Extended Data Figure 2: Autophagy in Xbp1-deficient intestinal epithelial cells is induced by eIF2α phosphorylation but independent of JNK or oxidative stress pathways.

a, Timeline for salubrinal experiment shown in Fig. 1j. b, Representative indirect immunofluorescence images of GFP–LC3 punctae accumulation (green) in small intestinal sections co-stained with anti-lysozyme antibody (red) treated as in a (n = 3). DAPI, blue. Scale bars, 10 μm. c, Immunoblot of crypt lysates after 3 days of tamoxifen or vehicle administration (n = 3). d, e, Immunoblot of primary intestinal epithelial cell scrapings (d); densitometry in e (n = 3; unpaired Student’s t-test; mean ± s.e.m.). f, g, Immunoblot of shXbp1 and shCtrl MODE-K cells. ER stress-induced Jun N-terminal kinase-1 (JNK1) has previously been connected in other cellular model systems to autophagy activation through phosphorylation of B-cell leukaemia 2 (Bcl-2) and its dissociation from beclin 1⁴³, as have oxidative stress/free radicals and haem oxygenase-1 (HO-1) activation⁴⁴. h, Intracellular ROS determined by dichlorofluorescein assay and mean fluorescent intensity (MFI) after vehicle or dichlorofluorescein diacetate (DCF-DA) treatment. i, Immunoblot of shXbp1 and shCtrl MODE-K cells after administration of the JNK inhibitor SP600125 (0, 5 or 25 μM) for 4 h. Note the absence of an effect of SP600125 treatment on the conversion of LC3-I to LC3-II or the levels of p-eIF2α, thereby excluding a major contribution of the JNK pathway to autophagy induction in the presence of intestinal-epithelial-cell-associated XBP1 deficiency. j, Immunoblot of shXbp1 and shCtrl MODE-K cells after N-acetylcysteine (NAC), glutathione (GSH) or vehicle treatment for 16 h. Note the absence of an effect of the free radical scavengers on either of these markers of UPR-induced autophagy (LC3-II or p-eIF2α). Results represent three (f–j) independent experiments. *P < 0.05.