Artificial extension of the adenovirus fiber shaft inhibits infectivity in coxsackievirus and adenovirus receptor-positive cell lines

Abstract

Recent studies demonstrate that virus-cellular receptor interactions are not the sole determinants of adenovirus (Ad) tropism. It has been shown that the fiber shaft length, which ranges from 6 to 23 beta-repeats in human Ads, also influences viral tropism. However, there is no report that investigates whether artificial extension of the shaft alters the infectivity profile of Ad. Therefore, we constructed Ad serotype 5 (Ad5) capsid-based longer-shafted Ad vectors by incorporating Ad2 shaft fragments of different lengths into the Ad5 shaft. We show that "longer-shafted" Ad vectors (up to 32 beta-repeats) could be rescued. We also show that longer-shafted Ad vectors had no impact on knob-CAR (coxsackievirus and Ad receptor) interaction compared to wild-type Ad. Nevertheless, gene transfer efficiencies of longer-shafted Ad vectors were lower in CAR-positive cell lines compared to wild-type Ad. We suggest that artificial extension of the shaft can inhibit infectivity in the context of CAR-positive cell lines without modification of knob-CAR interaction.

FIG. 1.

Design and construction of longer-shafted Ad vectors. According to a triple β-spiral model for Ad2 fiber shaft, Ad5 capsid-based longer-shafted Ad vectors were made by incorporating Ad2 shaft fragments of different lengths into the NheI site of the Ad5 shaft. In order to generate Ad2 shaft fragments, PCR was performed with four pairs of primers (Ad5longI, PLup and PLdownI; Ad5longII, PLup and PLdownII; Ad5longIII, PLup and PLdownIII; Ad5longIV, PLup and PLdownIV). The resulting longer-shafted Ad vectors have a total of 27, 32, 36, and 40 β-repeats, respectively.

FIG. 2.

Generation of longer-shafted Ad vectors. To confirm that the genomes of longer-shafted Ad vectors show the expected structure, DNA isolated from viral particles was analyzed. (A) PCR was performed to detect the length of total fiber region by using a pair of primers (PFup and PFdown). (B) DNA sequencing of complete shaft region was performed by using a pair of primers (PSup and PSdown). The results of sequencing are shown as schematic DNA alignment. Ad5G was used as a control.

FIG. 3.

Generation of longer-shafted Ad vectors. To verify that longer-shafted Ad vectors have the predicted different lengths of fiber proteins, Western blot analysis was performed. Purified Ad virions were boiled and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then detected with anti-fiber MAb 4D2. Ad5G or recombinant Ad5 fiber protein was used as a control.

FIG. 4.

Knob-CAR interaction of longer-shafted Ad vectors. Purified Ad virions of Ad5GL (control), Ad5longIGL, and Ad5longIIGL immobilized in the wells of an ELISA plate were incubated with various concentrations of sCAR-His₆ protein. Binding sCAR-His₆ protein was then detected with anti-sCAR-His₆ MAb 2A3.

FIG. 5.

Infectivity profiles of longer-shafted Ad vectors in cell lines expressing variable levels of CAR. Gene transfer efficiencies of Ad5GL (control), Ad5longIGL, and Ad5longIIGL were tested in cell lines expressing high levels of αV integrins but variable levels of CAR: HeLa cells (high CAR) (A), HUVEC (moderate CAR) (B), and RD cells (low CAR) (C). All cells were infected at 10¹, 10², and 10³ VPs/cell. At 48 h postinfection, the cell lysates were assayed for luciferase activity, expressed as relative light units (RLU) per milligram of cellular protein. The results are the means ± the standard deviations (SD) of triplicate experiments.

FIG. 6.

Infectivity profiles of longer-shafted Ad vectors after blocking of knob-CAR interaction. HeLa cells (A), HUVEC (B), and RD cells (C) were preincubated with or without recombinant Ad5 fiber knob protein at 100 μg/ml for 10 min and then infected with Ad5GL (control), Ad5longIGL, and Ad5longIIGL at 10³ or 10⁴ VPs/cell. At 30 h postinfection, the cell lysates were assayed for luciferase activity, expressed as relative light units (RLU) per milligram of cellular protein. The results are the means ± the standard deviations of triplicate experiments.

FIG. 7.

Infectivity profiles of longer-shafted RGD Ad vector. Gene transfer efficiencies of Ad5GL, Ad5longIIGL, Ad5RGDGL, and Ad5RGDlongIIGL were tested in cell lines expressing high levels of αV integrins but variable levels of CAR: HeLa cells (A), HUVEC (B), and RD cells (C). All cells were infected at 10¹, 10², and 10³ VPs/cell. At 48 h postinfection, the cell lysates were assayed for luciferase activity expressed as relative light units (RLU) per milligram of cellular protein. The results are the means ± the standard deviations of triplicate experiments.

FIG. 8.

Model for Ad infectivity based on fiber shaft extension. Artificial extension of Ad fiber shaft disturbs penton base-αV integrin interaction in a context whereby longer-shafted Ad vectors must exclusively use CAR for attachment. On the other hand, shaft extension does not affect infectivity in CAR-independent cell entry pathway.