Results obtained over the past 40 years have demonstrated that tumor cells of all types tested have an elevated growth requirement for methioninase compared with normal cells. Recombinant methioninase (rMETase) cloned from Pseudomonas putida has been found previously to be an effective antitumor agent attributable to deprivation of the extracellular methionine source of the tumor. To degrade intracellular methioninase, we have now developed an adenoviral vector inserted with the P. putida methioninase (MET) gene (rAd-MET). The in vitro efficacy of rAd-MET was tested on the OVCAR-8 human ovarian cancer cell line, the HT1080 human fibrosarcoma cell line, and human normal fibroblasts. rAd-MET transduction of OVACAR-8 and HT1080 resulted in high levels of methioninase expression up to 10% or more of the total protein of the cells, depending on the multiplicity of infection. The IC50 of rAd-MET for OVCAR-8 cells in 96-well plates was approximately 2 x 106 plaque-forming units (pfu)/well. The IC50 of control adenovirus (control-rAd) was 4 x 10(7) pfu/well, 20 times higher than rAd-MET. In the presence of the IC50 of 2 x 10(6) pfu/well of rAd-MET, the addition of 0.025 units/ml of rMETase, which is 25% of the IC50, resulted in a 90% inhibition of tumor cell number. This indicated that rAd-MET enhanced the efficacy of rMETase. In contrast, 2 x 10(6) pfu/well of control-rAd in combination with 0.025 units/ml of rMETase had an efficacy of only 10% inhibition of cell number. The synergistic effect of the combination of rMETase and rAd-MET was quantitated by calculating the combination index (CI). The CIs for all combinations of rAd-MET and rMETase tested on OVCAR-8 were <0.7 with a mean of 0.5, indicating synergy. Similar synergy of rAd-MET and rMETase was seen on HT1080 human fibrosarcoma cells with a mean of 0.74. In contrast, the CIs of all combinations of rMETase and control adenovirus concentrations tested on both cell lines had a mean CI of approximately 1, which indicated that this combination had only an additive effect. The normal fibroblasts, on the other hand, appeared relatively resistant to the MET gene because in the presence of rMETase, 2.5 x 10(7) pfu/well of rAd-MET or control rAd had almost an identical effect on cell survival. The selectively strong synergy of rAd-MET and rMETase on cancer cells allows reduced levels of each agent to be used, thus decreasing potential side effects.