Hepatitis C virus (HCV) is a member of the Flaviviridae
family. This important human pathogen specifically infects the liver. At present there is no HCV vaccine and although a number of drugs targeting HCV replicase enzymes are in development, recent trials have shown a rapid appearance of drug-resistant viruses 
. The conserved nature of HCV entry into host cells offers an alternative and attractive target for therapeutic intervention.
HCV initiates infection by attaching to the cell surface followed by clathrin-dependent internalization of virus particles; current evidence supports a role for scavenger receptor class B member I (SR-BI), tetraspanin CD81 and tight junction proteins claudin-1 and occludin in coordinating this process (reviewed in 
). SR-BI and CD81 bind HCV-encoded E1E2 glycoproteins with high affinity and have been reported to play a role in particle attachment to the cell 
. In contrast, there is limited information on whether claudin-1 or occludin interacts directly with HCV. The essential role of claudin-1 in the late stages of HCV entry 
suggests that there may be a requirement for the virus to bind receptor proteins in a defined sequence or that claudin-1 has another, as yet undetermined, function.
The claudin superfamily of four transmembrane domain (4TM) proteins oligomerize to form strands that comprise cellular tight junctions 
, thereby generating the seal required to maintain cellular homeostasis. Interactions between the first and second claudin extracellular loops (EC1 and EC2; ) allow protein associations both within the plasma membrane of a single cell and between adjacent cells (reviewed in 
); Förster resonance energy transfer (FRET) between tagged molecules suggests that protein dimers are the primary building block(s) of claudin strands 
. We 
and others 
have reported that claudin-1 associates with tetraspanin CD81; this receptor complex is present at the basolateral membrane of hepatoma cells 
and is essential for HCV entry in vitro
. Inhibiting protein kinase A 
or activating epidermal growth factor accessory protein 
limits claudin-1/CD81 complex formation and HCV entry. Furthermore, anti-claudin-1 antibodies inhibit HCV infection by reducing claudin-1 association with CD81 without perturbing tight junction integrity 
Claudin-1 forms monomers and higher order structures in yeast membranes.
Having established the biological relevance of claudin-1 interaction with CD81, our aim is to characterize these proteins in vitro
. Since primary human hepatocytes and the majority of human hepatoma cells express all four receptor proteins, and siRNA silencing approaches are frequently partial, it is difficult to study the role of claudin-1 in the HCV internalisation process in mammalian cells. We have therefore focused on purifying recombinant, full-length protein components of the HCV receptor complex; studies on CD81 have highlighted important differences in the interaction of HCV with full-length, cell-expressed forms of CD81 and recombinant truncated forms of the soluble second extracellular loop (EC2) 
. For example a soluble, mutant form of EC2 (F150S) does not interact with HCV sE2, while mutation of the same residue in full-length, cell-expressed CD81 has minimal effects on HCV sE2 binding 
. These results suggest that EC2 has a more robust structure in the full-length tetraspanin and, together with other data, indicate that regions in EC1 or the transmembrane domains may play a role in CD81 dimerisation 
We previously reported the purification of recombinant, full-length human CD81 from the methylotrophic yeast Pichia pastoris
: monomers, dimers and higher oligomers of CD81 were observed in recombinant P. pastoris
membranes comparable with endogenous protein in mammalian membranes. Immunofluorescent and flow cytometric staining of P. pastoris
protoplasts with monoclonal antibodies specific for CD81 EC2 demonstrated comparable conformation of the recombinant and native molecules. Recombinant CD81 isolated in a monodispersed form using n-octyl-β-d
-glucopyranoside (βOG), as determined by analytical ultracentrifugation (AUC), was shown to interact with HCV sE2, representing the first biophysical characterization of a functional, full-length, recombinant tetraspanin 
Here we report the production of milligram quantities of recombinant human claudin-1 using the yeast, P. pastoris. Yeast protoplasts expressing claudin-1 bound an antibody specific for a conformation-dependent epitope expressed on human hepatocytes, suggesting a native protein conformation. Claudin-1 could be isolated from yeast membranes in various oligomeric forms, dependent upon the detergent used. When isolated with βOG, it was monodispersed and dimeric, while isolation with profoldin-8 or n-decylphosphocholine (foscholine-10) resulted in dynamic mixtures of oligomers. When oligomeric preparations were reconstituted into liposomes, the resulting proteoliposomes inhibited HCV infection of hepatoma cells in a dose-dependent manner. Claudin-1-containing proteoliposomes also inhibited the infectivity of lentiviral pseudotypes expressing HCV E1E2 glycoproteins (HCVpp) and, unexpectedly, vesicular stomatitis virus expressing glycoprotein G (VSV-Gpp). In contrast, CD82-containing proteoliposomes had no effect on HCVpp or VSV-Gpp, indicating that proteoliposomes are not tractable to study the biological function of claudin-1. In the absence of a suitable biological assay, the known interaction of claudin-1 with CD81 was therefore examined in vitro.
Using dynamic light scattering (DLS), dynamic preparations of claudin-1 (in foscholine-10) were demonstrated to associate with CD81 in a defined molar ratio (1
2) of claudin-1
CD81 and in the absence of any other cellular components. In contrast, monodispersed claudin-1 (in βOG) failed to associate with CD81. Claudin-1/CD81 complexes were stabilized by the presence of cholesteryl hemisuccinate (CHEMS) consistent with literature reports that cholesterol promotes complex formation in mammalian cells 
. In summary, this study represents the first in-depth characterization of recombinant, full-length claudin-1; the data presented here should accelerate the structural analysis of the claudin superfamily and promote structure-aided design of therapeutic agents that target the early entry step of the HCV lifecycle.