Toll-like receptor (TLR) 9 requires proteolytic processing in the endolysosome to initiate signaling in response to DNA. However, recent studies conflict as to which proteases are required for receptor cleavage. We show that TLR9 proteolysis is a multistep process. The first step removes the majority of the ectodomain and can be performed by asparagine endopeptidase (AEP) or cathepsin family members. This initial cleavage event is followed by a trimming event that is solely cathepsin mediated and required for optimal receptor signaling. This dual requirement for AEP and cathepsins is observed in all cell types that we have analyzed, including mouse macrophages and dendritic cells. In addition, we show that TLR7 and TLR3 are processed in an analogous manner. These results define the core proteolytic steps required for TLR9 function and suggest that receptor proteolysis may represent a general regulatory strategy for all TLRs involved in nucleic acid recognition.

High throughput screening (HTS) is routinely used to identify bioactive small molecules. This requires physical compounds, which limits coverage of accessible chemical space. Computational approaches combined with vast on-demand chemical libraries can access far greater chemical space, provided that the predictive accuracy is sufficient to identify useful molecules. Through the largest and most diverse virtual HTS campaign reported to date, comprising 318 individual projects, we demonstrate that our AtomNet® convolutional neural network successfully finds novel hits across every major therapeutic area and protein class. We address historical limitations of computational screening by demonstrating success for target proteins without known binders, high-quality X-ray crystal structures, or manual cherry-picking of compounds. We show that the molecules selected by the AtomNet® model are novel drug-like scaffolds rather than minor modifications to known bioactive compounds. Our empirical results suggest that computational methods can substantially replace HTS as the first step of small-molecule drug discovery.

The vanilloid receptor VR1 has attracted great interest as a sensory transducer for capsaicin, protons, and heat, and as a therapeutic target. Here we characterize two novel VR1 antagonists, KJM429 [N-(4-tert-butylbenzyl)-N′-[4-(methylsulfonylamino)benzyl]thiourea] and JYL1421 [N-(4-tert-butylbenzyl)-N′-[3-fluoro-4-(methylsulfonylamino)benzyl]thiourea], with enhanced activity compared with capsazepine on rat VR1 expressed in Chinese hamster ovary (CHO) cells. JYL1421, the more potent of the two novel antagonists, inhibited [³H]resiniferatoxin binding to rVR1 with an affinity of 53.5 ± 6.5 nM and antagonized capsaicin-induced calcium uptake with an EC₅₀ of 9.2 ± 1.6 nM, reflecting 25- and 60-fold greater potencies than capsazepine. Both JYL1421 and KJM429 antagonized RTX as well as capsaicin and their mechanism was competitive. The responses to JYL1421 and KJM429 differed for calcium uptake by rVR1 induced by heat or pH. JYL1421 antagonized the response to both pH 6.0 and 5.5, whereas KJM429 antagonized at pH 6.0 but was an agonist at lower pH (<5.5). For heat, JYL1421 fully antagonized and KJM429 partially antagonized. Capsazepine showed only weak antagonism for both pH and heat. Responses of rVR1 to different activators could thus be differentially affected by different ligands. In cultured dorsal root ganglion neurons, JYL1421 and KJM429 likewise behaved as antagonists for capsaicin, confirming that the antagonism is not limited to heterologous expression systems. Finally, JYL1421 and KJM429 had little or no effect on ATP-induced calcium uptake in CHO cells lacking rVR1, unlike capsazepine. We conclude that JYL1421 is a competitive antagonist of rVR1, blocking response to all three of the agonists (capsaicin, heat, and protons) with enhanced potency relative to capsazepine.