Epoxomicin, which is a member of linear peptide epoxyketone natural product family, was first discovered by Bristol-Myers Squibb (BMS, Tokyo, Japan) from an unidentified Actinomycetes strain, No. Q996-17, based on its in vivo antitumor activity against murine B16 melanoma tumors. Despite its promising antitumor activity, epoxomicin did not draw much attention as a potential drug lead from the pharmaceutical industry, presumably because epoxomicin has seemingly poor drug-like properties, namely the presence of a peptide backbone and the labile epoxy ketone pharmacophore (). More importantly, at that time its mode of antitumor action was unknown.

Our initial interest in epoxomicin stemmed from its potent antitumor activity. We hoped to use the compound as a molecular probe to uncover novel signaling pathways that are crucial for tumor growth. Our first step towards this goal was to develop a strategy for identifying its intracellular target. Looking at the chemical structure of epoxomicin, we speculated that the epoxide ring, which is prone to nucleophilic attack, would aid in expediting the identification of epoxomicin-binding proteins via their covalent modification by the compound. The idea of covalently labeling target proteins by epoxomicin led us to design a biotin-tagged epoxomicin analog for use as an affinity reagent (). As had been done with natural products,^, our goal was to use biotin-tagged epoxomicin to label target proteins with biotin through an irreversible linkage, enabling target proteins to be efficiently pulled down via an avidin-coated resin. Finally, isolated target proteins could be eluted from the resin and visualized by SDS-PAGE in conjunction with staining and identified by western blotting as well as mass spectrometry ().

While investigating the activity of epoxomicin against the common proteases, we quickly realized that epoxomicin possesses unprecedented selectivity for the proteasome over previously known proteasome inhibitors, which commonly demonstrate cross-inhibition with non-proteasomal proteases. This unusually high specificity was rather surprising because we originally speculated that the inhibition of the proteasome by epoxomicin was due to a nucleophilic attack on the epoxide ring by the hydroxyl group of N-terminal threonine of the proteasome, resulting in the formation of an ether linkage (), as do proteasome inhibitors containing reactive electrophilic moieties. To our surprise, however, X-ray crystallographic studies performed in collaboration with Michael Groll and Robert Huber revealed that the unique specificity of epoxomicin is a result of the irreversible formation of a six-membered morpholino ring between the amino terminal catalytic Thr-1 of the 20S proteasome and the α’, β’-epoxy ketone pharmacophore of epoxomicin (). Mechanistically, this unusual ring formation is explained by a sequential double nucleophilic attack on the epoxy ketone residue of epoxomicin by two nucleophiles of the N-terminal Thr, the hydroxyl side chain and the free amine. Therefore, it is expected that only Ntn (N-terminal nucleophile) proteases such as the proteasome can form the morpholino ring by reaction with epoxy ketones, thereby rendering epoxomicin exceptionally specific for the proteasome over alternative mechanism-based proteasome inhibitors.

This unusually high specificity and potency led us to believe that epoxomicin may have promising potential for cancer treatment. Initially, there was some skepticism about whether the proteasome can be a viable anticancer target considering its essential functions in all eukaryotic cells. Contrary to this assumption, some studies began to suggest that a reasonable therapeutic window in which one can selectively kill cancer cells might exist.^, It is now known that rapidly proliferating cancer cells generally display elevated levels of proteasome activity and increased cell death upon proteasomal inhibition compared to normal cells. In order to understand and improve epoxomicin's anticancer activity, we put forth an effort to optimize the compound via a classical medicinal chemistry approach. This effort eventually led us to YU-101 (), which showed more potent inhibitory activity against the proteasome than both epoxomicin and bortezomib, a boronic acid-based synthetic proteasome inhibitor developed initially by the biotech company Proscript and later acquired and developed further by Millennium to the yield the FDA-approved drug Velcade®. While YU-101 was being developed, bortezomib rapidly progressed into clinical trials for the treatment of multiple myeloma. At that time, it was already clear that YU-101 had better proteasome inhibitory and antitumor activities than bortezomib. Importantly, it is now shown to be substantially more selective for the proteasome than bortezomib. Indeed, it was later found that severe and dose-limiting peripheral neuropathy associated with bortezomib is most likely due to cross-inhibition with HtrA2/Omi, a neuronal serine protease implicated in neuron survival. Carfilzomib, a YU-101 derivative developed by the South San Francisco-based biotech company Proteolix, does not inhibit serine proteases including HtrA2/Omi, most likely due to the properties of its epoxomicin-derived pharmacophore.