Alpha-defensin binding expands human adenovirus tropism

Abstract

Mammalian α-defensins are a family of abundant effector peptides of the mucosal innate immune system. Although primarily considered to be antimicrobial, α-defensins can increase rather than block infection by certain prominent bacterial and viral pathogens in cell culture and in vivo. We have shown previously that exposure of mouse and human adenoviruses (HAdVs) to α-defensins is able to overcome competitive inhibitors that block cell binding, leading us to hypothesize a defensin-mediated binding mechanism that is independent of known viral receptors. To test this hypothesis, we used genetic approaches to demonstrate that none of several primary receptors nor integrin co-receptors are needed for human α-defensin-mediated binding of HAdV to cells; however, infection remains integrin dependent. Thus, our studies have revealed a novel pathway for HAdV binding to cells that bypasses viral primary receptors. We speculate that this pathway functions in parallel with receptor-mediated entry and contributes to α-defensin-enhanced infection of susceptible cells. Remarkably, we also found that in the presence of α-defensins, HAdV tropism is expanded to non-susceptible cells, even when viruses are exposed to a mixture of both susceptible and non-susceptible cells. Therefore, we propose that in the presence of sufficient concentrations of α-defensins, such as in the lung or gut, integrin expression rather than primary receptor expression will dictate HAdV tropism in vivo. In summary, α-defensins may contribute to tissue tropism not only through the neutralization of susceptible viruses but also by allowing certain defensin-resistant viruses to bind to cells independently of previously described mechanisms.

Fig 1. HD5 promotes HAdV-D64 infection despite blockade of the primary receptor, sialic acid.

(A) Representative images of A549 cell monolayers in 96-well plates infected with HAdV-C5 or HAdV-D64 in the presence or absence of HD5 were obtained at a resolution of 50 μm. Grayscale intensity correlates with eGFP expression. (B–C) HAdV-D64 infection of A549 cells that were treated or not with (B) 100 μg/mL WGA or (C) 10 mU/mL neuraminidase (NA) in the presence or absence of HD5. Each point is an independent experiment, and bars are the mean ± SD of the percent infectivity compared to control, untreated cells infected in the absence of HD5 (100%).

Fig 2. HD5-mediated binding of HAdV to cells is independent of known viral receptors.

(A) Construction of C5/D64-HVR1/RGE. Capsid proteins are depicted for HAdV-C5 (blue) and HAdV-D64 (brown). The variable residues in penton base (RGD and RGE) and hexon hypervariable region 1 (HVR1) are indicated. (B) Strategy for editing CXADR exon 2 in A549 cells. The small-guide RNA (sgRNA) target site (blue) and proto-spacer adjacent motif (PAM, red) sequences are highlighted. Below are the WT reference allele and an analysis of alleles from the polyclonal CAR KO A549 cells used in this study. The percentage and number of reads for each allele are indicated on the right. The vertical dotted line is the predicted cleavage site, dashes indicate deletions, and the red box denotes an insertion. (C) Validation of CAR KO by flow cytometry analysis. The yellow line represents cells stained with the isotype control, the blue line represents WT cells stained with the anti-CAR antibody, and the red line represents CAR KO cells stained with the anti-CAR antibody. The figure is representative of three independent experiments with similar results. (D) Representative cell monolayers in 96 well plates of WT A549 and CAR KO A549 cells infected with three-fold serial dilutions of HAdV-C5 and HAdV-B35. Images were obtained at a resolution of 50 μm, and grayscale intensity correlates with eGFP expression. (E) Binding of AF488-labeled viruses to A549 or CAR KO A549 cells in the presence of HD5. Data are the fold change in geometric mean fluorescence (GMF) of the entire cell population relative to the no HD5 condition for each virus bound to WT cells. Each point is an independent experiment, and bars are the mean ± SD.

Fig 3. HD5-mediated infection requires integrin co-receptors.

(A) Representative images and (B) quantification of infection of (top) A549 and (bottom) CAR KO cells with the indicated viruses in the presence or absence of HD5. Images in A were obtained at a resolution of 50 μm, and grayscale intensity correlates with eGFP expression. In B, data in both graphs are normalized to infection of A549 cells in the absence of defensin (control infection) and are the means ± SD of at least 3 individual experiments.

Fig 4. HD5-mediated infection requires HD5 multimerization.

(A) Binding of AF488-labeled viruses to A549 or CAR KO cells in the presence of E21me HD5. Data are the fold change in geometric mean fluorescence (GMF) of the entire cell population relative to the no HD5 condition for each virus bound to WT cells. Each point is an independent experiment, and bars are the mean ± SD. (B) Representative images of A549 and CAR KO cells infected with the indicated viruses in the presence or absence of E21me HD5. Images were obtained at a resolution of 50 μm, and grayscale intensity correlates with eGFP expression. Quantification of infection of (C) A549 and (D) CAR KO cells with the indicated viruses in the presence or absence of E21me HD5. Data in both graphs are normalized to infection of A549 cells in the absence of defensin (control infection) and are the means ± SD of at least 3 individual experiments.

Fig 5. α-Defensin-mediated infection is not restricted to HD5.

(A) Representative images and (B) quantification of infection of (top) A549 and (bottom) CAR KO cells with the indicated viruses in the presence or absence of HNP1. Images in A were obtained at a resolution of 50 μm, and grayscale intensity correlates with eGFP expression. In B, data in both graphs are normalized to infection of A549 cells in the absence of defensin (control infection) and are the means ± SD of at least 3 individual experiments.

Fig 6. Generation and validation of an HD5-resistant, HAdV-B-based virus.

(A) Construction of B35/D37-HVR1/Vertex. Capsid proteins are depicted for HAdV-B35 (gray) and HAdV-D37 (blue). Hexon hypervariable region 1 (HVR1) and the variable residues in fiber (EDES and GYAR) are indicated. (B) Infection of A549 cells with the indicated viruses in the presence or absence of HD5. Data are normalized to infection in the absence of defensin (control infection) and are the means ± SD of at least 3 individual experiments. (C) Infection of CD46 KO HAP1 cells in the presence or absence of 20 μM HD5. Data is normalized to infection of control HAP1 cells in the absence of defensin. Each connected pair of points is an independent experiment. The dashed line is 100% (control) infection. (D) Representative images of the data in C were obtained at a resolution of 50 μm, and grayscale intensity correlates with eGFP expression.

Fig 7. α-Defensin-mediated infection expands the tropism of HAdV.

Infection of (A) A549 or (B) CMT-93 cells in monoculture or (C) A549 and (D) CMT-93 cells in mixed culture with B35/D37-HVR1/Vertex in the presence or absence of 20 μM HD5. Infected cells were enumerated by flow cytometry at 24 h post-infection. Data are the mean percentage of each cell type that are GFP positive ± SD of at least 3 individual experiments. (E) Representative flow cytometry plots of mixed cultures (left) without virus or infected with B35/D37-HVR1/Vertex in the (middle) absence or (right) presence of 20 μM HD5.