Constitutive proteasomes and immunoproteasomes shape the peptide repertoire presented by major histocompatibility complex class I (MHC-I) molecules by harboring different sets of catalytically active subunits. Here, we present the crystal structures of constitutive proteasomes and immunoproteasomes from mouse in the presence and absence of the epoxyketone inhibitor PR-957 (ONX 0914) at 2.9 Å resolution. Based on our X-ray data, we propose a unique catalytic feature for the immunoproteasome subunit b5i/LMP7. Comparison of ligand-free and ligand-bound proteasomes reveals conformational changes in the S1 pocket of b5c/X but not b5i, thereby explaining the selectivity of PR-957 for b5i. Time-resolved structures of yeast proteasome:PR-957 complexes indicate that ligand docking to the active site occurs only via the reactive head group and the P1 side chain. Together, our results support structure-guided design of inhibitory lead structures selective for immunoproteasomes that are linked to cytokine production and diseases like cancer and autoimmune disorders.

Proteolytic degradation is an essential cellular process which is primarily carried out by the 20S proteasome core particle (CP), a protease of 720 kDa and 28 individual subunits. As a result of its central functional role, the proteasome represents an attractive drug target that has been extensively investigated during the last decade and validated by the approval of bortezomib by the US Food and Drug Administration (FDA). Currently, several optimized second-generation proteasome inhibitors are being explored as anticancer drugs in clinical trials, and most of them target both constitutive proteasomes (cCPs) and immunoproteasomes (iCPs). However, selective inhibition of the iCPs, a distinct class of proteasomes predominantly expressed in immune cells, appears to be a promising therapeutic rationale for the treatment of autoimmune disorders. Although a few selective agents have already been identified, the recently determined crystal structure of the iCP will further promote the development and optimization of iCP-selective compounds.

Mammalian genomes encode seven catalytic proteasome subunits, namely, β1c, β2c, β5c (assembled into constitutive 20S proteasome core particles), β1i, β2i, β5i (incorporated into immunoproteasomes), and the thymoproteasome-specific subunit β5t. Extensive research in the past decades has yielded numerous potent proteasome inhibitors including compounds currently used in the clinic to treat multiple myeloma and mantle cell lymphoma. Proteasome inhibitors that selectively target combinations of β1c/β1i, β2c/β2i, or β5c/β5i are available, yet ligands truly selective for a single proteasome activity are scarce. In this work we report the development of cell-permeable β1i and β5i selective inhibitors that outperform existing leads in terms of selectivity and/or potency. These compounds are the result of a rational design strategy using known inhibitors as starting points and introducing structural features according to the X-ray structures of the murine constitutive and immunoproteasome 20S core particles.