ATP-Competitive Partial Antagonists—'PAIR's—Segregate IRE1α’s RNase-Mediated Biological Outputs

During mammalian cell growth and differentiation, the unfolded protein response (UPR) homeostatically adjusts endoplasmic reticulum (ER) protein-folding capacity to match changing cellular secretory demands. However, under high/chronic ER stress conditions the UPR triggers apoptosis. This dichotomy is promoted in part by differential activation levels of the ER transmembrane kinase/endoribonuclease (RNase) IRE1a. IRE1a kinase auto-phosphorylation operates as a rheostat to control downstream RNase-induced outputs that either sustain adaptive ER protein-folding or cause apoptosis. We have previously shown that IRE1a’s RNase activity can be activated or fully inactivated by ATP-competitive kinase inhibitors. Here we developed a new class of ATP-competitive kinase inhibitors that partially antagonize the RNase of IRE1a at full occupancy. An atomic level resolution co-crystal structure shows that these small molecule partial antagonists—which we named ‘PAIR’s for (Partial Antagonists of IRE1a RNase)—occupy the ATP-binding site of IRE1a and partially displace a helix (helix aC) in the kinase domain that forms part of a dimeric interface. In insulin-producing beta cells, PAIRs permit adaptive XBP1 mRNA splicing, while quelling destructive/terminal outputs from extra-XBP1 mRNA endonucleolytic decay, thus preventing apoptosis. Preservation of XBP1 splicing by PAIRs permits B lymphocytes to differentiate into immunoglobulin-producing plasma cells. In summary, we propose that an intermediate RNase-inhibitory “sweet spot,” achieved by PAIR-bound IRE1a, may capture a structural conformation naturally available to IRE1a that could represent a desirable therapeutic state for drugging this master UPR sensor/effector. Rat INS1 cells with doxycycline-inducible mouse IRE1α expression were used in the study. IRE1α overexpression was induced with doxycycline and treated with small molecule modulators of IRE1α kinase activity (vehicle, PAIR1 or KIRA9). Control samples were treated only to drug vehicles (i.e., received neither doxycycline nor kinase modulators). Total RNA was collected at 4 hours or 24 hours after doxycyline induction. Experiments were performed in triplicate.