Icolos: a workflow manager for structure-based post-processing of de novo generated small molecules.

SUMMARY: We present Icolos, a workflow manager written in Python as a tool for automating complex structure-based workflows for drug design. Icolos can be used as a standalone tool, for example in virtual screening campaigns, or can be used in conjunction with deep learning-based molecular generation facilitated for example by REINVENT, a previously published molecular de novo design package. In this publication, we focus on the internal structure and general capabilities of Icolos, using molecular docking experiments as an illustrative example. AVAILABILITY AND IMPLEMENTATION: The source code is freely available at https://github.com/MolecularAI/Icolos under the Apache 2.0 license. Tutorial notebooks containing minimal working examples can be found at https://github.com/MolecularAI/IcolosCommunity. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The curcumin analog HO-3867 selectively kills cancer cells by converting mutant p53 protein to transcriptionally active wildtype p53.

p53 is an important tumor-suppressor protein that is mutated in more than 50% of cancers. Strategies for restoring normal p53 function are complicated by the oncogenic properties of mutant p53 and have not met with clinical success. To counteract mutant p53 activity, a variety of drugs with the potential to reconvert mutant p53 to an active wildtype form have been developed. However, these drugs are associated with various negative effects such as cellular toxicity, nonspecific binding to other proteins, and inability to induce a wildtype p53 response in cancer tissue. Here, we report on the effects of a curcumin analog, HO-3867, on p53 activity in cancer cells from different origins. We found that HO-3867 covalently binds to mutant p53, initiates a wildtype p53-like anticancer genetic response, is exclusively cytotoxic toward cancer cells, and exhibits high anticancer efficacy in tumor models. In conclusion, HO-3867 is a p53 mutant-reactivating drug with high clinical anticancer potential.

2-Sulfonylpyrimidines: Mild alkylating agents with anticancer activity toward p53-compromised cells.

The tumor suppressor p53 has the most frequently mutated gene in human cancers. Many of p53's oncogenic mutants are just destabilized and rapidly aggregate, and are targets for stabilization by drugs. We found certain 2-sulfonylpyrimidines, including one named PK11007, to be mild thiol alkylators with anticancer activity in several cell lines, especially those with mutationally compromised p53. PK11007 acted by two routes: p53 dependent and p53 independent. PK11007 stabilized p53 in vitro via selective alkylation of two surface-exposed cysteines without compromising its DNA binding activity. Unstable p53 was reactivated by PK11007 in some cancer cell lines, leading to up-regulation of p53 target genes such as p21 and PUMA. More generally, there was cell death that was independent of p53 but dependent on glutathione depletion and associated with highly elevated levels of reactive oxygen species and induction of endoplasmic reticulum (ER) stress, as also found for the anticancer agent PRIMA-1(MET)(APR-246). PK11007 may be a lead for anticancer drugs that target cells with nonfunctional p53 or impaired reactive oxygen species (ROS) detoxification in a wide variety of mutant p53 cells.

Machine learning estimates of natural product conformational energies.

Machine learning has been used for estimation of potential energy surfaces to speed up molecular dynamics simulations of small systems. We demonstrate that this approach is feasible for significantly larger, structurally complex molecules, taking the natural product Archazolid A, a potent inhibitor of vacuolar-type ATPase, from the myxobacterium Archangium gephyra as an example. Our model estimates energies of new conformations by exploiting information from previous calculations via Gaussian process regression. Predictive variance is used to assess whether a conformation is in the interpolation region, allowing a controlled trade-off between prediction accuracy and computational speed-up. For energies of relaxed conformations at the density functional level of theory (implicit solvent, DFT/BLYP-disp3/def2-TZVP), mean absolute errors of less than 1 kcal/mol were achieved. The study demonstrates that predictive machine learning models can be developed for structurally complex, pharmaceutically relevant compounds, potentially enabling considerable speed-ups in simulations of larger molecular structures.

pROC-Chemotype Plots Enhance the Interpretability of Benchmarking Results in Structure-Based Virtual Screening.

