Discovery of DNL343: A Potent, Selective, and Brain-Penetrant eIF2B Activator Designed for the Treatment of Neurodegenerative Diseases.

Eukaryotic translation initiation factor 2B (eIF2B) is a key component of the integrated stress response (ISR), which regulates protein synthesis and stress granule formation in response to cellular insult. Modulation of the ISR has been proposed as a therapeutic strategy for treatment of neurodegenerative diseases such as vanishing white matter (VWM) disease and amyotrophic lateral sclerosis (ALS) based on its ability to improve cellular homeostasis and prevent neuronal degeneration. Herein, we report the small-molecule discovery campaign that identified potent, selective, and CNS-penetrant eIF2B activators using both structure- and ligand-based drug design. These discovery efforts culminated in the identification of DNL343, which demonstrated a desirable preclinical drug profile, including a long half-life and high oral bioavailability across preclinical species. DNL343 was progressed into clinical studies and is currently undergoing evaluation in late-stage clinical trials for ALS.

Discovery of a First-in-Class Small-Molecule Ligand for WDR91 Using DNA-Encoded Chemical Library Selection Followed by Machine Learning.

WD40 repeat-containing protein 91 (WDR91) regulates early-to-late endosome conversion and plays vital roles in endosome fusion, recycling, and transport. WDR91 was recently identified as a potential host factor for viral infection. We employed DNA-encoded chemical library (DEL) selection against the WDR domain of WDR91, followed by machine learning to predict ligands from the synthetically accessible Enamine REAL database. Screening of predicted compounds identified a WDR91 selective compound 1, with a KD of 6 ± 2 µM by surface plasmon resonance. The co-crystal structure confirmed the binding of 1 to the WDR91 side pocket, in proximity to cysteine 487, which led to the discovery of covalent analogues 18 and 19. The covalent adduct formation for 18 and 19 was confirmed by intact mass liquid chromatography-mass spectrometry. The discovery of 1, 18, and 19, accompanying structure-activity relationship, and the co-crystal structures provide valuable insights for designing potent and selective chemical tools against WDR91 to evaluate its therapeutic potential.

Deep Learning Approach for the Discovery of Tumor-Targeting Small Organic Ligands from DNA-Encoded Chemical Libraries.

DNA-Encoded Chemical Libraries (DELs) have emerged as efficient and cost-effective ligand discovery tools, which enable the generation of protein-ligand interaction data of unprecedented size. In this article, we present an approach that combines DEL screening and instance-level deep learning modeling to identify tumor-targeting ligands against carbonic anhydrase IX (CAIX), a clinically validated marker of hypoxia and clear cell renal cell carcinoma. We present a new ligand identification and hit-to-lead strategy driven by machine learning models trained on DELs, which expand the scope of DEL-derived chemical motifs. CAIX-screening datasets obtained from three different DELs were used to train machine learning models for generating novel hits, dissimilar to elements present in the original DELs. Out of the 152 novel potential hits that were identified with our approach and screened in an in vitro enzymatic inhibition assay, 70% displayed submicromolar activities (IC50 < 1 µM). To generate lead compounds that are functionalized with anticancer payloads, analogues of top hits were prioritized for synthesis based on the predicted CAIX affinity and synthetic feasibility. Three lead candidates showed accumulation on the surface of CAIX-expressing tumor cells in cellular binding assays. The best compound displayed an in vitro KD of 5.7 nM and selectively targeted tumors in mice bearing human renal cell carcinoma lesions. Our results demonstrate the synergy between DEL and machine learning for the identification of novel hits and for the successful translation of lead candidates for in vivo targeting applications.

Structure Based Design of Potent Selective Inhibitors of Protein Kinase D1 (PKD1).

We previously disclosed a series of type I 1/2 inhibitors of NF-κB inducing kinase (NIK). Inhibition of NIK by these compounds was found to be strongly dependent on the inclusion and absolute stereochemistry of a propargyl tertiary alcohol as it forms critical hydrogen bonds (H-bonds) with NIK. We report that inhibition of protein kinase D1 (PKD1) by this class of compounds is not dependent on H-bond interactions of this tertiary alcohol. This feature was leveraged in the design of highly selective inhibitors of PKD1 that no longer inhibit NIK. A structure-based hypothesis based on the position and flexibility of the α-C-helix of PKD1 vs NIK is presented.

chemalot and chemalot_knime: Command line programs as workflow tools for drug discovery.

