AI-Driven De Novo Design and Molecular Modeling for Discovery of Small-Molecule Compounds as Potential Drug Candidates Targeting SARS-CoV-2 Main Protease.

Over the past three years, significant progress has been made in the development of novel promising drug candidates against COVID-19. However, SARS-CoV-2 mutations resulting in the emergence of new viral strains that can be resistant to the drugs used currently in the clinic necessitate the development of novel potent and broad therapeutic agents targeting different vulnerable spots of the viral proteins. In this study, two deep learning generative models were developed and used in combination with molecular modeling tools for de novo design of small molecule compounds that can inhibit the catalytic activity of SARS-CoV-2 main protease (Mpro), an enzyme critically important for mediating viral replication and transcription. As a result, the seven best scoring compounds that exhibited low values of binding free energy comparable with those calculated for two potent inhibitors of Mpro, via the same computational protocol, were selected as the most probable inhibitors of the enzyme catalytic site. In light of the data obtained, the identified compounds are assumed to present promising scaffolds for the development of new potent and broad-spectrum drugs inhibiting SARS-CoV-2 Mpro, an attractive therapeutic target for anti-COVID-19 agents.

Elucidating the Role of Noncovalent Interactions in Favipiravir, a Drug Active against Various Human RNA Viruses; a 1H-14N NQDR/Periodic DFT/QTAIM/RDS/3D Hirshfeld Surfaces Combined Study.

Favipiravir (6-fluoro-3-hydroxypyrazine-2-carboxamide, FPV), an active pharmaceutical component of the drug discovered and registered in March 2014 in Japan under the name Avigan, with an indication for pandemic influenza, has been studied. The study of this compound was prompted by the idea that effective processes of recognition and binding of FPV to the nucleic acid are affected predominantly by the propensity to form intra- and intermolecular interactions. Three nuclear quadrupole resonance experimental techniques, namely 1H-14N cross-relaxation, multiple frequency sweeps, and two-frequency irradiation, followed by solid-state computational modelling (density functional theory supplemented by the quantum theory of atoms in molecules, 3D Hirshfeld Surfaces, and reduced density gradient) approaches were applied. The complete NQR spectrum consisting of nine lines indicating the presence of three chemically inequivalent nitrogen sites in the FPV molecule was detected, and the assignment of lines to particular sites was performed. The description of the nearest vicinity of all three nitrogen atoms was used to characterize the nature of the intermolecular interactions from the perspective of the local single atoms and to draw some conclusions on the nature of the interactions required for effective recognition and binding. The propensity to form the electrostatic N-H···O, N-H···N, and C-H···O intermolecular hydrogen bonds competitive with two intramolecular hydrogen bonds, strong O-H···O and very weak N-H···N, closing the 5-member ring and stiffening the structure, as well as π···π and F···F dispersive interactions, were analysed in detail. The hypothesis regarding the similarity of the interaction pattern in the solid and the RNA template was verified. It was discovered that the -NH2 group in the crystal participates in intermolecular hydrogen bonds N-H···N and N-H···O, in the precatalytic state only in N-H···O, while in the active state in N-H···N and N-H···O hydrogen bonds, which is of importance to link FVP to the RNA template. Our study elucidates the binding modes of FVP (in crystal, precatalytic, and active forms) in detail and should guide the design of more potent analogues targeting SARS-CoV-2. Strong direct binding of FVP-RTP to both the active site and cofactor discovered by us suggests a possible alternative, allosteric mechanism of FVP action, which may explain the scattering of the results of clinical trials or the synergistic effect observed in combined treatment against SARS-CoV-2.

Suramin Disturbs the Association of the N-Terminal Domain of SARS-CoV-2 Nucleocapsid Protein with RNA.

Suramin was originally used as an antiparasitic drug in clinics. Here, we demonstrate that suramin can bind to the N-terminal domain of SARS-CoV-2 nucleocapsid protein (N-NTD) and disturb its interaction with RNA. The BLI experiments showed that N-NTD interacts suramin with a dissociate constant (Kd = 2.74 µM) stronger than that of N-NTD with ssRNA-16 (Kd = 8.37 µM). Furthermore, both NMR titration experiments and molecular docking analysis suggested that suramin mainly binds to the positively charged cavity between the finger and the palm subdomains of N-NTD, and residues R88, R92, R93, I94, R95, K102 and A156 are crucial for N-NTD capturing suramin. Besides, NMR dynamics experiments showed that suramin-bound N-NTD adopts a more rigid structure, and the loop between ß2-ß3 exhibits fast motion on the ps-ns timescale, potentially facilitating suramin binding. Our findings not only reveal the molecular basis of suramin disturbing the association of SARS-CoV-2 N-NTD with RNA but also provide valuable structural information for the development of drugs against SARS-CoV-2.