Recently, we have reported a systematic comparison of molecular preparation protocols (using MOE or Maestro) in combination with two docking tools (GOLD or Glide), employing our DEKOIS 2.0 benchmark sets. Herein, we demonstrate how comparable settings of data preparation protocols can affect the profile and AUC of pROC curves based on variations in chemotype enrichment. We show how the recognition of different classes of chemotypes can affect the docking performance, particularly in the early enrichment, and monitor changes in this recognition behavior based on score normalization and rescoring strategies. For this, we have developed "pROC-Chemotype", which is an automated protocol that matches and visualizes ligand chemotype information together with potency classes in the pROC profiles obtained by docking. This tool enhances the understanding of the influence of chemotype recognition in early enrichment, but also reveals trends of impaired recognition of chemotype classes at the end of the score-ordered rank. Identifying such issues helps to devise score-normalization strategies to overcome this potential bias in an intuitive manner. Furthermore, strong perturbations in chemotype ranking between different methods can help to identify the underlying reasons (e.g., changes in the protonation/tautomerization state). It also assists in the selection of appropriate scoring functions that are capable to retrieve more potent and diverse hits. In summary, we demonstrate how this new tool can be utilized to identify and highlight chemotype-specific behavior, e.g., in dataset preparation. This can help to overcome some chemistry-related bias in virtual screening campaigns. pROC-Chemotype is made freely available at www.dekois.com.

Lithocholic acid is an endogenous inhibitor of MDM4 and MDM2.

The proteins MDM2 and MDM4 are key negative regulators of the tumor suppressor protein p53, which are frequently upregulated in cancer cells. They inhibit the transactivation activity of p53 by binding separately or in concert to its transactivation domain. MDM2 is also a ubiquitin ligase that leads to the degradation of p53. Accordingly, MDM2 and MDM4 are important targets for drugs to inhibit their binding to p53. We found from in silico screening and confirmed by experiment that lithocholic acid (LCA) binds to the p53 binding sites of both MDM2 and MDM4 with a fivefold preference for MDM4. LCA is an endogenous steroidal bile acid, variously reported to have both carcinogenic and apoptotic activities. The comparison of LCA effects on apoptosis in HCT116 p53(+/+) vs. p53(-/-) cells shows a predominantly p53-mediated induction of caspase-3/7. The dissociation constants are in the µM region, but only modest inhibition of binding of MDM2 and MDM4 is required to negate their upregulation because they have to compete with transcriptional coactivator p300 for binding to p53. Binding was weakened by structural changes in LCA, and so it may be a natural ligand of MDM2 and MDM4, raising the possibility that MDM proteins may be sensors for specific steroids.

Optimization of Potent Ligands for the E3 Ligase DCAF15 and Evaluation of Their Use in Heterobifunctional Degraders.

Unlocking novel E3 ligases for use in heterobifunctional PROTAC degraders is of high importance to the pharmaceutical industry. Over-reliance on the current suite of ligands used to recruit E3 ligases could limit the potential of their application. To address this, potent ligands for DCAF15 were optimized using cryo-EM supported, structure-based design to improve on micromolar starting points. A potent binder, compound 24, was identified and subsequently conjugated into PROTACs against multiple targets. Following attempts on degrading a number of proteins using DCAF15 recruiting PROTACs, only degradation of BRD4 was observed. Deconvolution of the mechanism of action showed that this degradation was not mediated by DCAF15, thereby highlighting both the challenges faced when trying to expand the toolbox of validated E3 ligase ligands for use in PROTAC degraders and the pitfalls of using BRD4 as a model substrate.

mRNA Display Identifies Potent, Paralog-Selective Peptidic Ligands for ARID1B.

The ARID1A and ARID1B subunits are mutually exclusive components of the BAF variant of SWI/SNF chromatin remodeling complexes. Loss of function mutations in ARID1A are frequently observed in various cancers, resulting in a dependency on the paralog ARID1B for cancer cell proliferation. However, ARID1B has never been targeted directly, and the high degree of sequence similarity to ARID1A poses a challenge for the development of selective binders. In this study, we used mRNA display to identify peptidic ligands that bind with nanomolar affinities to ARID1B and showed high selectivity over ARID1A. Using orthogonal biochemical, biophysical, and chemical biology tools, we demonstrate that the peptides engage two different binding pockets, one of which directly involves an ARID1B-exclusive cysteine that could allow covalent targeting by small molecules. Our findings impart the first evidence of the ligandability of ARID1B, provide valuable tools for drug discovery, and suggest opportunities for the development of selective molecules to exploit the synthetic lethal relationship between ARID1A and ARID1B in cancer.

Evaluation and optimization of virtual screening workflows with DEKOIS 2.0--a public library of challenging docking benchmark sets.

