Proximity Graph Networks: Predicting Ligand Affinity with Message Passing Neural Networks.

Message passing neural networks (MPNNs) on molecular graphs generate continuous and differentiable encodings of small molecules with state-of-the-art performance on protein-ligand complex scoring tasks. Here, we describe the proximity graph network (PGN) package, an open-source toolkit that constructs ligand-receptor graphs based on atom proximity and allows users to rapidly apply and evaluate MPNN architectures for a broad range of tasks. We demonstrate the utility of PGN by introducing benchmarks for affinity and docking score prediction tasks. Graph networks generalize better than fingerprint-based models and perform strongly for the docking score prediction task. Overall, MPNNs with proximity graph data structures augment the prediction of ligand-receptor complex properties when ligand-receptor data are available.

DSFworld: A flexible and precise tool to analyze differential scanning fluorimetry data.

Differential scanning fluorimetry (DSF) is a method to determine the apparent melting temperature (Tma) of a purified protein. In DSF, the raw unfolding curves from which Tma is calculated vary widely in shape and complexity. However, the tools available for calculating Tma are only compatible with the simplest of DSF curves, hindering many otherwise straightforward applications of the technology. To overcome this limitation, we designed new mathematical models for Tma calculation that accommodate common forms of variation in DSF curves, including the number of transitions, the presence of high initial signal, and temperature-dependent signal decay. When tested these models against DSFbase, an open-source database of 6235 raw, real-life DSF curves, these models outperformed the existing standard approaches of sigmoid fitting and maximum of the first derivative. To make these models accessible, we created an open-source software and website, DSFworld (https://gestwickilab.shinyapps.io/dsfworld/). In addition to these improved fitting capabilities, DSFworld also includes features that overcome the practical limitations of many analysis workflows, including automatic reformatting of raw data exported from common qPCR instruments, labeling of data based on experimental variables, and flexible interactive plotting. We hope that DSFworld will enable more streamlined and accurate calculation of Tma values for DSF experiments.

Protein-adaptive differential scanning fluorimetry using conformationally responsive dyes.

Differential scanning fluorimetry (DSF) is a technique that reports protein thermal stability via the selective recognition of unfolded states by fluorogenic dyes. However, DSF applications remain limited by protein incompatibilities with existing DSF dyes. Here we overcome this obstacle with the development of a protein-adaptive DSF platform (paDSF) that combines a dye library 'Aurora' with a streamlined procedure to identify protein-dye pairs on demand. paDSF was successfully applied to 94% (66 of 70) of proteins, tripling the previous compatibility and delivering assays for 66 functionally and biochemically diverse proteins, including 10 from severe acute respiratory syndrome coronavirus 2. We find that paDSF can be used to monitor biological processes that were previously inaccessible, demonstrated for the interdomain allostery of O-GlcNAc transferase. The chemical diversity and varied selectivities of Aurora dyes suggest that paDSF functionality may be readily extended. paDSF is a generalizable tool to interrogate protein stability, dynamics and ligand binding.

Conformationally responsive dyes enable protein-adaptive differential scanning fluorimetry.

Flexible in vitro methods alter the course of biological discoveries. Differential Scanning Fluorimetry (DSF) is a particularly versatile technique which reports protein thermal unfolding via fluorogenic dye. However, applications of DSF are limited by widespread protein incompatibilities with the available DSF dyes. Here, we enable DSF applications for 66 of 70 tested proteins (94%) including 10 from the SARS-CoV2 virus using a chemically diverse dye library, Aurora, to identify compatible dye-protein pairs in high throughput. We find that this protein-adaptive DSF platform (paDSF) not only triples the previous protein compatibility, but also fundamentally extends the processes observable by DSF, including interdomain allostery in O-GlcNAc Transferase (OGT). paDSF enables routine measurement of protein stability, dynamics, and ligand binding.

A campaign targeting a conserved Hsp70 binding site uncovers how subcellular localization is linked to distinct biological activities.

The potential of small molecules to localize within subcellular compartments is rarely explored. To probe this question, we measured the localization of Hsp70 inhibitors using fluorescence microscopy. We found that even closely related analogs had dramatically different distributions, with some residing predominantly in the mitochondria and others in the ER. CRISPRi screens supported this idea, showing that different compounds had distinct chemogenetic interactions with Hsp70s of the ER (HSPA5/BiP) and mitochondria (HSPA9/mortalin) and their co-chaperones. Moreover, localization seemed to determine function, even for molecules with conserved binding sites. Compounds with distinct partitioning have distinct anti-proliferative activity in breast cancer cells compared with anti-viral activity in cellular models of Dengue virus replication, likely because different sets of Hsp70s are required in these processes. These findings highlight the contributions of subcellular partitioning and chemogenetic interactions to small molecule activity, features that are rarely explored during medicinal chemistry campaigns.

Discovery and Optimization of Pyrazole Amides as Inhibitors of ELOVL1.

Accumulation of very long chain fatty acids (VLCFAs) due to defects in ATP binding cassette protein D1 (ABCD1) is thought to underlie the pathologies observed in adrenoleukodystrophy (ALD). Pursuing a substrate reduction approach based on the inhibition of elongation of very long chain fatty acid 1 enzyme (ELOVL1), we explored a series of thiazole amides that evolved into compound 27─a highly potent, central nervous system (CNS)-penetrant compound with favorable in vivo pharmacokinetics. Compound 27 selectively inhibits ELOVL1, reducing C26:0 VLCFA synthesis in ALD patient fibroblasts, lymphocytes, and microglia. In mouse models of ALD, compound 27 treatment reduced C26:0 VLCFA concentrations to near-wild-type levels in blood and up to 65% in the brain, a disease-relevant tissue. Preclinical safety findings in the skin, eye, and CNS precluded progression; the origin and relevance of these findings require further study. ELOVL1 inhibition is an effective approach for normalizing VLCFAs in models of ALD.

Electrochemically Enabled Chan-Lam Couplings of Aryl Boronic Acids and Anilines.

The Chan-Lam reaction remains a highly utilized transformation for C-N bond formation. However, anilines remain problematic substrates due to their lower nucleophilicity. To address this problem, we developed an electrochemically mediated Chan-Lam coupling of aryl boronic acids and amines utilizing a dual copper anode/cathode system. The mild conditions identified have enabled the preparation of a wide range of functionalized biarylanilines in good yields and chemoselectivities.