Ctions in the host are triggered by the viral infection, our findings suggest that the severity of influenza should be regulated by the host reaction associated with FasL expression, especially in the early phase of the infection. Since it was demonstrated that gld/gld mutation prevented the reduction of the survival rate(Fig. 1) but did not affect the virus titer in lung (Fig. S1), this perspective is strongly supported. Regarding the molecular function of FasL in lung inflammation mediated by lethal infection with PR/8 virus, it is known that FasL plays an effector role in killing the virus infected cells as well as the activated lymphocytes [2]. The reduction of CD3(+) T-cell population in the lungs of mice infected with a high titer of PR/ 8 virus was observed and this reduction was prevented by gld/gld mutation (Fig. S2 A and B). These data and previous report [22] suggested that the FasL/Fas signal should negatively regulate the host protection system by controlling the T-cell population rather than eliminate virus-infected cells in lethal influenza virus infection. In Fig. 4, it is demonstrated that in non-infected mice, Fas protein was expressed on several cell surfaces, but expression of FasL protein was detected on a rare population of lung cells. In B6 mice lethally infected with PR/8 virus, it was observed that expression of FasL was dramatically increased on several cell surfaces but Fas expression was not or slightly up-regulated. More importantly, this induction of FasL expression due to lethal infection was not observed in B6-IFNR-KO mice. These findings indicate that the FasL/Fas signal should be triggered by the induction 
of expression of FasL rather than Fas in mice infected with influenza A viruses, and this induction was regulated by typeI IFN mediated signal. Since, in the lung of control B6 mice lethally infected, higher induction of FasL expression in CD4(+), CD74(+), NK1.1(+) or CD11c(+) cells than other cell types was detected (Fig. 4, upper panel, light green color histogram), these cells should associate with the FasL mediated reduction of CD3(+) cell population in lung of mice lethally infected (Fig. S2). As shown in above studies, there are differences in kinetics of FasL mRNA expression between lethal and non-lethal virus infections (Fig. 3 A and C). It is also demonstrated that at 3DPI, IFN- ?is largely produced after the infection with a high titer of the virus compared to that with a low titer of the virus, and their amounts are equivalent at 5DPI (Fig. 5), suggesting that FasL expression in 
the virus-infected mice are controlled by Linolenic acid methyl ester type-I IFN depending on its time kinetics rather than its amount. Production of type-I IFN after influenza A virus infection is regulated by two different types of viral RNA recognizing receptor proteins, such as TLRs and RIG-I like proteins. While TLRs play their essential role for production of type-I IFN in macrophages or plasmacytoid dendritic cells (DC), RIG-I like proteins are critical for their production in conventional DC or fibroblasts [12,13]. In addition, it is proposed that in a AN 3199 respiratory RNA virus infection, alveolar macrophage is a main source for producing type-I IFN [23] and it is also reported that prevention of the recruitment of macrophages into the lungs protects mice against lethal PR/8 virus infection [24]. The differences in the time-kinetics of type-I IFN between the lethal and non-lethal infections might be due to the differences of mainly produci.Ctions in the host are triggered by the viral infection, our findings suggest that the severity of influenza should be regulated by the host reaction associated with FasL expression, especially in the early phase of the infection. Since it was demonstrated that gld/gld mutation prevented the reduction of the survival rate(Fig. 1) but did not affect the virus titer in lung (Fig. S1), this perspective is strongly supported. Regarding the molecular function of FasL in lung inflammation mediated by lethal infection with PR/8 virus, it is known that FasL plays an effector role in killing the virus infected cells as well as the activated lymphocytes [2]. The reduction of CD3(+) T-cell population in the lungs of mice infected with a high titer of PR/ 8 virus was observed and this reduction was prevented by gld/gld mutation (Fig. S2 A and B). These data and previous report [22] suggested that the FasL/Fas signal should negatively regulate the host protection system by controlling the T-cell population rather than eliminate virus-infected cells in lethal influenza virus infection. In Fig. 4, it is demonstrated that in non-infected mice, Fas protein was expressed on several cell surfaces, but expression of FasL protein was detected on a rare population of lung cells. In B6 mice lethally infected with PR/8 virus, it was observed that expression of FasL was dramatically increased on several cell surfaces but Fas expression was not or slightly up-regulated. More importantly, this induction of FasL expression due to lethal infection was not observed in B6-IFNR-KO mice. These findings indicate that the FasL/Fas signal should be triggered by the induction of expression of FasL rather than Fas in mice infected with influenza A viruses, and this induction was regulated by typeI IFN mediated signal. Since, in the lung of control B6 mice lethally infected, higher induction of FasL expression in CD4(+), CD74(+), NK1.1(+) or CD11c(+) cells than other cell types was detected (Fig. 4, upper panel, light green color histogram), these cells should associate with the FasL mediated reduction of CD3(+) cell population in lung of mice lethally infected (Fig. S2). As shown in above studies, there are differences in kinetics of FasL mRNA expression between lethal and non-lethal virus infections (Fig. 3 A and C). It is also demonstrated that at 3DPI, IFN- ?is largely produced after the infection with a high titer of the virus compared to that with a low titer of the virus, and their amounts are equivalent at 5DPI (Fig. 5), suggesting that FasL expression in the virus-infected mice are controlled by type-I IFN depending on its time kinetics rather than its amount. Production of type-I IFN after influenza A virus infection is regulated by two different types of viral RNA recognizing receptor proteins, such as TLRs and RIG-I like proteins. While TLRs play their essential role for production of type-I IFN in macrophages or plasmacytoid dendritic cells (DC), RIG-I like proteins are critical for their production in conventional DC or fibroblasts [12,13]. In addition, it is proposed that in a respiratory RNA virus infection, alveolar macrophage is a main source for producing type-I IFN [23] and it is also reported that prevention of the recruitment of macrophages into the lungs protects mice against lethal PR/8 virus infection [24]. The differences in the time-kinetics of type-I IFN between the lethal and non-lethal infections might be due to the differences of mainly produci.