The characteristics and specificities of spleen and peritoneal cytotoxic cells generated during lymphocytic choriomeningitis virus (LCMV) infection of C3H/St mice were examined. Activated natural killer (NK) cell activity was identified in fresh leukocyte populations from the 2nd to 8th days postinfection, whereas virus-specific cytotoxic T-cell activity was detected from the 6th to 14th days. When leukocytes were cultured overnight at 37 degrees C before assay, T-cell activity was still observed, but nonspecific activated NK cell-like cytotoxicity was only detected on the 6th and to a lesser degree the 8th day postinfection. Overnight culture of leukocytes taken earlier in the infection eliminated their NK cell activity. Similar activities were seen with spleen cell, plastic-adherent peritoneal cell, and nonadherent peritoneal cell populations. The virus-specific cytotoxicity observed with adherent peritoneal cells was due to contamination with cytotoxic T cells, as shown by H-2-restricted cytotoxicity and sensitivity to anti-theta antibody and complement. The nonspecific cultured day 6 effector cell from either the spleen or peritoneum displayed killing specificities and other physical properties identical to those of activated NK cells, but had sensitivities to anti-theta antibody and complement intermediate between activated day 3 NK cells and cytotoxic T cells. Culture stable NK-like cells were not found in athymic nude mice, suggesting a T-cell-dependent mechanism. Whereas LCMV spleen homogenates contained 10-fold-higher levels of interferon at day 2 than at day 6 postinfection, substantially more (nearly 20-fold) interferon was made in cultures of day 6 cells than day 2 cells. Spleen interferon was predominantly type I, whereas the culture interferon was predominantly type II, as shown by acid lability studies. Significant levels of interferon were produced by nylon-wool-passed day 6 spleen cells, and virtually all interferon production was eliminated by treatment of either day 2 or day 6 cells with antibody to theta antigen and complement, suggesting that T cells produced the interferon in vitro. Furthermore, athymic nude mice had no culture-stable NK cells 6 days postinfection, and spleen cells from them failed to produce significant levels of interferon in vitro. Addition of interferon (type I, fibroblast) to cultured C3H spleen cells affect the already elevated levels of cytotoxicity in day 6 cultures, suggesting that the NK cells in the day 6 culture were already activated. Our results suggest that T cells responding to LCMV infection secrete interferon type II which causes the continued activation of NK cells in culture. The resulting population of activated NK cells therefore appears to be relatively stable in culture and to express more theta antigen because of this T-cell dependence. Although one could mistakenly or allospecific cytotoxic T cells or cytotoxic macrophages, more careful examination shows that they are most likely activated NK cells...