental design and its inadequacies for studying the chronology of receptor usage in HCV entry.
Early endosomes are points of convergence for internalized molecules (22
), where endocytosed material can acquire other cellular or signaling components by fusing with endocytic vesicles (31
). The localization of CD81/claudin-1 in Rab5-positive early endosomes and the reported 20-min delay between HCV internalization and fusion (48
) suggests that HCV/CD81/claudin-1-containing vesicles may converge with other intracellular or endocytosed molecules prior to the fusion of virus and endosomal membranes. Our data support a model wherein HCV infects hepatocytes by priming CD81/claudin-1 endocytosis and exploits the dynamic intracellular trafficking of receptor proteins.