Deconvoluting low yield from weak potency in direct-to-biology workflows with machine learning.

High throughput and rapid biological evaluation of small molecules is an essential factor in drug discovery and development. Direct-to-biology (D2B), whereby compound purification is foregone, has emerged as a viable technique in time efficient screening, specifically for PROTAC design and biological evaluation. However, one notable limitation is the prerequisite of high yielding reactions to ensure the desired compound is indeed the compound responsible for biological activity. Herein, we report a machine learning based yield-assay deconfounder capable of deconvoluting low yield from low potency to identify false negatives. We validated this approach by identifying promising SARS-CoV-2 main protease inhibitors with nanomolar activity that rivaled potency observed from the standard D2B workflow. Furthermore, we show how our framework can be utilized in a broad, in silico screen to produce compounds of similar potency as a D2B assay.

A simple method for developing lysine targeted covalent protein reagents.

Peptide-based covalent probes can target shallow protein surfaces not typically addressable using small molecules, yet there is a need for versatile approaches to convert native peptide sequences into covalent binders that can target a broad range of residues. Here we report protein-based thio-methacrylate esters-electrophiles that can be installed easily on unprotected peptides and proteins via cysteine side chains, and react efficiently and selectively with cysteine and lysine side chains on the target. Methacrylate phosphopeptides derived from 14-3-3-binding proteins irreversibly label 14-3-3σ via either lysine or cysteine residues, depending on the position of the electrophile. Methacrylate peptides targeting a conserved lysine residue exhibit pan-isoform binding of 14-3-3 proteins both in lysates and in extracellular media. Finally, we apply this approach to develop protein-based covalent binders. A methacrylate-modified variant of the colicin E9 immunity protein irreversibly binds to the E9 DNAse, resulting in significantly higher thermal stability relative to the non-covalent complex. Our approach offers a simple and versatile route to convert peptides and proteins into potent covalent binders.

Open science discovery of potent noncovalent SARS-CoV-2 main protease inhibitors.

We report the results of the COVID Moonshot, a fully open-science, crowdsourced, and structure-enabled drug discovery campaign targeting the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease. We discovered a noncovalent, nonpeptidic inhibitor scaffold with lead-like properties that is differentiated from current main protease inhibitors. Our approach leveraged crowdsourcing, machine learning, exascale molecular simulations, and high-throughput structural biology and chemistry. We generated a detailed map of the structural plasticity of the SARS-CoV-2 main protease, extensive structure-activity relationships for multiple chemotypes, and a wealth of biochemical activity data. All compound designs (>18,000 designs), crystallographic data (>490 ligand-bound x-ray structures), assay data (>10,000 measurements), and synthesized molecules (>2400 compounds) for this campaign were shared rapidly and openly, creating a rich, open, and intellectual property-free knowledge base for future anticoronavirus drug discovery.

Turning high-throughput structural biology into predictive inhibitor design.

A common challenge in drug design pertains to finding chemical modifications to a ligand that increases its affinity to the target protein. An underutilized advance is the increase in structural biology throughput, which has progressed from an artisanal endeavor to a monthly throughput of hundreds of different ligands against a protein in modern synchrotrons. However, the missing piece is a framework that turns high-throughput crystallography data into predictive models for ligand design. Here, we designed a simple machine learning approach that predicts protein-ligand affinity from experimental structures of diverse ligands against a single protein paired with biochemical measurements. Our key insight is using physics-based energy descriptors to represent protein-ligand complexes and a learning-to-rank approach that infers the relevant differences between binding modes. We ran a high-throughput crystallography campaign against the SARS-CoV-2 main protease (MPro), obtaining parallel measurements of over 200 protein-ligand complexes and their binding activities. This allows us to design one-step library syntheses which improved the potency of two distinct micromolar hits by over 10-fold, arriving at a noncovalent and nonpeptidomimetic inhibitor with 120 nM antiviral efficacy. Crucially, our approach successfully extends ligands to unexplored regions of the binding pocket, executing large and fruitful moves in chemical space with simple chemistry.

Azapeptide activity-based probes for the SARS-CoV-2 main protease enable visualization of inhibition in infected cells.

