P-eIF2α is required for heat shock SGs in mammals, but not in flies. (A) S2R+ cells were treated with indicated dsRNAs for 96 h, followed by 500 μm arsenite for 2 h or 40°C heat shock for 1 h, then hybridized with Cy3-Oligo-dT(30). Samples were blinded and photographed, and cells in each image were scored for the presence of SGs. (B) Western blot of parallel samples to the data shown in A. (C) Quantitative real-time PCR analysis of PEK and GCN2 mRNA levels in knockdown cells. Error bars in A and C represent one standard deviation from the average of three independent experiments. (D) eIF2α Ser51Ala MEFs (denoted “M”) were co-plated with human-derived U2OS cells (denoted “H”), then either exposed to heat shock at 48°C for 30 min (top panels), pre-treated with CHX at 50 μg/mL for 30 min, and then heat-shocked (center panels), or treated with 50 μM pateamine A (patA) for 1 h (bottom panels). Samples were fixed and stained with anti-eIF3 (green), anti-eIF4E (red), and Hoechst (DNA, blue). Mouse cells can be distinguished from human cells by the distinct nuclear speckling pattern in DNA panels. Bar, 25 μm. (E) A model of eIF2α-dependent and -independent translational arrest and SG formation in Drosophila and mammals. In Drosophila, arsenite results in the phosphorylation of eIF2α via the kinase PEK, whereas in mammals the critical arsenite responsive kinase is HRI. In both cases, P-eIF2α inhibits translation initiation, and the initiation complex, associated mRNAs and RNA binding proteins are recruited to SGs. However, the formation of Drosophila heat shock SGs does not depend on P-eIF2α, whereas P-eIF2α is required for mammalian heat shock SGs. The mammalian eIF2α kinase(s) activated by heat shock is unknown.