Gutted adenoviral (Ad) vectors have a greater cloning capacity and elicit less immune response than conventional Ad vectors. Unfortunately, clinical use of gutted vectors has been slowed by production difficulties, including low yield and a tendency for recombinant virus to emerge. These two problems are related, because expansion of dilute vector stocks requires selective pressure against helper virus. The ability to rescue gutted virus at high titer would lessen the requirement for selective pressure, thereby limiting the advantage afforded to undesirable recombinants. We tested gutted virus rescue from plasmids and from synthetic terminal protein (TP)-DNA complexes by transfection/infection or cotransfection with various forms of helper viral DNA. Optimal rescue required cotransfection of gutted and helper genomes with identical origins of replication. Transfection/infection, which introduces unequal origins, was 30 times less efficient than cotransfection of genomes that had been released from plasmid DNA and bore identical origins. Cotransfection of TP-linked genomes was several times more efficient than that of unlinked genomes, yielding average gutted viral titers above 10(7) transducing units (TU)/ml. In addition, we found that limited expression of Cre recombinase doubled the yield of gutted virus. Using these techniques, gutted viruses can be rescued at titers greater than 3 x 10(7) TU/ml, about 100 times higher than is usually achieved. Finally, we found that high-titer lysates could be serially passaged on Cre-negative cells without loss of titer, further reducing selective pressure. These methods allow large-scale production of gutted virus in three or four serial passages, while minimizing exposure to Cre recombinase.