Integrin αvβ5 is a primary receptor for adenovirus in CAR-negative cells

Background: Viruses bind to specific cellular receptors in order to infect their hosts. The specific receptors a virus uses are important factors in determining host range, cellular tropism, and pathogenesis. For adenovirus, the existing model of entry requires two receptor interactions. First, the viral fiber protein binds Coxsackie and Adenovirus Receptor (CAR), its primary cellular receptor, which docks the virus to the cell surface. Next, viral penton base engages cellular integrins, coreceptors thought to be required exclusively for internalization and not contributing to binding. However, a number of studies reporting data which conflicts with this simple model have been published. These observations have led us to question the proposed two-step model for adenovirus infection. Results: In this study we report that cells which express little to no CAR can be efficiently transduced by adenovirus. Using competition experiments between whole virus and soluble viral fiber protein or integrin blocking peptides, we show virus binding is not dependent on fiber binding to cells but rather on penton base binding cellular integrins. Further, we find that binding to low CAR expressing cells is inhibited specifically by a blocking antibody to integrin alpha v beta 5, demonstrating that in these cells integrin alpha v beta 5 and not CAR is required for adenovirus attachment. The binding mediated by integrin alpha v beta 5 is extremely high affinity, in the picomolar range. Conclusions: Our data further challenges the model of adenovirus infection in which binding to primary receptor CAR is required in order for subsequent interactions between adenovirus and integrins to initiate viral entry. In low CAR cells, binding occurs through integrin alpha v beta 5, a receptor previously thought to be used exclusively in internalization. We show for the first time that integrin alpha v beta 5 can be used as an alternate binding receptor.