BACKGROUND: Analyzing files containing chemical information is at the core of cheminformatics. Each analysis may require a unique workflow. This paper describes the chemalot and chemalot_knime open source packages. Chemalot is a set of command line programs with a wide range of functionalities for cheminformatics. The chemalot_knime package allows command line programs that read and write SD files from stdin and to stdout to be wrapped into KNIME nodes. The combination of chemalot and chemalot_knime not only facilitates the compilation and maintenance of sequences of command line programs but also allows KNIME workflows to take advantage of the compute power of a LINUX cluster. RESULTS: Use of the command line programs is demonstrated in three different workflow examples: (1) A workflow to create a data file with project-relevant data for structure-activity or property analysis and other type of investigations, (2) The creation of a quantitative structure-property-relationship model using the command line programs via KNIME nodes, and (3) The analysis of strain energy in small molecule ligand conformations from the Protein Data Bank database. CONCLUSIONS: The chemalot and chemalot_knime packages provide lightweight and powerful tools for many tasks in cheminformatics. They are easily integrated with other open source and commercial command line tools and can be combined to build new and even more powerful tools. The chemalot_knime package facilitates the generation and maintenance of user-defined command line workflows, taking advantage of the graphical design capabilities in KNIME. Graphical abstract Example KNIME workflow with chemalot nodes and the corresponding command line pipe.

Measuring experimental cyclohexane-water distribution coefficients for the SAMPL5 challenge.

Small molecule distribution coefficients between immiscible nonaqueuous and aqueous phases-such as cyclohexane and water-measure the degree to which small molecules prefer one phase over another at a given pH. As distribution coefficients capture both thermodynamic effects (the free energy of transfer between phases) and chemical effects (protonation state and tautomer effects in aqueous solution), they provide an exacting test of the thermodynamic and chemical accuracy of physical models without the long correlation times inherent to the prediction of more complex properties of relevance to drug discovery, such as protein-ligand binding affinities. For the SAMPL5 challenge, we carried out a blind prediction exercise in which participants were tasked with the prediction of distribution coefficients to assess its potential as a new route for the evaluation and systematic improvement of predictive physical models. These measurements are typically performed for octanol-water, but we opted to utilize cyclohexane for the nonpolar phase. Cyclohexane was suggested to avoid issues with the high water content and persistent heterogeneous structure of water-saturated octanol phases, since it has greatly reduced water content and a homogeneous liquid structure. Using a modified shake-flask LC-MS/MS protocol, we collected cyclohexane/water distribution coefficients for a set of 53 druglike compounds at pH 7.4. These measurements were used as the basis for the SAMPL5 Distribution Coefficient Challenge, where 18 research groups predicted these measurements before the experimental values reported here were released. In this work, we describe the experimental protocol we utilized for measurement of cyclohexane-water distribution coefficients, report the measured data, propose a new bootstrap-based data analysis procedure to incorporate multiple sources of experimental error, and provide insights to help guide future iterations of this valuable exercise in predictive modeling.

An integrated suite of modeling tools that empower scientists in structure- and property-based drug design.

Structure- and property-based drug design is an integral part of modern drug discovery, enabling the design of compounds aimed at improving potency and selectivity. However, building molecules using desktop modeling tools can easily lead to poor designs that appear to form many favorable interactions with the protein's active site. Although a proposed molecule looks good on screen and appears to fit into the protein site X-ray crystal structure or pharmacophore model, doing so might require a high-energy small molecule conformation, which would likely be inactive. To help scientists make better design decisions, we have built integrated, easy-to-use, interactive software tools to perform docking experiments, de novo design, shape and pharmacophore based database searches, small molecule conformational analysis and molecular property calculations. Using a combination of these tools helps scientists in assessing the likelihood that a designed molecule will be active and have desirable drug metabolism and pharmacokinetic properties. Small molecule discovery success requires project teams to rapidly design and synthesize potent molecules with good ADME properties. Empowering scientists to evaluate ideas quickly and make better design decisions with easy-to-access and easy-to-understand software on their desktop is now a key part of our discovery process.

Back pocket flexibility provides group II p21-activated kinase (PAK) selectivity for type I 1/2 kinase inhibitors.

Structure-based methods were used to design a potent and highly selective group II p21-activated kinase (PAK) inhibitor with a novel binding mode, compound 17. Hydrophobic interactions within a lipophilic pocket past the methionine gatekeeper of group II PAKs approached by these type I 1/2 binders were found to be important for improving potency. A structure-based hypothesis and strategy for achieving selectivity over group I PAKs, and the broad kinome, based on unique flexibility of this lipophilic pocket, is presented. A concentration-dependent decrease in tumor cell migration and invasion in two triple-negative breast cancer cell lines was observed with compound 17.

The crystal structure of the catalytic domain of the NF-κB inducing kinase reveals a narrow but flexible active site.

The NF-κB inducing kinase (NIK) regulates the non-canonical NF-κB pathway downstream of important clinical targets including BAFF, RANKL, and LTß. Despite numerous genetic studies associating dysregulation of this pathway with autoimmune diseases and hematological cancers, detailed molecular characterization of this central signaling node has been lacking. We undertook a systematic cloning and expression effort to generate soluble, well-behaved proteins encompassing the kinase domains of human and murine NIK. Structures of the apo NIK kinase domain from both species reveal an active-like conformation in the absence of phosphorylation. ATP consumption and peptide phosphorylation assays confirm that phosphorylation of NIK does not increase enzymatic activity. Structures of murine NIK bound to inhibitors possessing two different chemotypes reveal conformational flexibility in the gatekeeper residue controlling access to a hydrophobic pocket. Finally, a single amino acid difference affects the ability of some inhibitors to bind murine and human NIK with the same affinity.

