Figure 1.
Claudin-1 forms monomers and higher order structures in yeast membranes.
(A) A schematic representation of the claudin-1 protein produced in this study. The membrane is represented in grey (the extra- and intracellular sides are labeled), the extracellular loops, EC1 and EC2, are marked and the His6 tag is indicated in the carboxy-terminal tail. The predicted molecular mass is 23.7 kDa. (B) Recombinant claudin-1 in P. pastoris membranes forms monomers, dimers and trimers as determined by non-reducing SDS-PAGE (1 µg protein loaded per well). Endogenous claudin-1 in Huh-7.5 hepatoma cells was analyzed for comparison (5 µg protein loaded per well) showing monomers and oligomers. Protein concentrations in the two cell types are different because the yeast cells overexpress recombinant claudin-1; this does not affect the antigenicity of the protein, as shown in Fig. 3. Molecular size markers are indicated in the intervening lane.
Figure 2.
Antigenicity of human claudin-1 in yeast protoplasts.
Conformation-dependent antibodies specific for claudin-1 (R&D Systems) and, as a control, CD81 (2s131) were used as tools to probe the antigenicity of yeast-expressed claudin-1. Secondary antibodies were Alexa Fluor 488 goat anti-mouse IgG (H+L) and Alexa Fluor 633 goat anti-rat IgG (H+L) (Invitrogen). The parental X33 strain with no heterologous protein expression was used as a negative control. (A) Fluorescence activated cell sorting shows specific antibody binding with anti-claudin-1 antibodies for X33 (0.93%) and claudin-1 (33.6%) protoplasts. For anti-CD81 antibodies, the corresponding values are X33 (1.9%) and claudin-1 (0.93%). (B) Confocal imaging shows specific antibody binding for protoplasts expressing human claudin-1. The protoplasts were immuno-flourescently labelled with anti-claudin-1 or a relevant isotype control. Settings were optimized for each fluorescent protein to obtain the highest signal to noise ratio, while controlling for cross talk. Relevant isotype control antibodies did not bind.
Figure 3.
Biophysical analysis of purified claudin-1.
(A) Coomassie-stained SDS-PAGE gel showing CD81 (lane 1 containing 2.4 µg protein), a control 4TM protein (CD82; lane 2; 1.5 µg protein) and claudin-1 (lane 3; 1.2 µg protein) solubilised in βOG and eluted from a nickel affinity column. Claudin-1 (highlighted with an arrow) runs as an apparent dimer. Molecular size markers are indicated. (B) Analytical ultracentrifugation trace for claudin-1 in βOG micelles collected on a Proteome Lab XL-I instrument. (C) ELISA data based on OD450 readings showing antibody reactivity (n = 2; error bars are the standard deviation) for βOG-extracted and purified claudin-1, CD81 and CD82. (D) Analytical ultracentrifugation trace for claudin-1 in profoldin-8 micelles collected on a Proteome Lab XL-I instrument. Data were analyzed using the continuous distribution, c(s), model, which calculates the distribution of sedimenting species taking into account their diffusion. Shown are the mass distributions of particles within the samples, c(s), as a function of the sedimentation co-efficient (S; measured in units of Svedberg, with 1 S = 10−13 s). Inset is a silver-stained SDS-PAGE gel showing claudin-1 solubilized in profoldin-8 and eluted from a nickel affinity column. Monomers and oligomers are highlighted with arrows.
Figure 4.
Purified claudin-1 does not bind HCV E2.
ELISA data based on OD450 readings (n = 3; error bars are the standard deviation) showing reactivity with HCV E2. Claudin-1 in both βOG micelles (monodispersed) and profoldin-8 micelles (a dynamic mixture of oligomers) were analyzed and compared with CD81 (a positive control for HCV E2 binding) and CD82 (a negative control for HCV E2 binding). A control GNA lectin gave an OD450 signal of 3.29±0.04 in the presence of HCV E2, and 0.04±0.01 in its absence.
Figure 5.
Biological analysis of claudin-1 in proteoliposomes.
(A) Proteoliposome preparations containing claudin-1 or a control protein (CD82) were diluted and evaluated for anti-His reactivity by ELISA; data are represented as optical density (OD) at 450 nm. Inset is an immunoblot of foscholine-10 extracted claudin-1 stained with an anti-hexahistine tag antibody. Monomers and oligomers are highlighted with arrows. Proteoliposome preparations were normalized for His OD units and evaluated for their effect on (B) HCV as well as (C) HCVpp and VSV-Gpp infectivity. Data are expressed relative to untreated control virus infection.
Figure 6.
Oligomeric claudin-1 associates with CD81 at a specific molar ratio of 1∶2.
A plot of relative intensity of differently-sized particles (RH reported in nm, derived from DLS measurements at room temperature) of claudin-1, CD81 or a control protein (CD82), alone or as mixtures, solubilized in detergent micelles and in the presence of CHEMS. The data are the average of the first 10 measurements of the DLS experiment. Inset are DLS heat maps, where the colour spectrum indicates the amplitude of the signal for a given hydrodynamic radius (RH) value as a function of time; red is high- and blue is low amplitude. (A) CD81 in βOG. (B) CD82 in CD. (C) claudin-1 in foscholine-10; claudin-1 particles have a broad radial distribution consistent with a dynamic pool of oligomers. (D) claudin-1 mixed with CD81 in a 1∶2 molar ratio; a distinct 30 nm particle is highlighted with a red asterisk. (E) CD82 mixed with CD81 in a 1∶2 molar ratio; in addition to a dominant peak at 6 nm, some higher oligomers are present, which do not form a distinct peak. (F) claudin-1 mixed with CD82 in a 1∶2 molar ratio; in addition to a dominant peak at 6 nm, higher oligomers are present at >40 nm.
Figure 7.
Claudin-1/CD81 complexes are stabilized by CHEMS.
(A) A plot of relative intensity of differently-sized particles (RH reported in nm) is shown for the claudin-1/CD81 1∶2 mix with (closed circles) and without (open circles) CHEMS. (B) The stable RH = 30 nm particle forms within a few minutes of mixing and is stable in the presence of CHEMS for 18 h. In the absence of EC1, aggregation occurs after 4 h.
Table 1.
Distribution of particle hydrodynamic radii (RH) derived from DLS measurements of purified recombinant claudin-1 and CD81.
Figure 8.
Proposed assembly of claudin-1/CD81 complexes (RH = 30 nm).
A cartoon assembled in ChemDraw illustrating seven uroplakin-like CD81 complexes (black), each containing six homodimers, arranged with 24 claudin-1 homodimers (red). The resulting molar ratio of 24∶42 claudin-1∶CD81 is close that observed experimentally (Table 1), and generates a RH = 30 nm particle. The model displayed is the most symmetric of those generated.