Assay optimization and shRNA screen design. (A) Y. enterocolitica WA inhibits NF-κB signaling through TNF-R. RE-luc2P-HEK293 cells were infected with Y. enterocolitica WA, at either MOI 0 (circles), 1 (square), or 5 (diamonds), in a 96-well plate. Cells were stimulated either 5 h post infection with 10 ng/ml TNF-α and incubated for an additional 18 h (closed symbols) or 2 h post-infection with 10 ng/ml TNF-α and incubated for another 5 hrs (open symbols). The arbitrary luciferase activity per well from a representative of two experiments (n=10/expt) is presented. Z’ was calculated using the SD and mean of luciferase activity from cells infected with Y. enterocolitica WA at MOI 5 versus cells not treated with bacteria (MOI 0) at each time point . The best Z’ value 0.65 was obtained for the 18 h time point at MOI 5. (B) For the shRNA screen, the kinome plasmid library was transfected in 96 well format, and cells were subjected to puromycin selection to enrich for populations expressing the inhibitory sequences. Chloramphenicol (170 μg/ml) was added 1 h post-infection to control extracellular bacteria counts. At 5 h post-infection, 10 ng/ml TNF-α was added to the cells and NF-κB-driven luciferase activity was determined 18 h later. (C) The hit selection cut-off was determined as ≥40% direct recovery in luciferase signal of Yersinia-infected cells (black squares) relative to non-hits (gray squares) and bacteria free samples (light gray diamonds). (D) The statistical significance of assay hit selection was evaluated using a standard z-score. Genes in which silencing resulted in assay reads with a score ≥3 standard deviations above the assay mean score were considered to be true hits with a strong effect on Yersinia-driven inhibition of NF-κB signaling (shown in black diamonds), compared to non-hits (gray diamonds).