ZIP-2/CEBP-2 translation-surveillance immunity in Caenorhabditis elegans

Caenorhabditis elegans can detect a pathogen by the damage it does to a core host process instead of by recognizing a microbial molecule. The best defined case is translation surveillance in the intestine. Pseudomonas aeruginosa Exotoxin A (ToxA) is endocytosed and ADP-ribosylates elongation factor 2, which inhibits host mRNA translation. Chemical translation inhibitors such as cycloheximide, and loss of other core host functions, trigger the same response. Enzymatically inactive ToxA does not. So the host senses the block in translation, not the toxin itself. Translation inhibition raises the level of the bZIP transcription factor ZIP-2 through a conserved upstream open reading frame (uORF) in the zip-2 mRNA. This fits a model in which a general fall in translation derepresses ZIP-2 synthesis. ZIP-2 acts with the C/EBP-gamma ortholog CEBP-2, a bZIP partner that binds it, to induce a set of intestinal infection-response genes. These include irg-1 and irg-2, which are induced independently of the PMK-1 p38 MAPK, DBL-1 TGF-beta and KGB-1 JNK pathways. zip-2 and cebp-2 mutants are more susceptible to P. aeruginosa and to ToxA. The molecular functions of the IRG-1 and IRG-2 proteins are unknown. Mitochondrial surveillance through ATFS-1 and the mitochondrial UPR is a parallel, separately regulated surveillance pathway.

Boundary. This is a concrete C. elegans module for the ZIP-2 branch of surveillance immunity: translation inhibition -> uORF-mediated derepression of ZIP-2 -> ZIP-2/CEBP-2-dependent transcription -> irg-1/irg-2 output. It has three parts. The biology is thin in one respect. The sensing step is carried by ZIP-2 itself, through a cis-acting uORF in its own mRNA, so part 1 and part 2 have the same protein. They are kept as separate parts because they are different mechanisms: translational control of ZIP-2 abundance in part 1, and ZIP-2/CEBP-2 transcription factor activity in part 2. Part 1 has no molecular function because the regulated entity is the zip-2 mRNA, not an activity of a protein. Exclusions and cross-references. GCN-2, the eIF2-alpha kinase of the worm integrated stress response, is not modeled as a member. No cached paper shows that GCN-2 or eIF2-alpha phosphorylation is needed for ZIP-2 induction, and the gcn-2 review places GCN-2 in amino-acid-starvation and mitochondrial-stress translational control. Whether the zip-2 uORF responds to eIF2-alpha phosphorylation or to a general fall in initiation is open (see knowledge_gaps). ATFS-1 and the mitochondrial UPR are a separate surveillance pathway that detects loss of mitochondrial import (PMID:25274306) and is not modeled here. ZIP-2 activation during aging is independent of ATFS-1 and CEBP-2 (PMID:32350153). ZIP-2 and CEBP-2 also take part in the ESRE network, which responds to mitochondrial damage in liquid-killing assays (PMID:28662060). That broader bZIP network is recorded as context. The PMK-1/ATF-7 pathway is modeled in MODULE:c_elegans_pmk1_intestinal_immunity. Both pathways protect against ToxA, but irg-1 and irg-2 induction does not need PMK-1. CEBP-2 also acts with ZIP-11 in a PMK-1-independent response (PMID:34804026), which is not modeled. Ontology. GO has no term for surveillance (effector-triggered) immunity in animals or for a response to host translation inhibition. The module therefore uses antibacterial innate immune response and defense response to Gram-negative bacterium as its process concepts. The surveillance character is described in text.