The rapidly evolving role of cryo-EM in drug design

Structural Characterization of Monoclonal Antibodies and Epitope Mapping by FFAP Footprinting

Covalent labeling in combination with mass spectrometry is a powerful approach used in structural biology to study protein structures, interactions, and dynamics. Recently, the toolbox of covalent labeling techniques has been expanded with fast fluoroalkylation of proteins (FFAP). FFAP is a novel radical labeling method that utilizes fluoroalkyl radicals generated from hypervalent Togni reagents for targeting aromatic residues. This report further demonstrates the benefits of FFAP as a new method for structural characterization of therapeutic antibodies and interaction interfaces of antigen-antibody complexes. The results obtained from human trastuzumab and its complex with human epidermal growth factor receptor 2 (HER2) correlate well with previously published structural data and demonstrate the potential of FFAP in structural biology.

Evaluation of AlphaFold2 Structures for Hit Identification across Multiple Scenarios

The introduction of AlphaFold2 (AF2) has sparked significant enthusiasm and generated extensive discussion within the scientific community, particularly among drug discovery researchers. Although previous studies have addressed the performance of AF2 structures in virtual screening (VS), a more comprehensive investigation is still necessary considering the paramount importance of structural accuracy in drug design. In this study, we evaluate the performance of AF2 structures in VS across three common drug discovery scenarios: targets with holo, apo, and AF2 structures; targets with only apo and AF2 structures; and targets exclusively with AF2 structures. We utilized both the traditional physics-based Glide and the deep-learning-based scoring function RTMscore to rank the compounds in the DUD-E, DEKOIS 2.0, and DECOY data sets. The results demonstrate that, overall, the performance of VS on AF2 structures is comparable to that on apo structures but notably inferior to that on holo structures across diverse scenarios. Moreover, when a target has solely AF2 structure, selecting the holo structure of the target from different subtypes within the same protein family produces comparable results with the AF2 structure for VS on the data set of the AF2 structures, and significantly better results than the AF2 structures on its own data set. This indicates that utilizing AF2 structures for docking-based VS may not yield most satisfactory outcomes, even when solely AF2 structures are available. Moreover, we rule out the possibility that the variations in VS performance between the binding pockets of AF2 and holo structures arise from the differences in their biological assembly composition.

Epitope Mapping of Japanese Encephalitis Virus Neutralizing Antibodies by Native Mass Spectrometry and Hydrogen/Deuterium Exchange

Japanese encephalitis virus (JEV) remains a global public health concern due to its epidemiological distribution and the existence of multiple strains. Neutralizing antibodies against this infection have shown efficacy in in vivo studies. Thus, elucidation of the epitopes of neutralizing antibodies can aid in the design and development of effective vaccines against different strains of JEV. Here, we describe a combination of native mass spectrometry (native-MS) and hydrogen/deuterium exchange mass spectrometry (HDX-MS) to complete screening of eight mouse monoclonal antibodies (MAbs) against JEV E-DIII to identify epitope regions. Native-MS was used as a first pass to identify the antibodies that formed a complex with the target antigen, and it revealed that seven of the eight monoclonal antibodies underwent binding. Native mass spectra of a MAb (JEV-27) known to be non-binding showed broad native-MS peaks and poor signal, suggesting the protein is a mixture or that there are impurities in the sample. We followed native-MS with HDX-MS to locate the binding sites for several of the complex-forming antibodies. This combination of two mass spectrometry-based approaches should be generally applicable and particularly suitable for screening of antigen-antibody and other protein-protein interactions when other traditional approaches give unclear results or are difficult, unavailable, or need to be validated.

Cryo-electron microscopy-based drug design

Structure-based drug design (SBDD) has gained popularity owing to its ability to develop more potent drugs compared to conventional drug-discovery methods. The success of SBDD relies heavily on obtaining the three-dimensional structures of drug targets. X-ray crystallography is the primary method used for solving structures and aiding the SBDD workflow; however, it is not suitable for all targets. With the resolution revolution, enabling routine high-resolution reconstruction of structures, cryogenic electron microscopy (cryo-EM) has emerged as a promising alternative and has attracted increasing attention in SBDD. Cryo-EM offers various advantages over X-ray crystallography and can potentially replace X-ray crystallography in SBDD. To fully utilize cryo-EM in drug discovery, understanding the strengths and weaknesses of this technique and noting the key advancements in the field are crucial. This review provides an overview of the general workflow of cryo-EM in SBDD and highlights technical innovations that enable its application in drug design. Furthermore, the most recent achievements in the cryo-EM methodology for drug discovery are discussed, demonstrating the potential of this technique for advancing drug development. By understanding the capabilities and advancements of cryo-EM, researchers can leverage the benefits of designing more effective drugs. This review concludes with a discussion of the future perspectives of cryo-EM-based SBDD, emphasizing the role of this technique in driving innovations in drug discovery and development. The integration of cryo-EM into the drug design process holds great promise for accelerating the discovery of new and improved therapeutic agents to combat various diseases.

