Accelerating Inhibitor Discovery for Multiple SARS-CoV-2 Targets with a Single, Sequence-Guided Deep Generative Framework (preprint)

The COVID-19 pandemic has highlighted the urgency for developing more efficient molecular discovery pathways. As exhaustive exploration of the vast chemical space is infeasible, discovering novel inhibitor molecules for emerging drug-target proteins is challenging, particularly for targets with unknown structure or ligands. We demonstrate the broad utility of a single deep generative framework toward discovering novel drug-like inhibitor molecules against two distinct SARS-CoV-2 targets — the main protease (Mpro) and the receptor binding domain (RBD) of the spike protein. To perform target-aware design, the framework employs a target sequence-conditioned sampling of novel molecules from a generative model. Micromolar-level in vitro inhibition was observed for two candidates (out of four synthesized) for each target. The most potent spike RBD inhibitor also emerged as a rare non-covalent antiviral with broad-spectrum activity against several SARS-CoV-2 variants in live virus neutralization assays. These results show that a broadly deployable machine intelligence framework can accelerate hit discovery across different emerging drug-targets.

xia2.multiplex: a multi-crystal data analysis pipeline (preprint)

In macromolecular crystallography radiation damage limits the amount of data that can be collected from a single crystal. It is often necessary to merge data sets from multiple crystals, for example small-wedge data collections on micro-crystals, in situ room-temperature data collections, and collection from membrane proteins in lipidic mesophase. Whilst indexing and integration of individual data sets may be relatively straightforward with existing software, merging multiple data sets from small wedges presents new challenges. Identification of a consensus symmetry can be problematic, particularly in the presence of a potential indexing ambiguity. Furthermore, the presence of non-isomorphous or poor-quality data sets may reduce the overall quality of the final merged data set. To facilitate and help optimise the scaling and merging of multiple data sets, we developed a new program, xia2.multiplex , which takes data sets individually integrated with DIALS and performs symmetry analysis, scaling and merging of multicrystal data sets. xia2.multiplex also performs analysis of various pathologies that typically affect multi-crystal data sets, including non-isomorphism, radiation damage and preferential orientation. After describing a number of use cases, we demonstrate the benefit of xia2.multiplex within a wider autoprocessing framework in facilitating a multi-crystal experiment collected as part of in situ room-temperature fragment screening experiments on the SARS-CoV-2 main protease.

COVID Moonshot: Open Science Discovery of SARS-CoV-2 Main Protease Inhibitors by Combining Crowdsourcing, High-Throughput Experiments, Computational Simulations, and Machine Learning (preprint)

Herein we provide a living summary of the data generated during the COVID Moonshot project focused on the development of SARS-CoV-2 main protease (Mpro) inhibitors. Our approach uniquely combines crowdsourced medicinal chemistry insights with high throughput crystallography, exascale computational chemistry infrastructure for simulations, and machine learning in triaging designs and predicting synthetic routes. This manuscript describes our methodologies leading to both covalent and non-covalent inhibitors displaying protease IC50 values under 150 nM and viral inhibition under 5 uM in multiple different viral replication assays. Furthermore, we provide over 200 crystal structures of fragment-like and lead-like molecules in complex with the main protease. Over 1000 synthesized and ordered compounds are also reported with the corresponding activity in Mpro enzymatic assays using two different experimental setups. The data referenced in this document will be continually updated to reflect the current experimental progress of the COVID Moonshot project, and serves as a citable reference for ensuing publications. All of the generated data is open to other researchers who may find it of use.

A cell-free antibody engineering platform rapidly generates SARS-CoV-2 neutralizing antibodies (preprint)

Antibody engineering technologies face increasing demands for speed, reliability and scale. We developed CeVICA, a cell-free antibody engineering platform that integrates a novel generation method and design for camelid heavy-chain antibody VHH domain-based synthetic libraries, optimized in vitro selection based on ribosome display and a computational pipeline for binder prediction based on CDR-directed clustering. We applied CeVICA to engineer antibodies against the Receptor Binding Domain (RBD) of the SARS-CoV-2 spike proteins and identified >800 predicted binder families. Among 14 experimentally-tested binders, 6 showed inhibition of pseudotyped virus infection. Antibody affinity maturation further increased binding affinity and potency of inhibition. Additionally, the unique capability of CeVICA for efficient and comprehensive binder prediction allowed retrospective validation of the fitness of our synthetic VHH library design and revealed direction for future refinement. CeVICA offers an integrated solution to rapid generation of divergent synthetic antibodies with tunable affinities in vitro and may serve as the basis for automated and highly parallel antibody generation.

PRAK-03202: A triple antigen VLP vaccine candidate against SARS CoV-2 (preprint)

The rapid development of safe and effective vaccines against SARS CoV-2 is the need of the hour for the coronavirus outbreak. Here, we have developed PRAK-03202, the world's first triple antigen VLP vaccine candidate in a highly characterized S. cerevisiae-based D-Crypt platform, which induced SARS CoV-2 specific neutralizing antibodies in BALB/c mice. Immunizations using three different doses of PRAK-03202 induces antigen specific (Spike, envelope and membrane proteins) humoral response and neutralizing potential. PBMCs from convalescent patients, when exposed to PRAK-03202, showed lymphocyte proliferation and elevated IFN-{gamma} levels suggestive of conservation of epitopes and induction of T helper 1 (Th1)-biased cellular immune responses. These data support the clinical development and testing of PRAK-03202 for use in humans.

Evidence for adaptive evolution in the receptor-binding domain of seasonal coronaviruses (preprint)

Seasonal coronaviruses (OC43, 229E, NL63 and HKU1) are endemic to the human population, regularly infecting and reinfecting humans while typically causing asymptomatic to mild respiratory infections. It is not known to what extent reinfection by these viruses is due to waning immune memory or antigenic drift of the viruses. Here, we address the influence of antigenic drift on immune evasion of seasonal coronaviruses. We provide evidence that at least two of these viruses, OC43 and 229E, are undergoing adaptive evolution in regions of the viral spike protein that are exposed to human humoral immunity. This suggests that reinfection may be due, in part, to positively-selected genetic changes in these viruses that enable them to escape recognition by the immune system. It is possible that, as with seasonal influenza, these adaptive changes in antigenic regions of the virus would necessitate continual reformulation of a vaccine made against them.