Large-Scale Virtual Screening for the Discovery of SARS-CoV-2 Papain-like Protease (PLpro) Non-covalent Inhibitors.

The rapid global spread of the SARS-CoV-2 virus facilitated the development of novel direct-acting antiviral agents (DAAs). The papain-like protease (PLpro) has been proposed as one of the major SARS-CoV-2 targets for DAAs due to its dual role in processing viral proteins and facilitating the host's immune suppression. This dual role makes identifying small molecules that can effectively neutralize SARS-CoV-2 PLpro activity a high-priority task. However, PLpro drug discovery faces a significant challenge due to the high mobility and induced-fit effects in the protease's active site. Herein, we virtually screened the ZINC20 database with Deep Docking (DD) to identify prospective noncovalent PLpro binders and combined ultra-large consensus docking with two pharmacophore (ph4)-filtering strategies. The analysis of active compounds revealed their somewhat-limited diversity, likely attributed to the induced-fit nature of PLpro's active site in the crystal structures, and therefore, the use of rigid docking protocols poses inherited limitations. The top hits were assessed against recombinant viral proteins and live viruses, demonstrating desirable inhibitory activities. The best compound VPC-300195 (IC50: 15 µM) ranks among the top noncovalent PLpro inhibitors discovered through in silico methodologies. In the search for novel SARS-CoV-2 PLpro-specific chemotypes, the identified inhibitors could serve as diverse templates for the development of effective noncovalent PLpro inhibitors.

Comprehensive classification of proteins based on structures that engage lipids by COMPOSEL.

Structures can now be predicted for any protein using programs like AlphaFold and Rosetta, which rely on a foundation of experimentally determined structures of architecturally diverse proteins. The accuracy of such artificial intelligence and machine learning (AI/ML) approaches benefits from the specification of restraints which assist in navigating the universe of folds to converge on models most representative of a given protein's physiological structure. This is especially pertinent for membrane proteins, with structures and functions that depend on their presence in lipid bilayers. Structures of proteins in their membrane environments could conceivably be predicted from AI/ML approaches with user-specificized parameters that describe each element of the architecture of a membrane protein accompanied by its lipid environment. We propose the Classification Of Membrane Proteins based On Structures Engaging Lipids (COMPOSEL), which builds on existing nomenclature types for monotopic, bitopic, polytopic and peripheral membrane proteins as well as lipids. Functional and regulatory elements are also defined in the scripts, as shown with membrane fusing synaptotagmins, multidomain PDZD8 and Protrudin proteins that recognize phosphoinositide (PI) lipids, the intrinsically disordered MARCKS protein, caveolins, the ß barrel assembly machine (BAM), an adhesion G-protein coupled receptor (aGPCR) and two lipid modifying enzymes - diacylglycerol kinase DGKε and fatty aldehyde dehydrogenase FALDH. This demonstrates how COMPOSEL communicates lipid interactivity as well as signaling mechanisms and binding of metabolites, drug molecules, polypeptides or nucleic acids to describe the operations of any protein. Moreover COMPOSEL can be scaled to express how genomes encode membrane structures and how our organs are infiltrated by pathogens such as SARS-CoV-2.

Distinguishing features of fold-switching proteins.

Though many folded proteins assume one stable structure that performs one function, a small-but-increasing number remodel their secondary and tertiary structures and change their functions in response to cellular stimuli. These fold-switching proteins regulate biological processes and are associated with autoimmune dysfunction, severe acute respiratory syndrome coronavirus-2 infection, and more. Despite their biological importance, it is difficult to computationally predict fold switching. With the aim of advancing computational prediction and experimental characterization of fold switchers, this review discusses several features that distinguish fold-switching proteins from their single-fold and intrinsically disordered counterparts. First, the isolated structures of fold switchers are less stable and more heterogeneous than single folders but more stable and less heterogeneous than intrinsically disordered proteins (IDPs). Second, the sequences of single fold, fold switching, and intrinsically disordered proteins can evolve at distinct rates. Third, proteins from these three classes are best predicted using different computational techniques. Finally, late-breaking results suggest that single folders, fold switchers, and IDPs have distinct patterns of residue-residue coevolution. The review closes by discussing high-throughput and medium-throughput experimental approaches that might be used to identify new fold-switching proteins.

Discovery of natural-derived Mpro inhibitors as therapeutic candidates for COVID-19: Structure-based pharmacophore screening combined with QSAR analysis.

