Misfolded protein overload in the endoplasmic reticulum (ER) is a frequent cause of ER-stress. Many degenerative and metabolic disorders are caused by chronic stress in this organelle that result in the death of irreplaceable and essential cells. In spite of its pathological significance, our understanding of the underlying cell death mechanisms remains rudimentary. What we know is that, in order for cells to die, the stressed ER must signal to the nucleus or the mitochondria. That is because “apoptotic cell death” requires expression of genes encoded in the nucleus, or post-translational modification of proteins that act on the mitochondria. Even “necrotic cell death,” which was once thought to be a passive process, was recently shown to require dedicated signaling proteins.1 In terms of signaling pathways activated from the stressed ER, particularly well-understood are the three branches of the unfolded protein response (UPR) 2 (Fig. 1). One of them is initiated by IRE1, which activates its RNase function upon binding to misfolded peptides in the ER and catalyzes an unconventional splicing of the XBP1 mRNA. This generates an active XBP1 transcription factor that induces a number of ER quality control genes. In addition to activating XBP1, IRE1 can activate Jun N-terminal kinase (JNK). In certain experimental settings, this IRE1-JNK axis can induce apoptosis.3 However, IRE1α-knockout mouse embryonic fibroblasts are not more susceptible to ER stress-induced apoptosis.4,5 Therefore, the IRE1-JNK-induced cell death most likely occurs in a context-specific manner. Another UPR branch is initiated by PERK, a kinase that activates the transcription factor ATF4. Although ATF4 induces many quality control genes necessary for survival, it also induces CHOP, a transcription factor that aggravates stress in the ER by inducing oxidative protein folding enzymes and thereby increasing reactive oxygen species (ROS) production in the ER. In macrophages, such conditions lead to Ca2+ leakage, Cam kinase II activation and JNK activation for apoptosis.5 A third branch of the UPR is mediated by an ER-tethered transcription factor ATF6, which is thought to induce chaperone expression, and is not implicated in cell death regulation. Recently, we began using Drosophila to understand how cells die in response to excessive overload of mutant rhodopsin-1 that causes ER stress.6 In using Drosophila, our goal was to bring improvements in a number of aspects. We felt the need to perform unbiased screens. Also, it is important to establish physiological significance. This is because there may be many different ways of killing cells, depending upon an experimental setup, but not all those mechanisms may be relevant to the diseases we are interested in. There is also likely a cell type difference: a neuron may succumb to ER stress in a way that is different from macrophages or fibroblasts. With these in mind, we performed an RNAi screen and found CDK5 as a gene required for cell death in response to mutant rhodopsin-1 expression in the fly eye. In an independent screen, we found MEKK1, an upstream regulator of JNK, also required for cell death. Finally, we tested their pathological significance by inactivating CDK5 or MEKK1 in a Drosophila model for autosomal dominant retinitis pigmentosa (ADRP), where an endogenous allele of rhodopsin-1 causes ER stress that underlies age-related retinal degeneration.7 Interestingly, we found that the CDK5/MEKK1 pathway is not part of the known UPR pathways (Fig. 1): loss of CDK5 or MEKK1 neither affected XBP1 mRNA splicing nor influenced ATF4 activation. Conversely, mutations in IRE1 or ATF4 did not block apoptosis by mutant rhodopsin-1. How is this pathway regulated, and what does it do? As ROS and Ca2+ are known upstream activators of mammalian CDK5, these molecules are good candidates for linking the stressed ER to CDK5. We would like to make it clear that ER-stress-induced cell death mechanisms may vary depending upon the cell type and stress conditions. CDK5 is a kinase primarily active in post-mitotic cells, which are normally resistant to apoptosis. Therefore, cells with no endogenous CDK5 activity will not die through this pathway. What is the consequence of CDK5/MEKK1 activation? We note that many cells become resistant to death by turning their pro-apoptotic loci into a heterochromatin-like state.8 At the same time, MEKK1 is known to affect heterochromatin structures.9 Thus, we speculate that, perhaps the main function of CDK5/MEKK1 signaling is to remodel the chromatin structure of normally apoptosis-resistant post-mitotic cells, to make their pro-apoptotic loci more sensitive for activation. Figure 1. Signaling pathways from the stressed ER. The green arrow indicates the three known Unfolded Protein Response pathways mediated by transcription factors ATF4, ATF6 and XBP1. The predominant effect of these pathways is to induce quality ...