Fig 1.
G6PD deficiency increases the replication level of coronavirus via down-regulation of TNF-α-induced COX-2 expression and its downstream metabolite PGE2 production in A549 cells.
(A) Scramble control and G6PD-kd A549 were treated with 15 ng/ml TNF-α for the indicated time, and the expression of COX-2 mRNA was investigated by quantitative PCR. Data are reported as the means ±SD, n = 3. *p<0.05. (B) Scramble control and G6PD-kd A549 were treated with 15 ng/ml TNF-α for 24 h. COX-2 promoter activity was determined by the luciferase assay. Data are reported as the means ±SD, n = 3. *p<0.05. (C) The expression level of COX-2 protein upon 15 ng/ml TNF-α treatment at different time courses was shown, and β-actin was present as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *p<0.05 and **p<0.01 represent levels of significant difference when comparing scramble control with TNF-α treatment at the corresponding time points. (D) PGE2 secretion by 15 ng/ml TNF-α stimulation was detected by ELISA. Data are reported as the means ±SD, n = 3. *p<0.05. (E) Upper panel: Scramble control and G6PD-kd A549 cells were infected with coronavirus (0.1 MOI) for 8 h, and the infected cells were harvested for analyzing viral mRNA expression. Data are reported as the fold change normalized to infected scramble control cells. Data are reported as the means ±SD, n = 3. *p<0.05. Lower panel: Scramble control and G6PD-kd A549 cells were infected with HCoV-229E (0.1 MOI) for 24 h then viral particle was harvested and virus titer was determined using plaque assay. (F) Scramble control and G6PD-kd A549 cells were infected with coronavirus (0.1 MOI) for 8 h upon 15 ng/ml TNF-α with or without 10 μM celecoxib co-pretreatment, and the infected cells were harvested for analyzing viral mRNA expression. Data are reported as the fold normalized to infected control cells. Data are reported as the means ±SD, n = 3. *p<0.05.
Fig 2.
G6PD knockdown impairs the phosphorylation of MAPKs signaling.
(A, B) The expression level of COX-2 was assessed by western blot in scramble control and G6PD-kd A549 cells treated with 15 ng/ml TNF-α or combined with MAPK inhibitor (SB203580, p38 inhibitor; U0126, MEK1/2 inhibitor) pre-treatment for 3 h. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *,#p<0.05 indicate significant difference between cells with or without inhibitor pretreatment upon TNF-α stimulation. (C, D) The phosphorylation level of p38 (C), ERK1/2 (D) were determined by western blot in scramble control and G6PD-kd A549 cells stimulated with 1.5 or 15 ng/ml TNF-α in different time courses. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *p<0.05 indicates significant difference comparing scramble control with TNF-α treatment at the corresponding time points.
Fig 3.
G6PD knockdown dysregulates the activation of c-JUN and NF-κB signaling.
(A) Scramble control and G6PD-kd A549 cells were pretreated with Tanshinone IIA (TSIIA), AP-1 inhibitor for 2 h and then treated with 15 ng/ml TNF-α for 3 h. The expression level of COX-2 was assessed by western blotting assay. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *,#p<0.05 indicate significant difference between cells with or without inhibitor pretreatment upon TNF-α stimulation. (B) The expression level of COX-2 was determined by western blot under 15 ng/ml TNF-α stimulation or combined with pre-treatment of NF-κB inhibitor, Helenalin, for 3 h in scramble control and G6PD-kd A549 cells. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *,#p<0.05 indicate significant difference between cells with or without inhibitor pretreatment upon TNF-α stimulation. (C) The phosphorylation level of c-JUN was determined by western blot in scramble control and G6PD-kd A549 cells stimulated with 1.5 or 15 ng/ml TNF-α in different time courses. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *p<0.05 indicates significant difference between scramble control and G6PD-kd cells upon TNF-α treatment at the corresponding time points. (D) The phosphorylation levels of p65 and IκBα were investigated in scramble control and G6PD-kd A549 cells upon TNF-α treatment (15 ng/ml) in different time courses. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *p<0.05 and **p<0.01 represent levels of significant difference between scramble control and G6PD-kd cells upon TNF-α treatment at the corresponding time points.
Fig 4.
TNF-α-triggered p38 MAPK activation and COX-2 expression are mediated by NOX signaling.
(A) The mRNA level of NOX2 was determined in universal negative control (NC) or NOX2-targeting siRNA transfected scramble control and G6PD-kd A549 cells. After transient transfection, cells were harvested for analyzing NOX2 mRNA expression. Data are the means ±SD, n = 3. *p<0.05 indicates significant difference between cells with or without NOX2 siRNA pretreatment. (B) The phosphorylation level of p38 was determined in scramble control and G6PD-kd A549 cells upon 15 ng/ml TNF-α stimulation combined with pre-treatment of universal negative control (NC) or NOX2-targeting siRNA. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *,#p<0.05 indicate significant difference between cells with or without NOX2 siRNA pretreatment upon TNF-α stimulation. (C) The expression level of COX-2 was determined under 15 ng/ml TNF-α stimulation for 3 h in universal negative control (NC) or NOX2-targeting siRNA transfected scramble control and G6PD-kd A549 cells. β-Actin expression was shown as the loading control. Numbers represent the relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *,#p<0.05 indicate significant difference between cells with or without NOX2 siRNA pretreatment upon TNF-α stimulation.
Fig 5.
G6PD is required for the TNF-α-induced activation of NOX signaling and antiviral response.
(A) The activity of NOX was measured by a lucigenin chemiluminescence assay. Data are the means ±SD, n = 3. *p<0.05. (B) The mRNA level of NOX subunits were determined by quantitative PCR in scramble control and G6PD-kd A549 cells. Data are the means ±SD, n = 3. *p<0.05. (C) The expression level of NOX subunits were determined by western blot in scramble control and G6PD-kd A549 cells. β-Actin expression was shown as the loading control. Numbers represent relative fold differences of protein levels on the basis of densitometer quantitation. Data are means ±SD of three separate experiments, *p<0.05 vs. scramble control. (D) Scramble control and G6PD-kd A549 cells were infected with coronavirus (0.1 MOI) for 8 h upon universal negative control (NC) or NOX2-targeting siRNA pretreatment, and the infected cells were harvested for analyzing viral mRNA expression. Data are the means ±SD, n = 3. *p<0.05.
Fig 6.
Proposed schematic representation of the signaling pathway involved in the TNF-α-induced NOX/MAPK/c-JUN/NF-κB/COX-2 signaling impaired by G6PD knockdown in A549 cells.
Physiologically, stimulus by a low dosage of TNF-α causes cellular ROS production through the activation of NOX and downstream signaling (MAPK/NF-κB). The activated NOX/MAPK/c-JUN/NF-κB signaling is concomitant with the increased expression of COX-2 and production of PGE2. G6PD knockdown results in the reduced generation of cellular NADPH and an impairment of TNF-α-induced NOX activation. Consequently, COX-2 expression and PGE2 production are also less in G6PD-knockdown cells, suggesting the participation of G6PD in the TNF-α-induced inflammatory response against viral infection.