Our structure of the adenovirus-12 fiber knob domain in complex with a soluble fragment of CAR, the cellular receptor for group B coxsackieviruses and many adenoviruses, challenges established models of the mechanism of virus-receptor binding. The CAR binding site on knob is formed from surface loops, which are the most variable parts of the knob protein and account for serological differences, rather than from more highly conserved regions that were proposed as candidate binding sites based on an earlier crystal structure of the Ad5 knob. The overlap of a receptor binding site of conserved specificity and antigenically variable regions strongly suggests that aspects of the molecular mechanism of virus-receptor binding may be unique and without precedent in non-viral systems. Both the knob-CAR and HIV gp120-CD4 interfaces contain unusual large water-filled cavities, allowing for indirect water-mediated binding in addition to direct contact of amino acids across the interface. This feature may be part of a mechanism to buffer binding site specificity against overlapping antigenic variation. We will analyze the mechanism of knob-CAR binding in detail, using the combined approaches of x-ray crystallography, mutagenesis and biochemistry. The structures of fiber knobs from different adenovirus serotypes will be solved alone and in complex with CAR to study the impact of antigenic variation on the structure and activity of the receptor binding site. The contribution of individual knob amino acids that contact CAR directly or indirectly through cavity-bound water molecules will be assessed by mutagenesis and quantitative binding assays. Knob-specific monoclonal antibodies which interfere with knob-CAR binding will be isolated to characterize the antigenic structure of the receptor binding sites. Knowledge gained from our studies may be broadly applicable to understanding mechanisms of receptor binding in other virus systems, including HIV. Our results also may have application to the development of vaccines and anti- viral drugs, and they will impact efforts to retarget the tropism of adenovirus-based vectors for gene therapy.