Hydrogen Deuterium Exchange and other Mass Spectrometry-based Approaches for Epitope Mapping

Antigen-antibody interactions are a fundamental subset of protein-protein interactions responsible for the "survival of the fittest". Determining the interacting interface of the antigen, called an epitope, and that on the antibody, called a paratope, is crucial to antibody development. Because each antigen presents multiple epitopes (unique footprints), sophisticated approaches are required to determine the target region for a given antibody. Although X-ray crystallography, Cryo-EM, and nuclear magnetic resonance can provide atomic details of an epitope, they are often laborious, poor in throughput, and insensitive. Mass spectrometry-based approaches offer rapid turnaround, intermediate structural resolution, and virtually no size limit for the antigen, making them a vital approach for epitope mapping. In this review, we describe in detail the principles of hydrogen deuterium exchange mass spectrometry in application to epitope mapping. We also show that a combination of MS-based approaches can assist or complement epitope mapping and push the limit of structural resolution to the residue level. We describe in detail the MS methods used in epitope mapping, provide our perspective about the approaches, and focus on elucidating the role that HDX-MS is playing now and in the future by organizing a discussion centered around several improvements in prototype instrument/applications used for epitope mapping. At the end, we provide a tabular summary of the current literature on HDX-MS-based epitope mapping.

A high-throughput single-particle imaging platform for antibody characterization and a novel competition assay for therapeutic antibodies

Monoclonal antibodies (mAbs) play an important role in diagnostics and therapy of infectious diseases. Here we utilize a single-particle interferometric reflectance imaging sensor (SP-IRIS) for screening 30 mAbs against Ebola, Sudan, and Lassa viruses (EBOV, SUDV, and LASV) to find out the ideal capture antibodies for whole virus detection using recombinant vesicular stomatitis virus (rVSV) models expressing surface glycoproteins (GPs) of EBOV, SUDV, and LASV. We also make use of the binding properties on SP-IRIS to develop a model for mapping the antibody epitopes on the GP structure. mAbs that bind to mucin-like domain or glycan cap of the EBOV surface GP show the highest signal on SP-IRIS, followed by mAbs that target the GP1-GP2 interface at the base domain. These antibodies were shown to be highly efficacious against EBOV infection in non-human primates in previous studies. For LASV detection, 8.9F antibody showed the best performance on SP-IRIS. This antibody binds to a unique region on the surface GP compared to other 15 mAbs tested. In addition, we demonstrate a novel antibody competition assay using SP-IRIS and rVSV-EBOV models to reveal the competition between mAbs in three successful therapeutic mAb cocktails against EBOV infection. We provide an explanation as to why ZMapp cocktail has higher efficacy compared to the other two cocktails by showing that three mAbs in this cocktail (13C6, 2G4, 4G7) do not compete with each other for binding to EBOV GP. In fact, the binding of 13C6 enhances the binding of 2G4 and 4G7 antibodies. Our results establish SP-IRIS as a versatile tool that can provide high-throughput screening of mAbs, multiplexed and sensitive detection of viruses, and evaluation of therapeutic antibody cocktails.

A Hybrid Docking and Machine Learning Approach to Enhance the Performance of Virtual Screening Carried out on Protein-Protein Interfaces

