PERK 1,2 . By contrast, the effects of CDDO on CHOP were not blunted by any individual EIF2α kinase deficiency (Extended Data Fig. 1g) , possibly owing to the limited specificity of related compounds 11 . The comparative interrogation of CHOP regulators following three distinct cellular insults allowed us to differentiate global regulators of CHOP biology (Extended Data Fig. 1h-m) from such selectively operating in the context of CCCP-induced mitochondrial depolarization (Extended Data Figs. 2-4). In particular, stringent filtering for genes that prominently scored with CCCP, but not TM or CDDO, highlighted the transcriptional regulators TAF4 and GABPB1, glycolysis factors SLC2A1 and TPI1, and RNA binding proteins RBM27 and CLUH (KIAA0664). Moreover, the signature contained the mitochondrial proteins ATP5IF1 (ATPIF1) and OMA1. Most strikingly, it revealed a strong requirement for HRI (EIF2AK1) and the scarcely studied protein DELE1 (KIAA0141) (Fig. 1a, Extended Data Fig. 5a-b). Cellular dynamics of DELE1Given the scant knowledge on DELE1 and the unexpected involvement of HRI, we first sought to validate their requirement in a panel of cell systems including non-transformed cells, and indeed could confirm their importance in all cases (Extended Data Fig. 5c). Furthermore, CHOP induction also depended on DELE1 and HRI for other types of mitochondrial stress, including inhibition of complex V (oligomycin), TRAP1 (GTPP), and genetic ablation of LONP1 (Extended Data Fig. 5d-f). Failure to induce CHOP after stimulation with CCCP was preceded by a defect in EIF2α phosphorylation in HRI-or DELE1-deficient cells, suggesting that like HRI, DELE1 operates upstream of this event (Extended Data Fig. 5g). Strikingly, expression of HRI in DELE1 knockout cells was able to partially restore CHOP induction, whereas DELE1 expression in HRI-deficient cells was unproductive (Fig. 2a, Extended Data Fig. 6a-b). This indicated that DELE1 requires HRI to trigger CHOP but not vice versa, suggesting that DELE1 may act upstream of both EIF2α and HRI. Given that DELE1 is a mitochondrial protein 12 (Extended Data Fig. 6c), whereas HRI resides in the cytoplasm, we next wondered whether the activity of DELE1 towards HRI might be regulated by its localization. To test this hypothesis, we investigated if artificially rerouting DELE1 to the cytosol would bypass the need for a mitochondrial insult to provoke CHOP expression. Indeed, expression of a DELE1 mutant lacking the mitochondrial targeting sequence (DELE1 ∆MTS ) yielded a predominantly cytoplasmic protein that readily induced CHOP expression independently of CCCP (Fig. 2b-c, Extended Data Fig. 6d-e). This constitutively active version of DELE1 still required HRI to induce CHOP, underscoring its likely role as an activator of HRI. Based on these findings, we asked whether the activity of wild-type DELE1 might be regulated via a similar mechanism. Indeed, while DELE1 localized to mitochondria in unperturbed cells, it could be detected in the cytosol upon CCCP treatment (Fig. 2d). We did not ob...

The gene regulatory network governing anterior–posterior axis formation in Drosophila is a well-established paradigm to study transcription in developmental biology. The rapid temporal dynamics of gene expression during early stages of development, however, are difficult to track with standard techniques. We optimized the bright and fast-maturing fluorescent protein mNeonGreen as a real-time, quantitative reporter of enhancer expression. We derive enhancer activity from the reporter fluorescence dynamics with high spatial and temporal resolution, using a robust reconstruction algorithm. By comparing our results with data obtained with the established MS2-MCP system, we demonstrate the higher detection sensitivity of our reporter. We used the reporter to quantify the activity of variants of a simple synthetic enhancer, and observe increased activity upon reduction of enhancer–promoter distance or addition of binding sites for the pioneer transcription factor Zelda. Our reporter system constitutes a powerful tool to study spatio-temporal gene expression dynamics in live embryos.

Hunchback (Hb) is considered a context-dependent transcription factor, able to activate or repress different enhancers during Drosophila Melanogaster embryo segmentation. The mechanism driving the context dependent activity of Hb is however not well understood. Here, we design 20 synthetic enhancers to elucidate the effect of Hb binding sites in Drosophila segmentation and quantitatively measure their activity. We obtain the spatiotemporal activity dynamics of all synthetic enhancers in-vivo, by using a quantitative and sensitive reporter system that we recently developed. Our data reveal a dual role for Hb binding sites in segmentation enhancers: on one hand, Hb act as a typical short range repressor by binding to its cognate sequences; on the other hand, we report a novel effect of a sequence containing multiple Hb binding sites, which is able to increase enhancer activity independently from Hb binding. This sequence, which contains multiple Poly-dA stretches, increases the activity of enhancers driven by different activators, possibly by disfavoring nucleosome occupancy.