The main protease (Mpro ) is an essential enzyme for the life cycle of SARS-CoV-2 and a validated target for treatment of COVID-19 infection. Structure-based pharmacophore modeling combined with QSAR calculations were employed to identify new chemical scaffolds of Mpro inhibitors from natural products repository. Hundreds of pharmacophore models were manually built from their corresponding X-ray crystallographic structures. A pharmacophore model that was validated by receiver operating characteristic (ROC) curve analysis and selected using the statistically optimum QSAR equation was implemented as a 3D-search tool to mine AnalytiCon Discovery database of natural products. Captured hits that showed the highest predicted inhibitory activities were bioassayed. Three active Mpro inhibitors (pseurotin A, lactupicrin, and alpinetin) were successfully identified with IC50 values in low micromolar range.

Development of a charged model of the SARS-CoV-2 viral surface.

A recent study provided experimental evidence of inactivation of viral activity after radio-frequency (RF) exposures in the 6-12 GHz band that was hypothesized to be caused by vibrations of an acoustic dipole mode in the virus that excited the viral membrane to failure. Here, we develop an atomic-scale molecular dynamics (MD) model of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viral surface to estimate the electric fields necessary to rupture the viral membrane via dipole shaking of the virus. We computed the absorption spectrum of the system via unbiased MD simulations and found no particular strong absorption in the GHz band. We investigated the mechanical resiliency of the viral membrane by introducing uniaxial strains in the system and observed no pore formation in the membrane for strains up to 50%. Because the computed absorption spectrum was found to be essentially flat, and the strain required to break the viral membrane was >0.5, the field strength associated with rupture of the virus was greater than the dielectric breakdown value of air. Thus, RF disinfection of enveloped viruses would occur only once sufficient heat was transferred to the virus via a thermal mechanism and not by direct action (shaking) of the RF field oscillations on the viral membrane.

Molecular Generative Model via Retrosynthetically Prepared Chemical Building Block Assembly.

Deep generative models are attracting attention as a smart molecular design strategy. However, previous models often render molecules with low synthesizability, hindering their real-world applications. Here, a novel graph-based conditional generative model which makes molecules by tailoring retrosynthetically prepared chemical building blocks until achieving target properties in an auto-regressive fashion is proposed. This strategy improves the synthesizability and property control of the resulting molecules and also helps learn how to select appropriate building blocks and bind them together to achieve target properties. By applying a negative sampling method to the selection process of building blocks, this model overcame a critical limitation of previous fragment-based models, which can only use molecules from the training set during generation. As a result, the model works equally well with unseen building blocks without sacrificing computational efficiency. It is demonstrated that the model can generate potential inhibitors with high docking scores against the 3CL protease of SARS-COV-2.

The Cytoplasmic Domain of the SARS-CoV-2 Envelope Protein Assembles into a ß-Sheet Bundle in Lipid Bilayers.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) envelope (E) protein forms a pentameric ion channel in the lipid membrane of the endoplasmic reticulum Golgi intermediate compartment (ERGIC) of the infected cell. The cytoplasmic domain of E interacts with host proteins to cause virus pathogenicity and may also mediate virus assembly and budding. To understand the structural basis of these functions, here we investigate the conformation and dynamics of an E protein construct (residues 8-65) that encompasses the transmembrane domain and the majority of the cytoplasmic domain using solid-state NMR. 13C and 15N chemical shifts indicate that the cytoplasmic domain adopts a ß-sheet-rich conformation that contains three ß-strands separated by turns. The five subunits associate into an umbrella-shaped bundle that is attached to the transmembrane helices by a disordered loop. Water-edited NMR spectra indicate that the third ß-strand at the C terminus of the protein is well hydrated, indicating that it is at the surface of the ß-bundle. The structure of the cytoplasmic domain cannot be uniquely determined from the inter-residue correlations obtained here due to ambiguities in distinguishing intermolecular and intramolecular contacts for a compact pentameric assembly of this small domain. Instead, we present four structural topologies that are consistent with the measured inter-residue contacts. These data indicate that the cytoplasmic domain of the SARS-CoV-2 E protein has a strong propensity to adopt ß-sheet conformations when the protein is present at high concentrations in lipid bilayers. The equilibrium between the ß-strand conformation and the previously reported α-helical conformation may underlie the multiple functions of E in the host cell and in the virion.

Improved AlphaFold modeling with implicit experimental information.