The COVID-19 pandemic has revealed the vulnerability of the modern, global society. With expected waves of future infections by SARS-CoV-2, treatment options for infected individuals will be crucial in order to decrease mortality and hospitalizations. The SARS-CoV-2 main protease is a validated drug target, for which the first inhibitor has been approved for use in patients. To facilitate future work on this drug target, we designed a solid-phase synthesis route towards azapeptide activity-based probes that are capped with a cysteine-reactive electrophile for covalent modification of the active site of Mpro. This design led to the most potent ABP for Mpro and one of the most potent inhibitors reported thus far. We demonstrate that this ABP can be used to visualize Mpro activity and target engagement by drugs in infected cells.

Sulfamate Acetamides as Self-Immolative Electrophiles for Covalent Ligand-Directed Release Chemistry.

Electrophiles for covalent inhibitors that are suitable for in vivo administration are rare. While acrylamides are prevalent in FDA-approved covalent drugs, chloroacetamides are considered too reactive for such purposes. We report sulfamate-based electrophiles that maintain chloroacetamide-like geometry with tunable reactivity. In the context of the BTK inhibitor ibrutinib, sulfamate analogues showed low reactivity with comparable potency in protein labeling, in vitro, and cellular kinase activity assays and were effective in a mouse model of CLL. In a second example, we converted a chloroacetamide Pin1 inhibitor to a potent and selective sulfamate acetamide with improved buffer stability. Finally, we show that sulfamate acetamides can be used for covalent ligand-directed release (CoLDR) chemistry, both for the generation of "turn-on" probes as well as for traceless ligand-directed site-specific labeling of proteins. Taken together, this chemistry represents a promising addition to the list of electrophiles suitable for in vivo covalent targeting.

Structure-guided design and characterization of a clickable, covalent PARP16 inhibitor.

PARP16-the sole ER-resident PARP family member-is gaining attention as a potential therapeutic target for cancer treatment. Nevertheless, the precise function of the catalytic activity of PARP16 is poorly understood. This is primarily due to the lack of inhibitors that are selective for PARP16 over other PARP family members. Herein, we describe a structure-guided strategy for generating a selective PARP16 inhibitor by incorporating two selectivity determinants into a phthalazinone pan-PARP inhibitor scaffold: (i) an acrylamide-based inhibitor (DB008) designed to covalently react with a non-conserved cysteine (Cys169, human numbering) in the NAD+ binding pocket of PARP16 and (ii) a dual-purpose ethynyl group designed to bind in a unique hydrophobic cavity adjacent to the NAD+ binding pocket as well as serve as a click handle. DB008 exhibits good selectivity for PARP16 versus other PARP family members. Copper-catalyzed azide-alkyne cycloaddition (CuAAC) confirmed that covalent labeling of PARP16 by DB008 in cells is dependent on Cys169. DB008 exhibits excellent proteome-wide selectivity at concentrations required to achieve saturable labeling of endogenous PARP16. In-cell competition labeling experiments using DB008 provided a facile strategy for evaluating putative PARP16 inhibitors. Lastly, we found that PARP16 is sequestered into a detergent-insoluble fraction under prolonged amino acid starvation, and surprisingly, treatment with PARP16 inhibitors prevented this effect. These results suggest that the catalytic activity of PARP16 regulates its solubility in response to nutrient stress.

Kidney Failure Alters Parathyroid Pin1 Phosphorylation and Parathyroid Hormone mRNA-Binding Proteins, Leading to Secondary Hyperparathyroidism.

BACKGROUND: Secondary hyperparathyroidism (SHP) is a common complication of CKD that increases morbidity and mortality. In experimental SHP, increased parathyroid hormone (PTH) expression is due to enhanced PTH mRNA stability, mediated by changes in its interaction with stabilizing AUF1 and destabilizing KSRP. The isomerase Pin1 leads to KSRP dephosphorylation, but in SHP parathyroid Pin1 activity is decreased and hence phosphorylated KSRP fails to bind PTH mRNA, resulting in high PTH mRNA stability and levels. The up- and downstream mechanisms by which CKD stimulates the parathyroid glands remain elusive. METHODS: Adenine-rich high-phosphate diets induced CKD in rats and mice. Parathyroid organ cultures and transfected cells were incubated with Pin1 inhibitors for their effect on PTH expression. Mass spectrometry was performed on both parathyroid and PTH mRNA pulled-down proteins. RESULTS: CKD led to changes in rat parathyroid proteome and phosphoproteome profiles, including KSRP phosphorylation at Pin1 target sites. Furthermore, both acute and chronic kidney failure led to parathyroid-specific Pin1 Ser16 and Ser71 phosphorylation, which disrupts Pin1 activity. Pharmacologic Pin1 inhibition, which mimics the decreased Pin1 activity in SHP, increased PTH expression ex vivo in parathyroid glands in culture and in transfected cells through the PTH mRNA-protein interaction element and KSRP phosphorylation. CONCLUSIONS: Kidney failure leads to loss of parathyroid Pin1 activity by inducing Pin1 phosphorylation. This predisposes parathyroids to increase PTH production through impaired PTH mRNA decay that is dependent on KSRP phosphorylation at Pin1-target motifs. Pin1 and KSRP phosphorylation and the Pin1-KSRP-PTH mRNA axis thus drive SHP.

