Clinical efficacy of adenovirus-mediated cancer gene therapy has been limited thus far. To improve its oncolytic effect, a replication-competent adenoviral vector was previously constructed to express high levels of p53 at a late time point in the viral life cycle. p53 expression from this vector improved tumor cell killing and viral spread in vitro. However, p53 function is antagonized by cellular mdm2 and adenoviral E1b-55kD, both of which are known to bind to and inactivate p53. Therefore, a new vector (Adp53W23S) that expresses a modified p53 transgene, which does not bind to E1b-55kD and mdm2, was constructed. The modified p53 protein was demonstrated to have a substantially prolonged half-life, and its localization was predominantly nuclear. Viral replication was unaffected by expression of the modified p53 and cancer cell killing was improved in vitro. However, in a xenograft model, efficacy was not significantly different from control virus. In conclusion, expression of a degradation-resistant p53 transgene late in the life cycle of a replication-competent adenovirus improves p53 stability and cancer cell killing in vitro. However, other factors, such as the adenoviral E1b-19kD and E1a proteins, which oppose p53 function, and limitations to viral spread need to be addressed to further improve in vivo efficacy.