The death of cultured insect cells after baculovirus infection is a time-dependent event. Without a quantitative model, it is difficult to characterize its kinetics. Our group has shown that the cell survival rate can be characterized by use of the n-target theory, which involves only two parameters : the number of hypothetical inactivation targets (n) and the first-order death rate (k). In this study, we used different recombinant viruses to examine the effect of heterologous protein expression on the cell survival rate. The proteins expressed were beta-galactosidase, human T-cell leukemia virus type I p40(x), human interleukin-2, and human tissue plasminogen activator (tPA). The survival rate was affected by protein expression, but the n value remained constant if the protein expression level was high (above 30 mg/L). Low-level expression of secreted, glycosylated tPA resulted in a reduced n value, which was restored to the normal value when the tPA signal peptide and prosequence were deleted. In addition, if the n value was normal (10-11), the level of protein expression correlated negatively with the death rate. However, if the n value was reduced by unfavorable culture conditions or foreign protein expression, the expression level correlated positively with the death rate. A dimensionless plot with kt as the dimensionless time shows that alteration of the k value while retaining constant n is equivalent to a rescaling of time. Therefore, the survival curves with constant n reduce to a single curve on the dimensionless plot. Although the underlying mechanisms affecting n and k are still unknown, this study shows that they can serve as convenient phenotypes for investigating virus-host interactions.