Machine-learning prediction algorithms such as AlphaFold and RoseTTAFold can create remarkably accurate protein models, but these models usually have some regions that are predicted with low confidence or poor accuracy. We hypothesized that by implicitly including new experimental information such as a density map, a greater portion of a model could be predicted accurately, and that this might synergistically improve parts of the model that were not fully addressed by either machine learning or experiment alone. An iterative procedure was developed in which AlphaFold models are automatically rebuilt on the basis of experimental density maps and the rebuilt models are used as templates in new AlphaFold predictions. We show that including experimental information improves prediction beyond the improvement obtained with simple rebuilding guided by the experimental data. This procedure for AlphaFold modeling with density has been incorporated into an automated procedure for interpretation of crystallographic and electron cryo-microscopy maps.

Structure of the Pandemic.

During global pandemics, the spread of information needs to be faster than the spread of the virus in order to ensure the health and safety of human populations worldwide. In our current crisis, the demand for SARS-CoV-2 drugs and vaccines highlights the importance of biological targets and their three-dimensional shape. In particular, structural biology as a field was poised to quickly respond to crises due to previous experience and expertise and because of its early adoption of open access practices.

Sensitivity of SARS-CoV-2 B.1.1.7 to mRNA vaccine-elicited antibodies.

Transmission of SARS-CoV-2 is uncontrolled in many parts of the world; control is compounded in some areas by the higher transmission potential of the B.1.1.7 variant1, which has now been reported in 94 countries. It is unclear whether the response of the virus to vaccines against SARS-CoV-2 on the basis of the prototypic strain will be affected by the mutations found in B.1.1.7. Here we assess the immune responses of individuals after vaccination with the mRNA-based vaccine BNT162b22. We measured neutralizing antibody responses after the first and second immunizations using pseudoviruses that expressed the wild-type spike protein or a mutated spike protein that contained the eight amino acid changes found in the B.1.1.7 variant. The sera from individuals who received the vaccine exhibited a broad range of neutralizing titres against the wild-type pseudoviruses that were modestly reduced against the B.1.1.7 variant. This reduction was also evident in sera from some patients who had recovered from COVID-19. Decreased neutralization of the B.1.1.7 variant was also observed for monoclonal antibodies that target the N-terminal domain (9 out of 10) and the receptor-binding motif (5 out of 31), but not for monoclonal antibodies that recognize the receptor-binding domain that bind outside the receptor-binding motif. Introduction of the mutation that encodes the E484K substitution in the B.1.1.7 background to reflect a newly emerged variant of concern (VOC 202102/02) led to a more-substantial loss of neutralizing activity by vaccine-elicited antibodies and monoclonal antibodies (19 out of 31) compared with the loss of neutralizing activity conferred by the mutations in B.1.1.7 alone. The emergence of the E484K substitution in a B.1.1.7 background represents a threat to the efficacy of the BNT162b2 vaccine.

Fragment-based computational design of antibodies targeting structured epitopes.

De novo design methods hold the promise of reducing the time and cost of antibody discovery while enabling the facile and precise targeting of predetermined epitopes. Here, we describe a fragment-based method for the combinatorial design of antibody binding loops and their grafting onto antibody scaffolds. We designed and tested six single-domain antibodies targeting different epitopes on three antigens, including the receptor-binding domain of the SARS-CoV-2 spike protein. Biophysical characterization showed that all designs are stable and bind their intended targets with affinities in the nanomolar range without in vitro affinity maturation. We further discuss how a high-resolution input antigen structure is not required, as similar predictions are obtained when the input is a crystal structure or a computer-generated model. This computational procedure, which readily runs on a laptop, provides a starting point for the rapid generation of lead antibodies binding to preselected epitopes.

Structure and Computational Studies of New Sulfonamide Compound: {(4-nitrophenyl)sulfonyl}tryptophan.

Synthesis of sulfonamide through an indirect method that avoids contamination of the product with no need for purification has been carried out using the indirect process. Here, we report the synthesis of a novel sulfonamide compound, ({4-nitrophenyl}sulfonyl)tryptophan (DNSPA) from 4-nitrobenzenesulphonylchloride and L-tryptophan precursors. The slow evaporation method was used to form single crystals of the named compound from methanolic solution. The compound was characterized by X-ray crystallographic analysis and spectroscopic methods (NMR, IR, mass spectrometry, and UV-vis). The sulfonamide N-H NMR signal at 8.07-8.09 ppm and S-N stretching vibration at 931 cm-1 indicate the formation of the target compound. The compound crystallized in the monoclinic crystal system and P21 space group with four molecules of the compound in the asymmetric unit. Molecular aggregation in the crystal structure revealed a 12-molecule aggregate synthon sustained by O-H⋯O hydrogen bonds and stabilised by N-H⋯O intermolecular contacts. Experimental studies were complemented by DFT calculations at the B3LYP/6-311++G(d,p) level of theory. The computed structural and spectroscopic data are in good agreement with those obtained experimentally. The energies of interactions between the units making up the molecule were calculated. Molecular docking studies showed that DNSPA has a binding energy of -6.37 kcal/mol for E. coli DNA gyrase (5MMN) and -6.35 kcal/mol for COVID-19 main protease (6LU7).

