Identification of coagulation factor (F)X binding sites on the adenovirus serotype 5 hexon: effect of mutagenesis on FX interactions and gene transfer

Abstract

Recent studies have demonstrated the importance of coagulation factor X (FX) in adenovirus (Ad) serotype 5-mediated liver transduction in vivo. FX binds to the adenovirus hexon hypervariable regions (HVRs). Here, we perform a systematic analysis of FX binding to Ad5 HVRs 5 and 7, identifying domains and amino acids critical for this interaction. We constructed a model of the Ad5-FX interaction using crystallographic and cryo-electron microscopic data to identify contact points. Exchanging Ad5 HVR5 or HVR7 from Ad5 to Ad26 (which does not bind FX) diminished FX binding as analyzed by surface plasmon resonance, gene delivery in vitro, and liver transduction in vivo. Exchanging Ad5-HVR5 for Ad26-HVR5 produced deficient virus maturation. Importantly, defined mutagenesis of just 2 amino acids in Ad5-HVR5 circumvented this and was sufficient to block liver gene transfer. In addition, mutation of 4 amino acids in Ad5-HVR7 or a single mutation at position 451 also blocked FX-mediated effects in vitro and in vivo. We therefore define the regions and amino acids on the Ad5 hexon that bind with high affinity to FX thereby better defining adenovirus infectivity pathways. These vectors may be useful for gene therapy applications where evasion of liver transduction is a prerequisite.

Figure 1

Modeling of the Ad5:FX interaction to identify FX contact points in the hexon protein. Models of the interaction between hexon and FX were used to determine candidate contact residues in the Ad5 hexon. (A) An atomic resolution representation of one such model is shown in top view (ie, viewed from the virion exterior). Hexon is shown in blue. FX is shown in purple. (B) Polypeptide chains within the trimeric capsomere are named and colored (A, red; B, yellow; and C, purple) relative to the orientation of FX. Missing regions in the crystal structure of hexon are represented with blue sticks and highlighted with black arrows. (C-E) Contact residues are shown for each of the 3 models calculated and named/colored according to their chain assignment relative to FX. Amino acids targeted for mutagenesis have labels that are colored according to regions identified in panel F. In each panel, hexon is represented such that FX is in approximately the same orientation as in panel C. (F) Amino acid alignment of Ad5 and Ad26 hexon HVR sequences to highlight domains (HVR5, red box; HVR7, blue box) and amino acids targeted for mutagenesis studies (point mutations highlighted by *).

Figure 2

Swapping HVR5 and 7 of Ad5 to Ad26 and its effect on FX binding by SPR. (A) Viewlite images of the Ad5 hexon showing HVR5 (blue), HVR7 (green), as well as the remaining HVR amino acids (gray). (B) Virus particle titers and VP/pfu ratios for each virus produced. (C) SPR analysis of virus interaction with immobilized FX (500 RU). Sensorgrams of 10¹¹ VP/mL injected at a flow rate of 30 μL/minute.

Figure 3

In vitro cell binding and transduction of HVR5 and 7 mutant adenoviruses. (A) Cell binding at 4°C to SKOV3 cells in the presence or absence of physiologic concentrations of FX. (B) Fold change in FX-assisted virus association with cell versus absence of FX. (C) SKOV3 cell transduction at 48 hours postinfection after a 3-hour exposure to each virus in the presence or absence of FX in serum-free conditions (P < .01 vs no FX conditions).

Figure 4

In vivo analysis of Ad5HVR5(Ad26) and Ad5HVR7(Ad26). Male MF-1 mice were injected with 10¹⁰ VP of each virus via the tail vein. Mice were killed at 48 hours, and livers were harvested and processed for quantification of β-galactosidase activity and staining (P < .001 vs Ad5).

Figure 5

Effect of point mutagenesis on FX binding by SPR. (A) Viewlite images of mutations in the Ad5 hexon for each virus. (B) Virus titers and VP/pfu ratios for each virus produced. (C) SPR analysis of virus interaction with immobilized FX (500 RU). Sensorgrams of 10¹¹ VP/mL injected at a flow rate of 30 μL/minute.

Figure 6

In vitro analysis of Ad5 hexon point mutants. (A) Cell binding at 4°C to SKOV3 cells in the presence or absence of physiologic concentrations of FX, (*P < .01 vs no FX conditions). (B) Fold change in FX-assisted virus association with cell versus absence of FX. (C) SKOV3 cell transduction at 48-hour postinfection after a 3-hour exposure to each virus in the presence or absence of FX in serum-free conditions (P < .001 vs no FX conditions).

Figure 7

In vivo analysis of Ad5 hexon point mutants. MF-1 mice were injected with 10¹⁰ VP of each virus via the tail vein. Mice were killed at 48 hours, and livers were harvested and processed for quantification of β-galactosidase activity and staining (P < .001 vs Ad5).