Cis- and trans-acting components involved in adenovirus type 5 DNA packaging

Late in the infectious cycle of adenoviruses (Ad), viral DNA is packaged into preformed, empty capsids in a selective manner and with left-to-right polarity. This process is thought to involve a coordinate interaction of cis-acting sequence elements and trans-acting packaging factors. Whereas a packaging domain is known to be located within the Ad type 5 left end terminus, and has previously been characterized, no information is available about trans-acting packaging components. The focus of this dissertation is of dual nature. Cis-acting packaging elements (A repeats) were further characterized mutationally and genetically. Information gained from these studies was subsequently used for the design of probes for the identification of putative trans-acting components of the adenovirus packaging machinery in vitro.;In previous analyses, the Ad 5 packaging domain was dissected into at least seven functionally redundant packaging elements termed A repeats I to VII, following a loosely conserved consensus motif. Despite their functional redundancy, A repeats are governed by a functional hierarchy. Our mutational studies have focused on the four most critical repeats and have defined a new consensus motif for packaging elements, which is bipartite in nature and evolutionarily conserved. Furthermore, the construction of viral mutants carrying minimally sized packaging domains in place of the entire packaging domain allowed for a rating of the functional independence of three of these repeats relative to one another (AI ;Functional minimal packaging domains were subsequently used as probes in gel mobility shift assays for the detection of specific protein binding in vitro. Two distinct cellular activities were identified that interact with A repeats, the transcription factor COUP-TF and an unknown complex (P-complex). P-complex affinity for different A repeats correlated with the ability of the respective fragments to function as minimal packaging domains in vivo, making this activity an attractive candidate for a bona-fide packaging factor. In contrast, COUP-TF binding did not parallel our in vivo data. COUP-TF overexpression dramatically reduced virus yield, which is, at least in part, the result of an inhibition of late gene expression.