The application of molecular benchmarking sets helps to assess the actual performance of virtual screening (VS) workflows. To improve the efficiency of structure-based VS approaches, the selection and optimization of various parameters can be guided by benchmarking. With the DEKOIS 2.0 library, we aim to further extend and complement the collection of publicly available decoy sets. Based on BindingDB bioactivity data, we provide 81 new and structurally diverse benchmark sets for a wide variety of different target classes. To ensure a meaningful selection of ligands, we address several issues that can be found in bioactivity data. We have improved our previously introduced DEKOIS methodology with enhanced physicochemical matching, now including the consideration of molecular charges, as well as a more sophisticated elimination of latent actives in the decoy set (LADS). We evaluate the docking performance of Glide, GOLD, and AutoDock Vina with our data sets and highlight existing challenges for VS tools. All DEKOIS 2.0 benchmark sets will be made accessible at http://www.dekois.com.

Design and Evaluation of a Low Hydrogen Bond Donor Count Fragment Screening Set to Aid Hit Generation of PROTACs Intended for Oral Delivery.

The development of orally bioavailable PROTACs presents a significant challenge due to the inflated physicochemical properties of such heterobifunctional molecules. Molecules occupying this "beyond rule of five" space often demonstrate limited oral bioavailability due to the compounding effects of elevated molecular weight and hydrogen bond donor count (among other properties), but it is possible to achieve sufficient oral bioavailability through physicochemical optimization. Herein, we disclose the design and evaluation of a low hydrogen bond donor count (≤1 HBD) fragment screening set to aid hit generation of PROTACs intended for an oral route of delivery. We demonstrate that application of this library can enhance fragment screens against PROTAC proteins of interest and ubiquitin ligases, yielding fragment hits containing ≤1 HBD suitable for optimizing toward orally bioavailable PROTACs.

DEKOIS: demanding evaluation kits for objective in silico screening--a versatile tool for benchmarking docking programs and scoring functions.

For widely applied in silico screening techniques success depends on the rational selection of an appropriate method. We herein present a fast, versatile, and robust method to construct demanding evaluation kits for objective in silico screening (DEKOIS). This automated process enables creating tailor-made decoy sets for any given sets of bioactives. It facilitates a target-dependent validation of docking algorithms and scoring functions helping to save time and resources. We have developed metrics for assessing and improving decoy set quality and employ them to investigate how decoy embedding affects docking. We demonstrate that screening performance is target-dependent and can be impaired by latent actives in the decoy set (LADS) or enhanced by poor decoy embedding. The presented method allows extending and complementing the collection of publicly available high quality decoy sets toward new target space. All present and future DEKOIS data sets will be made accessible at www.dekois.com.

Put a ring on it: application of small aliphatic rings in medicinal chemistry.

Aliphatic three- and four-membered rings including cyclopropanes, cyclobutanes, oxetanes, azetidines and bicyclo[1.1.1]pentanes have been increasingly exploited in medicinal chemistry for their beneficial physicochemical properties and applications as functional group bioisosteres. This review provides a historical perspective and comparative up to date overview of commonly applied small rings, exemplifying key principles with recent literature examples. In addition to describing the merits and advantages of each ring system, potential hazards and liabilities are also illustrated and explained, including any significant chemical or metabolic stability and toxicity risks.

DockStream: a docking wrapper to enhance de novo molecular design.

Recently, we have released the de novo design platform REINVENT in version 2.0. This improved and extended iteration supports far more features and scoring function components, which allows bespoke and tailor-made protocols to maximize impact in small molecule drug discovery projects. A major obstacle of generative models is producing active compounds, in which predictive (QSAR) models have been applied to enrich target activity. However, QSAR models are inherently limited by their applicability domains. To overcome these limitations, we introduce a structure-based scoring component for REINVENT. DockStream is a flexible, stand-alone molecular docking wrapper that provides access to a collection of ligand embedders and docking backends. Using the benchmarking and analysis workflow provided in DockStream, execution and subsequent analysis of a variety of docking configurations can be automated. Docking algorithms vary greatly in performance depending on the target and the benchmarking and analysis workflow provides a streamlined solution to identifying productive docking configurations. We show that an informative docking configuration can inform the REINVENT agent to optimize towards improving docking scores using public data. With docking activated, REINVENT is able to retain key interactions in the binding site, discard molecules which do not fit the binding cavity, harness unused (sub-)pockets, and improve overall performance in the scaffold-hopping scenario. The code is freely available at https://github.com/MolecularAI/DockStream .

