The initial steps leading to papillomavirus infection occur on the basement membrane prior to cell surface binding

Abstract

Using a murine challenge model, we previously determined that human papillomavirus (HPV) pseudovirions initially bind preferentially to the cervicovaginal basement membrane (BM) at sites of trauma. We now report that the capsids undergo a conformational change while bound to the BM that results in L2 cleavage by a proprotein convertase (PC), furin, and/or PC5/6, followed by the exposure of an N-terminal cross-neutralization L2 epitope and transfer of the capsids to the epithelial cell surface. Prevention of this exposure by PC inhibition results in detachment of the pseudovirions from the BM and their eventual loss from the tissue, thereby preventing infection. Pseudovirions whose L2 had been precleaved by furin can bypass the PC inhibition of binding and infectivity. Cleavage of heparan sulfate proteoglycans (HSPG) with heparinase III prevented infection and BM binding by the precleaved pseudovirions, but did not prevent them from binding robustly to cell surfaces. These results indicate that the infectious process has evolved so that the initial steps take place on the BM, in contrast to the typical viral infection that is initiated by binding to the cell surface. The data are consistent with a dynamic model of in vivo HPV infection in which a conformational change and PC cleavage on the BM allows transfer of virions from HSPG attachment factors to an L1-specific receptor on basal keratinocytes migrating into the site of trauma.

Fig. 1.

Binding of capsids to the BM and timecourse of L2 epitope exposure. To demonstrate the initial association of the PV capsids with the BM murine genital, tract sections of tissue harvested at 2 h following virus instillation were stained with a rat anti-L1 polyclonal serum (A) and a rabbit antiserum that recognizes laminin 5, a BM marker protein (B). The remaining panels compare the L1 staining (C, E, and G) and the staining with a rabbit antiserum raised against the 17/36 peptide of L2 (D, F, and H), in the same sections. Panels C and D show staining of tissue harvested at 2 h following virus instillation. Panels E and F show the distribution at 4 h postinstillation and panels G and H show the pattern of staining at the 18 h time point.

Fig. 2.

In vivo PC inhibition. The effect of the PC inhibitor, decanoyl-RVKR-cmk on infection of wild-type, untreated pseudovirus was determined. The average luminescence values obtained 48 h following virus instillation is shown. There were five mice in each treatment group; no virus, virus with buffer only or virus with inhibitor, as indicated below the x axis. Error bars indicate standard error of the mean.

Fig. 3.

Capsid association with the vaginal epithelium. Pseudovirion capsids were evaluated for their ability to interact with the BM and epithelium following various treatments. Untreated pseudovirus in PC-inhibitor-treated tissue is shown in panels A–D. Detection of capsids with rabbit anti-L1 serum is shown in panels A, B, and C (4, 18, and 30 h, respectively). The progressive loss of capsids over time is evident. The arrowhead in panel A points to capsids on the BM. Panel D shows the inability to detect anti-L2 17/36 staining at 18 h under these conditions. FPC pseudovirus binding, also detected with the rabbit anti-L1 serum, is shown in panels E–J. Binding to untreated tissues is shown in panels E and F (4 and 18 h, respectively). In each panel a region of BM association is indicated by an arrowhead and cell association is indicated by an arrow. It is of note that some cell association is evident in panel E (see magnified insert). The association throughout the epithelium is evident in panel F. FPC binding to heparinase-treated tissues is shown in panels G and H (4 and 18 h, respectively). Please note the enlarged area in panel G demonstrates the cell association of virus. The association throughout the epithelium is evident in panel H. FPC virus binding to PC inhibitor-treated tissue is shown in panels I and J (4 and 18 h, respectively). The typical BM pattern is obvious in panel I and the distribution throughout the epithelium is evident in J. L1-only VLP binding at 4 h is shown in panels K and L (untreated tissue and heparinase-treated tissue, respectively). The enlarged areas demonstrate the BM (arrowhead) and cell association (arrow) in panel K and the absence of BM binding in L.

Fig. 4.

Inhibition of FPC infection in vivo. Infection of FPC was evaluated in untreated mice or under conditions of PC inhibition by treatment with the PC inhibitor decanoyl-RVKR-cmk, indicated as RVKR. The luminescence values are shown in the left panel. The ability of heparinase III to inhibit FPC infection was determined. Results are shown in the right panel. Average radiance is shown for each condition. Each group consisted of five mice. Error bars indicate standard error of the mean.

Fig. 5.

Model for events leading to in vivo infection. Untreated, mature capsids initially bind to HSPG on the BM. This interaction leads to a conformational change that allows the accessibility of the L2 N terminus to PC cleavage (either furin shed from the wounded epithelium and/or PC5/6 attached to HSPG on the BM). This conformation allows a decreased binding affinity of the capsid for the HSPG. If PC cleavage is prevented the capsids are lost from the BM. If PC cleavage is allowed the 17–36 region of L2 is exposed on the BM. Following this conformational change, capsids can be transferred to a non-HSPG epithelial cell receptor. The transfer occurs preferentially to the basal cells as they migrate over the BM during the wound healing process. Following cell division, infection would ensue. The bottom panel summarizes the binding patterns of the three capsid preparations used in this study under various conditions as indicated; either in untreated tissue (untx), heparinase-treated tissue (hepx), or in tissue treated with the PC inhibitor, decanoyl-RVKR-cmk (RVKR). ND indicates that the experiment was not performed.