Electronic Structures and Reactivities of COVID-19 Drugs: A DFT Study.

These days, the world is facing the threat of pandemic Coronavirus Disease 2019 (COVID-19). Although a vaccine has been found to combat the pandemic, it is essential to find drugs for an effective treatment method against this disease as soon as possible. In this study, electronic and thermodynamic properties, molecular electrostatic potential (MEP) analysis, and frontier molecular orbitals (FMOs) of nine different covid drugs were studied with Density Functional Theory (DFT). In addition, the relationship between the electronic structures of these drugs and their biological effectiveness was examined. All parameters were computed at the B3LYP/6-311++g(d,p) level. The Solvent effect was evaluated using conductor-like polarizable continuum model (CPCM) as the solvation model. It was observed that electrophilic indexes were important to understand the efficiencies of these drugs in COVID-19 disease. Paxlovid, hydroxyquinone, and nitazoxanide were found as the most thermodynamically stable molecules. Thermodynamic parameters also demonstrated that these drugs were more stable in the aqueous media. Global descriptors and the reactivity of these drugs were found to be related. Nitazoxanide molecule exhibited the highest dipole moment. The high dipole moments of drugs can cause hydrophilic interactions that increase their effectiveness in an aqueous solution.

Structural heterogeneity and precision of implications drawn from cryo-electron microscopy structures: SARS-CoV-2 spike-protein mutations as a test case.

Protein structures may be used to draw functional implications at the residue level, but how sensitive are these implications to the exact structure used? Calculation of the effects of SARS-CoV-2 S-protein mutations based on experimental cryo-electron microscopy structures have been abundant during the pandemic. To understand the precision of such estimates, we studied three distinct methods to estimate stability changes for all possible mutations in 23 different S-protein structures (3.69 million ΔΔG values in total) and explored how random and systematic errors can be remedied by structure-averaged mutation group comparisons. We show that computational estimates have low precision, due to method and structure heterogeneity making results for single mutations uninformative. However, structure-averaged differences in mean effects for groups of substitutions can yield significant results. Illustrating this protocol, functionally important natural mutations, despite individual variations, average to a smaller stability impact compared to other possible mutations, independent of conformational state (open, closed). In summary, we document substantial issues with precision in structure-based protein modeling and recommend sensitivity tests to quantify these effects, but also suggest partial solutions to the problem in the form of structure-averaged "ensemble" estimates for groups of residues when multiple structures are available.

Biophysical and structural characterizations of the effects of mutations on the structure-activity relationships of SARS-CoV-2 spike protein.

Mutations on the spike (S) protein of SARS-CoV-2 could induce structural changes that help increase viral transmissibility and enhance resistance to antibody neutralization. Here, we report a robust workflow to prepare recombinant S protein variants and its host receptor angiotensin-convert enzyme 2 (ACE2) by using a mammalian cell expression system. The functional states of the S protein variants are investigated by cryo-electron microscopy (cryo-EM) and negative staining electron microscopy (NSEM) to visualize their molecular structures in response to mutations, receptor binding, antibody binding, and environmental changes. The folding stabilities of the S protein variants can be deduced from morphological changes based on NSEM imaging analysis. Differential scanning calorimetry provides thermodynamic information to complement NSEM. Impacts of the mutations on host receptor binding and antibody neutralization are in vitro by kinetic binding analyses in addition to atomic insights gleaned from cryo-electron microscopy (cryo-EM). This experimental strategy is generally applicable to studying the molecular basis of host-pathogen interactions.

The impact of AlphaFold2 on experimental structure solution.

AlphaFold2 is a machine-learning based program that predicts a protein structure based on the amino acid sequence. In this article, we report on the current usages of this new tool and give examples from our work in the Coronavirus Structural Task Force. With its unprecedented accuracy, it can be utilized for the design of expression constructs, de novo protein design and the interpretation of Cryo-EM data with an atomic model. However, these methods are limited by their training data and are of limited use to predict conformational variability and fold flexibility; they also lack co-factors, post-translational modifications and multimeric complexes with oligonucleotides. They also are not always perfect in terms of chemical geometry. Nevertheless, machine learning-based fold prediction is a game changer for structural bioinformatics and experimentalists alike, with exciting developments ahead.