Supporting SARS-CoV-2 Papain-Like Protease Drug Discovery: In silico Methods and Benchmarking.

The coronavirus disease 19 (COVID-19) is a rapidly growing pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Its papain-like protease (SARS-CoV-2 PLpro) is a crucial target to halt virus replication. SARS-CoV PLpro and SARS-CoV-2 PLpro share an 82.9% sequence identity and a 100% sequence identity for the binding site reported to accommodate small molecules in SARS-CoV. The flexible key binding site residues Tyr269 and Gln270 for small-molecule recognition in SARS-CoV PLpro exist also in SARS-CoV-2 PLpro. This inspired us to use the reported small-molecule binders to SARS-CoV PLpro to generate a high-quality DEKOIS 2.0 benchmark set. Accordingly, we used them in a cross-benchmarking study against SARS-CoV-2 PLpro. As there is no SARS-CoV-2 PLpro structure complexed with a small-molecule ligand publicly available at the time of manuscript submission, we built a homology model based on the ligand-bound SARS-CoV structure for benchmarking and docking purposes. Three publicly available docking tools FRED, AutoDock Vina, and PLANTS were benchmarked. All showed better-than-random performances, with FRED performing best against the built model. Detailed performance analysis via pROC-Chemotype plots showed a strong enrichment of the most potent bioactives in the early docking ranks. Cross-benchmarking against the X-ray structure complexed with a peptide-like inhibitor confirmed that FRED is the best-performing tool. Furthermore, we performed cross-benchmarking against the newly introduced X-ray structure complexed with a small-molecule ligand. Interestingly, its benchmarking profile and chemotype enrichment were comparable to the built model. Accordingly, we used FRED in a prospective virtual screen of the DrugBank database. In conclusion, this study provides an example of how to harness a custom-made DEKOIS 2.0 benchmark set as an approach to enhance the virtual screening success rate against a vital target of the rapidly emerging pandemic.

Targeting Cavity-Creating p53 Cancer Mutations with Small-Molecule Stabilizers: the Y220X Paradigm.

We have previously shown that the thermolabile, cavity-creating p53 cancer mutant Y220C can be reactivated by small-molecule stabilizers. In our ongoing efforts to unearth druggable variants of the p53 mutome, we have now analyzed the effects of other cancer-associated mutations at codon 220 on the structure, stability, and dynamics of the p53 DNA-binding domain (DBD). We found that the oncogenic Y220H, Y220N, and Y220S mutations are also highly destabilizing, suggesting that they are largely unfolded under physiological conditions. A high-resolution crystal structure of the Y220S mutant DBD revealed a mutation-induced surface crevice similar to that of Y220C, whereas the corresponding pocket's accessibility to small molecules was blocked in the structure of the Y220H mutant. Accordingly, a series of carbazole-based small molecules, designed for stabilizing the Y220C mutant, also bound to and stabilized the folded state of the Y220S mutant, albeit with varying affinities due to structural differences in the binding pocket of the two mutants. Some of the compounds also bound to and stabilized the Y220N mutant, but not the Y220H mutant. Our data validate the Y220S and Y220N mutants as druggable targets and provide a framework for the design of Y220S or Y220N-specific compounds as well as compounds with dual Y220C/Y220S specificity for use in personalized cancer therapy.

Electrostatic Complementarity as a Fast and Effective Tool to Optimize Binding and Selectivity of Protein-Ligand Complexes.

Electrostatic interactions between small molecules and their respective receptors are essential for molecular recognition and are also key contributors to the binding free energy. Assessing the electrostatic match of protein-ligand complexes therefore provides important insights into why ligands bind and what can be changed to improve binding. Ideally, the ligand and protein electrostatic potentials at the protein-ligand interaction interface should maximize their complementarity while minimizing desolvation penalties. In this work, we present a fast and efficient tool to calculate and visualize the electrostatic complementarity (EC) of protein-ligand complexes. We compiled benchmark sets demonstrating electrostatically driven structure-activity relationships (SAR) from literature data, including kinase, protein-protein interaction, and GPCR targets, and used these to demonstrate that the EC method can visualize, rationalize, and predict electrostatically driven ligand affinity changes and help to predict compound selectivity. The methodology presented here for the analysis of EC is a powerful and versatile tool for drug design.

