Harris HJ, Clerte C, Farquhar MJ, Goodall M, Hu K, Rassam P, Dosset P, Wilson GK, Balfe P, IJzendoorn SC, Milhiet PE and McKeating JA.
Cellular Microbiology, 15:430-445, 2013
Cellular Microbiology website
| Get a PDF file of this paper
back to publications page
We investigated potential differences in lipid mobility beween polarized and non-polarized HepG2 membranes using fluorescent lipid probes and monitoring their movement using Fluoresence Recovery After Photobleaching (FRAP). We found significantly slower Diffusion coefficients (DC) in polarized HepG2 cells (DC=0.17μm2/s) compared to non-polarized cells (DC=0.43-0.45μm2/s) (Fig.1). This change in the membrane may contribute to the reduced HCVpp entry observed in polarized cells.
TIRF microscopy of cell surface bound anti-CD81 Fabs enabled us to quantify CD81 dynamics at the basal membrane of living cells. Three diffusion modes were observed using a neural network: (1) pure Brownian diffusion (55% of total trajectories) with a median DC value of 0.17 μm2/s; (2) pure confined or restricted diffusion (30%) with a lower median DC value of 0.03 μm2/s and an average confinement diameter of 367nm; and (3) mixed diffusion, a combination of Brownian and confined (15%). Cell polarization reduced the proportion of CD81 Brownian movement (40%), whilst increasing the frequency of mixed diffusion (29%) (Fig.2). These studies show that polarization slows CD81 diffusion and increases its confinement.
The same approach was used to study HCV pseudo particle (HCVpp) trafficking. In non-polarized HepG2 cells, the dominant mode of HCVpp diffusion was Brownian (~80%), with few mixed (15%) or confined (5%) trajectories (Fig.3). Polarization led to an increase in these mixed and confined trajectories and a concomitant decrease in Brownian diffusion Furthermore, HCVpp Brownian diifusion was halved on polarized HepG2 cells compared to non-polarized cells, from 0.14μm2/s to 0.07μm2/s respectively, leading us to conclude that HepG2 polarization has similar effects on both HCVpp and CD81 movement.
In summary, our data suggest that CD81 needs to freely diffuse in the cellular membrane to form transient associations with claudin-1 (Harris et al., 2010, Harris et al., 2008). Our recent observation that HCV promotes CD81 and claudin-1 co-endocytosis (Davis 2012) supports a model where virus-CD81 complexes move laterally across the plasma membrane prior to internalization (Farquhar et al., 2012). The observation in this paper that polarization reduces the dynamics of both CD81 and HCVpp at the basolateral membrane suggests a new pathway by which polarized epithelia may restrict pathogen invasion, and highlights the importance of studying pathogen-host cell interactions in realistic polarized cell systems.
Many airway and enteric viruses target the polarized epithelial apex during host invasion. In contrast, hepatitis C virus (HCV) enters the liver via the sinusoidal blood and engages receptors at the basal surface of hepatocytes in the polarized liver parenchyma.
We previously reported that hepatoma polarization limits HCV entry by an unknown mechanism. Real-time FRAP and single particle tracking of viral receptors and HCV pseudoparticles (HCVpp) demonstrate the dynamic nature of these proteins at the cell surface. We observed a significant reduction in lipid and CD81 mobility at the basal membrane of polarized HepG2 cells that influenced HCVpp diffusion. We demonstrate a role for the CD81 C-terminus in regulating both protein mobility and HCVpp lateral diffusion. Collectively, these studies provide a new mechanism for polarized epithelia membrane protein dynamics to limit pathogen infection.
Figure 1. Effect of HepG2 polarization on lipid dynamics.
Quantitative FRAP measurement of the dynamice of the fluorescent lipids DiO-C18 and DiD-C18 in polarized (day 7 post plating with a polarization index of 43%) and non-polarized HepG2 cells (day 1 post plating with a polarization index of 2%). Representative normalized FRAP recovery curves in polarized (Pol, black) and non- polarized (Non Pol, grey) cells are shown. Median diffusion coefficient (DC) of DiO-C18 (A) and DiD-C18 (B) in polarized and non-polarized HepG2 cells, where each point represents one bleached region of interest and the vertical black line represent the median value. A minimum of 10 cells and 100 ROIs were measured per parameter. Non-parametric Mann Whitney T tests were used to determine the significance of the differences seen (***, p<0.0001).
Figure 2. Single particle tracking of CD81 in polarized and non-polarized HepG2 cells.
(A) Anti-CD81 2s66 Fab was pre-incubated with cells for 1 hour prior to FRAP imaging. The Fab had no significant effect on CD81 mobility (Mf) or diffusion (DC). (B) The TIRF image traces CD81 trajectories on the non-polarized HepG2 cell surface, the cartoon depicts representative trajectories with Brownian (B), mixed (M) or confined (C) properties. (C) The relative frequencies of the anti-CD81 2s66 Fab trajectories in polarized (Pol - white bars) and non-polarized (Non Pol - black bars) cells are shown. (D) Scatter plots for diffusion coefficients (DC) calculated from the Brownian trajectories in polarized and non-polarized cells. Each point represents one trajectory, with the black line indicating the median DC, a minimum of 10 cells and 250 trajectories were measured. Non-parametric Mann Whitney T tests were used to determine the degree of significance (***, p<0.0001).
Figure 3: Single particle tracking of viral particles in polarized and non-polarized HepG2 cells. (A) The relative frequencies of Brownian (B), mixed (M) and confined (C) HCVpp diffusion at the polarized (Pol - white bars) and non-polarized (Non Pol - black bars) HepG2 cell membrane are shown. (B) Scatter plots for HCVpp diffusion coefficients bound to polarized or non-polarized HepG2 cells were calculated from the Brownian trajectories in polarized and non-polarized cells. Each point represents one trajectory, with the black line indicating the median DC, a minimum of 10 cells and 250 trajectories were measured. Non-parametric Mann Whitney T tests were used to determine the degree of significance ( **, p< 0.001; ***, < 0.0001).