The modulation of protein-protein interactions (PPIs) by small chemical compounds is challenging. PPIs play a critical role in most cellular processes and are involved in numerous disease pathways. As such, novel strategies that assist the design of PPI inhibitors are of major importance. We previously reported that the knowledge-based DLIGAND2 scoring tool was the best-rescoring function for improving receptor-based virtual screening (VS) performed with the Surflex docking engine applied to several PPI targets with experimentally known active and inactive compounds. Here, we extend our investigation by assessing the vs. potential of other types of scoring functions with an emphasis on docking-pose derived solvent accessible surface area (SASA) descriptors, with or without the use of machine learning (ML) classifiers. First, we explored rescoring strategies of Surflex-generated docking poses with five GOLD scoring functions (GoldScore, ChemScore, ASP, ChemPLP, ChemScore with Receptor Depth Scaling) and with consensus scoring. The top-ranked poses were post-processed to derive a set of protein and ligand SASA descriptors in the bound and unbound states, which were combined to derive descriptors of the docked protein-ligand complexes. Further, eight ML models (tree, bagged forest, random forest, Bayesian, support vector machine, logistic regression, neural network, and neural network with bagging) were trained using the derivatized SASA descriptors and validated on test sets. The results show that many SASA descriptors are better than Surflex and GOLD scoring functions in terms of overall performance and early recovery success on the used dataset. The ML models were superior to all scoring functions and rescoring approaches for most targets yielding up to a seven-fold increase in enrichment factors at 1% of the screened collections. In particular, the neural networks and random forest-based ML emerged as the best techniques for this PPI dataset, making them robust and attractive vs. tools for hit-finding efforts. The presented results suggest that exploring further docking-pose derived SASA descriptors could be valuable for structure-based virtual screening projects, and in the present case, to assist the rational design of small-molecule PPI inhibitors.

Cryo-EM structure-based selection of computed ligand poses enables design of MTA-synergic PRMT5 inhibitors of better potency

Projected potential of 2.5-4.0 Å cryo-EM structures for structure-based drug design is not well realized yet. Here we show that a 3.1 Å structure of PRMT5 is suitable for selecting computed poses of a chemical inhibitor and its analogs for enhanced potency. PRMT5, an oncogenic target for various cancer types, has many inhibitors manifesting little cooperativity with MTA, a co-factor analog accumulated in MTAP-/- cells. To achieve MTA-synergic inhibition, a pharmacophore from virtual screen leads to a specific inhibitor (11-2 F). Cryo-EM structures of 11-2 F / MTA-bound human PRMT5/MEP50 complex and its apo form resolved at 3.1 and 3.2 Å respectively show that 11-2 F in the catalytic pocket shifts the cofactor-binding pocket away by ~2.0 Å, contributing to positive cooperativity. Computational analysis predicts subtype specificity of 11-2 F among PRMTs. Structural analysis of ligands in the binding pockets is performed to compare poses of 11-2 F and its redesigned analogs and identifies three new analogs predicted to have significantly better potency. One of them, after synthesis, is ~4 fold more efficient in inhibiting PRMT5 catalysis than 11-2 F, with strong MTA-synergy. These data suggest the feasibility of employing near-atomic resolution cryo-EM structures and computational analysis of ligand poses for small molecule therapeutics.

Putting on molecular weight: Enabling cryo-EM structure determination of sub-100-kDa proteins

Significant advances in the past decade have enabled high-resolution structure determination of a vast variety of proteins by cryogenic electron microscopy single particle analysis. Despite improved sample preparation, next-generation imaging hardware, and advanced single particle analysis algorithms, small proteins remain elusive for reconstruction due to low signal-to-noise and lack of distinctive structural features. Multiple efforts have therefore been directed at the development of size-increase techniques for small proteins. Here we review the latest methods for increasing effective molecular weight of proteins <100 ​kDa through target protein binding or target protein fusion - specifically by using nanobody-based assemblies, fusion tags, and symmetric scaffolds. Finally, we summarize these state-of-the-art techniques into a decision-tree to facilitate the design of tailored future approaches, and thus for further exploration of ever-smaller proteins that make up the largest part of the human genome.

Are Deep Learning Structural Models Sufficiently Accurate for Free-Energy Calculations? Application of FEP+ to AlphaFold2-Predicted Structures

The availability of AlphaFold2 has led to great excitement in the scientific community─particularly among drug hunters─due to the ability of the algorithm to predict protein structures with high accuracy. However, beyond globally accurate protein structure prediction, it remains to be determined whether ligand binding sites are predicted with sufficient accuracy in these structures to be useful in supporting computationally driven drug discovery programs. We explored this question by performing free-energy perturbation (FEP) calculations on a set of well-studied protein-ligand complexes, where AlphaFold2 predictions were performed by removing all templates with >30% identity to the target protein from the training set. We observed that in most cases, the ΔΔG values for ligand transformations calculated with FEP, using these prospective AlphaFold2 structures, were comparable in accuracy to the corresponding calculations previously carried out using crystal structures. We conclude that under the right circumstances, AlphaFold2-modeled structures are accurate enough to be used by physics-based methods such as FEP in typical lead optimization stages of a drug discovery program.