Optimization of Covalent MKK7 Inhibitors via Crude Nanomole-Scale Libraries.

High-throughput nanomole-scale synthesis allows for late-stage functionalization (LSF) of compounds in an efficient and economical manner. Here, we demonstrated that copper-catalyzed azide-alkyne cycloaddition could be used for the LSF of covalent kinase inhibitors at the nanoscale, enabling the synthesis of hundreds of compounds that did not require purification for biological assay screening, thus reducing experimental time drastically. We generated crude libraries of inhibitors for the kinase MKK7, derived from two different parental precursors, and analyzed them via the high-throughput In-Cell Western assay. Select inhibitors were resynthesized, validated via conventional biological and biochemical methods such as western blots and liquid chromatography-mass spectrometry (LC-MS) labeling, and successfully co-crystallized. Two of these compounds showed over 20-fold increased inhibitory activity compared to the parental compound. This study demonstrates that high-throughput LSF of covalent inhibitors at the nanomole-scale level can be an auspicious approach in improving the properties of lead chemical matter.

Expanding the Repertoire of Low-Molecular-Weight Pentafluorosulfanyl-Substituted Scaffolds.

The pentafluorosulfanyl (-SF5 ) functional group is of increasing interest as a bioisostere in medicinal chemistry. A library of SF5 -containing compounds, including amide, isoxazole, and oxindole derivatives, was synthesised using a range of solution-based and solventless methods, including microwave and ball-mill techniques. The library was tested against targets including human dihydroorotate dehydrogenase (HDHODH). A subsequent focused approach led to synthesis of analogues of the clinically used disease modifying anti-rheumatic drugs (DMARDs), Teriflunomide and Leflunomide, considered for potential COVID-19 use, where SF5 bioisostere deployment led to improved inhibition of HDHODH compared with the parent drugs. The results demonstrate the utility of the SF5 group in medicinal chemistry.

Site-Specific Labeling of Endogenous Proteins Using CoLDR Chemistry.

Chemical modifications of native proteins can affect their stability, activity, interactions, localization, and more. However, there are few nongenetic methods for the installation of chemical modifications at a specific protein site in cells. Here we report a covalent ligand directed release (CoLDR) site-specific labeling strategy, which enables the installation of a variety of functional tags on a target protein while releasing the directing ligand. Using this approach, we were able to label various proteins such as BTK, K-RasG12C, and SARS-CoV-2 PLpro with different tags. For BTK we have shown selective labeling in cells of both alkyne and fluorophores tags. Protein labeling by traditional affinity methods often inhibits protein activity since the directing ligand permanently occupies the target binding pocket. We have shown that using CoLDR chemistry, modification of BTK by these probes in cells preserves its activity. We demonstrated several applications for this approach including determining the half-life of BTK in its native environment with minimal perturbation, as well as quantification of BTK degradation by a noncovalent proteolysis targeting chimera (PROTAC) by in-gel fluorescence. Using an environment-sensitive "turn-on" fluorescent probe, we were able to monitor ligand binding to the active site of BTK. Finally, we have demonstrated efficient CoLDR-based BTK PROTACs (DC50 < 100 nM), which installed a CRBN binder onto BTK. This approach joins very few available labeling strategies that maintain the target protein activity and thus makes an important addition to the toolbox of chemical biology.

Covalent flexible peptide docking in Rosetta.