A structure-guided molecular chaperone approach for restoring the transcriptional activity of the p53 cancer mutant Y220C.

Aim: The p53 cancer mutation Y220C creates a conformationally unstable protein with a unique elongated surface crevice that can be targeted by molecular chaperones. We report the structure-guided optimization of the carbazole-based stabilizer PK083. Materials & methods: Biophysical, cellular and x-ray crystallographic techniques have been employed to elucidate the mode of action of the carbazole scaffolds. Results: Targeting an unoccupied subsite of the surface crevice with heterocycle-substituted PK083 analogs resulted in a 70-fold affinity increase to single-digit micromolar levels, increased thermal stability and decreased rate of aggregation of the mutant protein. PK9318, one of the most potent binders, restored p53 signaling in the liver cancer cell line HUH-7 with homozygous Y220C mutation. Conclusion: The p53-Y220C mutant is an excellent paradigm for the development of mutant p53 rescue drugs via protein stabilization. Similar rescue strategies may be applicable to other cavity-creating p53 cancer mutations.

SLMP53-2 Restores Wild-Type-Like Function to Mutant p53 through Hsp70: Promising Activity in Hepatocellular Carcinoma.

Half of human cancers harbor TP53 mutations that render p53 inactive as a tumor suppressor. In these cancers, reactivation of mutant p53 (mutp53) through restoration of wild-type-like function constitutes a valuable anticancer therapeutic strategy. In order to search for mutp53 reactivators, a small library of tryptophanol-derived oxazoloisoindolinones was synthesized and the potential of these compounds as mutp53 reactivators and anticancer agents was investigated in human tumor cells and xenograft mouse models. By analysis of their anti-proliferative effect on a panel of p53-null NCI-H1299 tumor cells ectopically expressing highly prevalent mutp53, the compound SLMP53-2 was selected based on its potential reactivation of multiple structural mutp53. In mutp53-Y220C-expressing hepatocellular carcinoma (HCC) cells, SLMP53-2-induced growth inhibition was mediated by cell cycle arrest, apoptosis, and endoplasmic reticulum stress response. In these cells, SLMP53-2 restored wild-type-like conformation and DNA-binding ability of mutp53-Y220C by enhancing its interaction with the heat shock protein 70 (Hsp70), leading to the reestablishment of p53 transcriptional activity. Additionally, SLMP53-2 displayed synergistic effect with sorafenib, the only approved therapy for advanced HCC. Notably, it exhibited potent antitumor activity in human HCC xenograft mouse models with a favorable toxicological profile. Collectively, SLMP53-2 is a new mutp53-targeting agent with promising antitumor activity, particularly against HCC.

Mutant p53 as a therapeutic target for the treatment of triple-negative breast cancer: Preclinical investigation with the anti-p53 drug, PK11007.

The identification of a targeted therapy for patients with triple-negative breast cancer (TNBC) is one of the most urgent needs in breast cancer therapeutics. The p53 gene is mutated in approximately 80% of patients with TNBC, and is a potential therapeutic target for patients with this form of breast cancer. The 2-sulfonylpyrimidine compound, PK11007, preferentially decreases viability in p53-compromised cancer cell lines. We investigated PK11007 as a potential new treatment for TNBC. IC50 values for inhibition of proliferation in a panel of 17 breast cell lines by PK11007 ranged from 2.3 to 42.2 µM. There were significantly lower IC50 values for TNBC than for non-TNBC cell lines (p = 0.03) and for p53-mutated cell lines compared with p53 WT cells (p = 0.003). Response to PK11007 however, was independent of the estrogen receptor (ER) or HER2 status of the cell lines. In addition to inhibiting cell proliferation, PK11007 induced apoptosis in p53 mutant cell lines. Using RNAseq and gene ontology analysis, we found that PK11007 altered the expression of genes enriched in pathways involved in regulated cell death, regulation of apoptosis, signal transduction, protein refolding and locomotion. The observations that PK11007 inhibited cell proliferation, induced apoptosis and altered genes involved in cell death are all consistent with the ability of PK11007 to reactivate mutant p53. Based on our data, we conclude that targeting mutant p53 with PK11007 is a potential approach for treating p53-mutated breast cancer, including the subgroup with TN disease.