DEGAS: sharing and tracking target compound ideas with external collaborators.

To minimize the risk of failure in clinical trials, drug discovery teams must propose active and selective clinical candidates with good physicochemical properties. An additional challenge is that today drug discovery is often conducted by teams at different geographical locations. To improve the collaborative decision making on which compounds to synthesize, we have implemented DEGAS, an application which enables scientists from Genentech and from collaborating external partners to instantly access the same data. DEGAS was implemented to ensure that only the best target compounds are made and that they are made without duplicate effort. Physicochemical properties and DMPK model predictions are computed for each compound to allow the team to make informed decisions when prioritizing. The synthesis progress can be easily tracked. While developing DEGAS, ease of use was a particular goal in order to minimize the difficulty of training and supporting remote users.

SKATE: a docking program that decouples systematic sampling from scoring.

SKATE is a docking prototype that decouples systematic sampling from scoring. This novel approach removes any interdependence between sampling and scoring functions to achieve better sampling and, thus, improves docking accuracy. SKATE systematically samples a ligand's conformational, rotational and translational degrees of freedom, as constrained by a receptor pocket, to find sterically allowed poses. Efficient systematic sampling is achieved by pruning the combinatorial tree using aggregate assembly, discriminant analysis, adaptive sampling, radial sampling, and clustering. Because systematic sampling is decoupled from scoring, the poses generated by SKATE can be ranked by any published, or in-house, scoring function. To test the performance of SKATE, ligands from the Asetex/CDCC set, the Surflex set, and the Vertex set, a total of 266 complexes, were redocked to their respective receptors. The results show that SKATE was able to sample poses within 2 A RMSD of the native structure for 98, 95, and 98% of the cases in the Astex/CDCC, Surflex, and Vertex sets, respectively. Cross-docking accuracy of SKATE was also assessed by docking 10 ligands to thymidine kinase and 73 ligands to cyclin-dependent kinase.

A second look at mini-protein stability: analysis of FSD-1 using circular dichroism, differential scanning calorimetry, and simulations.

Mini-proteins that contain <50 amino acids often serve as model systems for studying protein folding because their small size makes long timescale simulations possible. However, not all mini-proteins are created equal. The stability and structure of FSD-1, a 28-residue mini-protein that adopted the betabetaalpha zinc-finger motif independent of zinc binding, was investigated using circular dichroism, differential scanning calorimetry, and replica-exchange molecular dynamics. The broad melting transition of FSD-1, similar to that of a helix-to-coil transition, was observed by using circular dichroism, differential scanning calorimetry, and replica-exchange molecular dynamics. The N-terminal beta-hairpin was found to be flexible. The FSD-1 apparent melting temperature of 41 degrees C may be a reflection of the melting of its alpha-helical segment instead of the entire protein. Thus, despite its attractiveness due to small size and purposefully designed helix, sheet, and turn structures, the status of FSD-1 as a model system for studying protein folding should be reconsidered.

Conformational manifold of alpha-aminoisobutyric acid (Aib) containing alanine-based tripeptides in aqueous solution explored by vibrational spectroscopy, electronic circular dichroism spectroscopy, and molecular dynamics simulations.

Replacement of the alpha-proton of an alanine residue to generate alpha-aminoisobutyric acid (Aib) in alanine-based oligopeptides favors the formation of a 3(10) helix when the length of the oligopeptide is about four to six residues. This research was aimed at experimentally identifying the structural impact of an individual Aib residue in an alanine context of short peptides in water and Aib's influence on the conformation of nearest-neighbor residues. The amide I band profile of the IR, isotropic and anisotropic Raman, and vibrational circular dichroism (VCD) spectra of Ac-Ala-Ala-Aib-OMe, Ac-Ala-Aib-Ala-OMe, and Ac-Aib-Ala-Ala-OMe were measured and analyzed in terms of different structural models by utilizing an algorithm that exploits the excitonic coupling between amide I' modes. The conformational search was guided by the respective 1H NMR and electronic circular dichroism spectra of the respective peptides, which were also recorded. From these analyses, all peptides adopted multiple conformations. Aib predominantly sampled the right-handed and left-handed 3(10)-helix region and to a minor extent the bridge region between the polyproline (PPII) and the helical regions of the Ramachandran plot. Generally, alanine showed the anticipated PPII propensity, but its conformational equilibrium was shifted towards helical conformations in Ac-Aib-Ala-Ala-OMe, indicating that Aib can induce helical conformations of neighboring residues positioned towards the C-terminal direction of the peptide. An energy landscape exploration by molecular dynamics simulations corroborated the results of the spectroscopic studies. They also revealed the dynamics and pathways of potential conformational transitions of the corresponding Aib residues.