Electrophilic peptides that form an irreversible covalent bond with their target have great potential for binding targets that have been previously considered undruggable. However, the discovery of such peptides remains a challenge. Here, we present Rosetta CovPepDock, a computational pipeline for peptide docking that incorporates covalent binding between the peptide and a receptor cysteine. We applied CovPepDock retrospectively to a dataset of 115 disulfide-bound peptides and a dataset of 54 electrophilic peptides. It produced a top-five scoring, near-native model, in 89% and 100% of the cases when docking from the native conformation, and 20% and 90% when docking from an extended peptide conformation, respectively. In addition, we developed a protocol for designing electrophilic peptide binders based on known non-covalent binders or protein-protein interfaces. We identified 7154 peptide candidates in the PDB for application of this protocol. As a proof-of-concept we validated the protocol on the non-covalent complex of 14-3-3σ and YAP1 phosphopeptide. The protocol identified seven highly potent and selective irreversible peptide binders. The predicted binding mode of one of the peptides was validated using X-ray crystallography. This case-study demonstrates the utility and impact of CovPepDock. It suggests that many new electrophilic peptide binders can be rapidly discovered, with significant potential as therapeutic molecules and chemical probes.

Targeting Pin1 renders pancreatic cancer eradicable by synergizing with immunochemotherapy.

Pancreatic ductal adenocarcinoma (PDAC) is characterized by notorious resistance to current therapies attributed to inherent tumor heterogeneity and highly desmoplastic and immunosuppressive tumor microenvironment (TME). Unique proline isomerase Pin1 regulates multiple cancer pathways, but its role in the TME and cancer immunotherapy is unknown. Here, we find that Pin1 is overexpressed both in cancer cells and cancer-associated fibroblasts (CAFs) and correlates with poor survival in PDAC patients. Targeting Pin1 using clinically available drugs induces complete elimination or sustained remissions of aggressive PDAC by synergizing with anti-PD-1 and gemcitabine in diverse model systems. Mechanistically, Pin1 drives the desmoplastic and immunosuppressive TME by acting on CAFs and induces lysosomal degradation of the PD-1 ligand PD-L1 and the gemcitabine transporter ENT1 in cancer cells, besides activating multiple cancer pathways. Thus, Pin1 inhibition simultaneously blocks multiple cancer pathways, disrupts the desmoplastic and immunosuppressive TME, and upregulates PD-L1 and ENT1, rendering PDAC eradicable by immunochemotherapy.

Cross-reactive antibodies against human coronaviruses and the animal coronavirome suggest diagnostics for future zoonotic spillovers.

The spillover of animal coronaviruses (aCoVs) to humans has caused SARS, MERS, and COVID-19. While antibody responses displaying cross-reactivity between SARS-CoV-2 and seasonal/common cold human coronaviruses (hCoVs) have been reported, potential cross-reactivity with aCoVs and the diagnostic implications are incompletely understood. Here, we probed for antibody binding against all seven hCoVs and 49 aCoVs represented as 12,924 peptides within a phage-displayed antigen library. Antibody repertoires of 269 recovered COVID-19 patients showed distinct changes compared to 260 unexposed pre-pandemic controls, not limited to binding of SARS-CoV-2 antigens but including binding to antigens from hCoVs and aCoVs with shared motifs to SARS-CoV-2. We isolated broadly reactive monoclonal antibodies from recovered COVID-19 patients that bind a shared motif of SARS-CoV-2, hCoV-OC43, hCoV-HKU1, and several aCoVs, demonstrating that interspecies cross-reactivity can be mediated by a single immunoglobulin. Employing antibody binding data against the entire CoV antigen library allowed accurate discrimination of recovered COVID-19 patients from unexposed individuals by machine learning. Leaving out SARS-CoV-2 antigens and relying solely on antibody binding to other hCoVs and aCoVs achieved equally accurate detection of SARS-CoV-2 infection. The ability to detect SARS-CoV-2 infection without knowledge of its unique antigens solely from cross-reactive antibody responses against other hCoVs and aCoVs suggests a potential diagnostic strategy for the early stage of future pandemics. Creating regularly updated antigen libraries representing the animal coronavirome can provide the basis for a serological assay already poised to identify infected individuals following a future zoonotic transmission event.

An automatic pipeline for the design of irreversible derivatives identifies a potent SARS-CoV-2 Mpro inhibitor.