Targeting Amyloid Fibrils by Passive Immunotherapy in Systemic Amyloidosis

Systemic amyloidoses are characterized by the unrelenting deposition of autologous proteins as highly ordered fibrils in target organs. The ensuing, potentially fatal organ dysfunction is the result of the combined damage caused by the proteotoxic effect of prefibrillar species and by the cytotoxicity and the structural alterations produced by the amyloid fibrils. Current therapy is focused on eliminating the amyloid protein, thus extinguishing the amyloid cascade at its origin. While this approach may end the cell damage caused by prefibrillar aggregates and prevent further amyloid accumulation, the noxious effects of the amyloid fibrils persist and may hamper the recovery of organ function, which is the ultimate goal of therapy as it is necessary to improve the quality of life and extend survival. Preclinical studies indicate that the clearance of amyloid deposits can be accelerated by specific antibodies targeting amyloid fibrils that activate complement-mediated macrophages and giant cell phagocytosis, possibly promoting the recovery of organ function. Measuring the therapeutic effect of anti-amyloid agents is still a matter of research. In recent years, several monoclonal antibodies targeting amyloid deposits have been tested in clinical trials with mixed outcomes. Recent encouraging results from phase I/II trials, new anti-amyloid agents, and new antibody engineering offer hope that effective amyloid removal will be accomplished in the near future, accelerating organ recovery and improving quality of life and survival.

AlphaFold2 and CryoEM: Revisiting CryoEM modeling in near-atomic resolution density maps

With the advent of new artificial intelligence and machine learning algorithms, predictive modeling can, in some cases, produce structures on par with experimental methods. The combination of predictive modeling and experimental structure determination by electron cryomicroscopy (cryoEM) offers a tantalizing approach for producing robust atomic models of macromolecular assemblies. Here, we apply AlphaFold2 to a set of community standard data sets and compare the results with the corresponding reference maps and models. Moreover, we present three unique case studies from previously determined cryoEM density maps of viruses. Our results show that AlphaFold2 can not only produce reasonably accurate models for analysis and additional hypotheses testing, but can also potentially yield incorrect structures if not properly validated with experimental data. Whereas we outline numerous shortcomings and potential pitfalls of predictive modeling, the obvious synergy between predictive modeling and cryoEM will undoubtedly result in new computational modeling tools.

On the reduction in the effects of radiation damage to two-dimensional crystals of organic and biological molecules at liquid-helium temperature

We have studied the fading of electron diffraction spots from two-dimensional (2D) crystals of paraffin (C44H90), purple membrane (bacteriorhodopsin) and aquaporin 4 (AQP4) at stage temperatures between 4K and 100K. We observed that the diffraction spots at resolutions between 3 Å and 20 Å fade more slowly at liquid-helium temperatures compared to liquid-nitrogen temperatures, by a factor of between 1.2 and 1.8, depending on the specimens. If the reduction in the effective rate of radiation damage for 2D crystals at liquid-helium temperature (as measured by spot fading) can be shown to extend to macromolecular assemblies embedded in amorphous ice, this would suggest that valuable improvements to electron cryomicroscopy (cryoEM) of biological specimens could be made by reducing the temperature of the specimens under irradiation below what is obtainable using standard liquid-nitrogen cryostats.

Beyond the Backbone: The Next Generation of Pathwalking Utilities for Model Building in CryoEM Density Maps

Single-particle electron cryomicroscopy (cryoEM) has become an indispensable tool for studying structure and function in macromolecular assemblies. As an integral part of the cryoEM structure determination process, computational tools have been developed to build atomic models directly from a density map without structural templates. Nearly a decade ago, we created Pathwalking, a tool for de novo modeling of protein structure in near-atomic resolution cryoEM density maps. Here, we present the latest developments in Pathwalking, including the addition of probabilistic models, as well as a companion tool for modeling waters and ligands. This software was evaluated on the 2021 CryoEM Ligand Challenge density maps, in addition to identifying ligands in three IP3R1 density maps at ~3 Å to 4.1 Å resolution. The results clearly demonstrate that the Pathwalking de novo modeling pipeline can construct accurate protein structures and reliably localize and identify ligand density directly from a near-atomic resolution map.