Designing covalent inhibitors is increasingly important, although it remains challenging. Here, we present covalentizer, a computational pipeline for identifying irreversible inhibitors based on structures of targets with non-covalent binders. Through covalent docking of tailored focused libraries, we identify candidates that can bind covalently to a nearby cysteine while preserving the interactions of the original molecule. We found âˆ¼11,000 cysteines proximal to a ligand across 8,386 complexes in the PDB. Of these, the protocol identified 1,553 structures with covalent predictions. In a prospective evaluation, five out of nine predicted covalent kinase inhibitors showed half-maximal inhibitory concentration (IC50) values between 155 nM and 4.5 µM. Application against an existing SARS-CoV Mpro reversible inhibitor led to an acrylamide inhibitor series with low micromolar IC50 values against SARS-CoV-2 Mpro. The docking was validated by 12 co-crystal structures. Together these examples hint at the vast number of covalent inhibitors accessible through our protocol.

Intracellular Protein-Drug Interactions Probed by Direct Mass Spectrometry of Cell Lysates.

Understanding protein-ligand interactions in a cellular context is an important goal in molecular biology and biochemistry, and particularly for drug development. Investigators must demonstrate that drugs penetrate cells and specifically bind their targets. Towards that end, we present a native mass spectrometry (MS)-based method for analyzing drug uptake and target engagement in eukaryotic cells. This method is based on our previously introduced direct-MS method for rapid analysis of proteins directly from crude samples. Here, direct-MS enables label-free studies of protein-drug binding in human cells and is used to determine binding affinities of lead compounds in crude samples. We anticipate that this method will enable the application of native MS to a range of problems where cellular context is important, including protein-protein interactions, drug uptake and binding, and characterization of therapeutic proteins.

Proteolysis Targeting Chimeras for BTK Efficiently Inhibit B-Cell Receptor Signaling and Can Overcome Ibrutinib Resistance in CLL Cells.

Proteolysis targeting chimeras (PROTACs) are small molecules that form ternary complexes between their target and E3 ligase, resulting in ubiquitination and proteasomal degradation of the target protein. Using our own designed Bruton's tyrosine kinase (BTK) PROTAC compounds, we show herein efficient BTK degradation in chronic lymphocytic leukemia (CLL) cells. The reversible non-covalent compound (NC-1) was the most potent and therefore we focused on this PROTAC to investigate its subsequent effects on the BCR pathway. NC-1 decreased baseline BTK phosphorylation as well as activation of BTK and other signaling molecules downstream of the BCR pathway, following IgM engagement. These effects were also obtained in samples from CLL patients with clinical resistance to ibrutinib and mutations at C481. NC-1 treatment further decreased baseline CD69 surface levels, completely abrogated its upregulation following IgM activation, decreased CLL cells migration toward SDF-1 and overcame stromal anti-apoptotic protection. In conclusion, our results indicate that targeting BTK using the PROTAC strategy could be a potential novel therapeutic approach for CLL.

Sulfopin is a covalent inhibitor of Pin1 that blocks Myc-driven tumors in vivo.

The peptidyl-prolyl isomerase, Pin1, is exploited in cancer to activate oncogenes and inactivate tumor suppressors. However, despite considerable efforts, Pin1 has remained an elusive drug target. Here, we screened an electrophilic fragment library to identify covalent inhibitors targeting Pin1's active site Cys113, leading to the development of Sulfopin, a nanomolar Pin1 inhibitor. Sulfopin is highly selective, as validated by two independent chemoproteomics methods, achieves potent cellular and in vivo target engagement and phenocopies Pin1 genetic knockout. Pin1 inhibition had only a modest effect on cancer cell line viability. Nevertheless, Sulfopin induced downregulation of c-Myc target genes, reduced tumor progression and conferred survival benefit in murine and zebrafish models of MYCN-driven neuroblastoma, and in a murine model of pancreatic cancer. Our results demonstrate that Sulfopin is a chemical probe suitable for assessment of Pin1-dependent pharmacology in cells and in vivo, and that Pin1 warrants further investigation as a potential cancer drug target.

Discovery of SARS-CoV-2 main protease inhibitors using a synthesis-directed de novo design model.

The SARS-CoV-2 main viral protease (Mpro) is an attractive target for antivirals given its distinctiveness from host proteases, essentiality in the viral life cycle and conservation across coronaviridae. We launched the COVID Moonshot initiative to rapidly develop patent-free antivirals with open science and open data. Here we report the use of machine learning for de novo design, coupled with synthesis route prediction, in our campaign. We discover novel chemical scaffolds active in biochemical and live virus assays, synthesized with model generated routes.