Contributions of Hyperactive Mutations in Mpro from SARS-CoV-2 to Drug Resistance

The appearance and spread of mutations that cause drug resistance in rapidly evolving diseases, including infections by the SARS-CoV-2 virus, are major concerns for human health. Many drugs target enzymes, and resistance-conferring mutations impact inhibitor binding or enzyme activity. Nirmatrelvir, the most widely used inhibitor currently used to treat SARS-CoV-2 infections, targets the main protease (Mpro) preventing it from processing the viral polyprotein into active subunits. Our previous work systematically analyzed resistance mutations in Mpro that reduce binding to inhibitors; here, we investigate mutations that affect enzyme function. Hyperactive mutations that increase Mpro activity can contribute to drug resistance but have not been thoroughly studied. To explore how hyperactive mutations contribute to resistance, we comprehensively assessed how all possible individual mutations in Mpro affect enzyme function using a mutational scanning approach with a fluorescence resonance energy transfer (FRET)-based yeast readout. We identified hundreds of mutations that significantly increased the Mpro activity. Hyperactive mutations occurred both proximal and distal to the active site, consistent with protein stability and/or dynamics impacting activity. Hyperactive mutations were observed 3 times more than mutations which reduced apparent binding to nirmatrelvir in recent studies of laboratory-grown viruses selected for drug resistance. Hyperactive mutations were also about three times more prevalent than nirmatrelvir binding mutations in sequenced isolates from circulating SARS-CoV-2. Our findings indicate that hyperactive mutations are likely to contribute to the natural evolution of drug resistance in Mpro and provide a comprehensive list for future surveillance efforts.

Direct delivery of stabilized Cas-embedded base editors achieves efficient and accurate editing of clinically relevant targets

Over the last 5 years, cytosine base editors (CBEs) have emerged as a promising therapeutic tool for specific editing of single nucleotide variants and disrupting specific genes associated with disease. Despite this promise, the currently available CBE's have the significant liabilities of off-target and bystander editing activities, in part due to the mechanism by which they are delivered, causing limitations in their potential applications. In this study we engineeredhighly stabilized Cas-embedded CBEs (sCE_CBEs) that integrate several recent advances, andthat are highly expressible and soluble for direct delivery into cells as ribonucleoprotein (RNP) complexes. Our resulting sCE_CBE RNP complexes efficiently and specifically target TC dinucleotides with minimal off-target or bystander mutations. Additional uracil glycosylase inhibitor (UGI) protein in trans further increased C-to-T editing efficiency and target purity in a dose-dependent manner, minimizing indel formation to untreated levels. A single electroporation was sufficient to effectively edit the therapeutically relevant locus for sickle cell disease in hematopoietic stem and progenitor cells (HSPC) in a dose dependent manner without cellular toxicity. Significantly, these sCE_CBE RNPs permitted for the transplantation of edited HSPCs confirming highly efficient editing in engrafting hematopoietic stem cells in mice. The success of the designed sCBE editors, with improved solubility and enhanced on-target editing, demonstrates promising agents for cytosine base editing at other disease-related sites in HSPCs and other cell types.

Addressing the dNTP bottleneck restricting prime editing activity

Prime editing efficiency is modest in cells that are quiescent or slowly proliferating where intracellular dNTP levels are tightly regulated. MMLV-reverse transcriptase - the prime editor polymerase subunit - requires high intracellular dNTPs levels for efficient polymerization. We report that prime editing efficiency in primary cells and in vivo is increased by mutations that enhance the enzymatic properties of MMLV-reverse transcriptase and can be further complemented by targeting SAMHD1 for degradation.

Systematic Analyses of the Resistance Potential of Drugs Targeting SARS-CoV-2 Main Protease

Drugs that target the main protease (Mpro) of SARS-CoV-2 are effective therapeutics that have entered clinical use. Wide-scale use of these drugs will apply selection pressure for the evolution of resistance mutations. To understand resistance potential in Mpro, we performed comprehensive surveys of amino acid changes that can cause resistance to nirmatrelvir (Pfizer), and ensitrelvir (Xocova) in a yeast screen. We identified 142 resistance mutations for nirmatrelvir and 177 for ensitrelvir, many of which have not been previously reported. Ninety-nine mutations caused apparent resistance to both inhibitors, suggesting likelihood for the evolution of cross-resistance. The mutation with the strongest drug resistance score against nirmatrelvir in our study (E166V) was the most impactful resistance mutation recently reported in multiple viral passaging studies. Many mutations that exhibited inhibitor-specific resistance were consistent with the distinct interactions of each inhibitor in the substrate binding site. In addition, mutants with strong drug resistance scores tended to have reduced function. Our results indicate that strong pressure from nirmatrelvir or ensitrelvir will select for multiple distinct-resistant lineages that will include both primary resistance mutations that weaken interactions with drug while decreasing enzyme function and compensatory mutations that increase enzyme activity. The comprehensive identification of resistance mutations enables the design of inhibitors with reduced potential of developing resistance and aids in the surveillance of drug resistance in circulating viral populations.

FruitFire: a luciferase based on a fruit fly metabolic enzyme

Firefly luciferase is homologous to fatty acyl-CoA synthetases from insects that are not bioluminescent. Here, we determined the crystal structure of the fruit fly fatty acyl-CoA synthetase CG6178 to 2.5 Å. Based on this structure, we mutated a steric protrusion in the active site to create the artificial luciferase FruitFire, which prefers the synthetic luciferin CycLuc2 to d-luciferin by >1000-fold. FruitFire enabled in vivo bioluminescence imaging in the brains of mice using the pro-luciferin CycLuc2-amide. The conversion of a fruit fly enzyme into a luciferase capable of in vivo imaging underscores the potential for bioluminescence with a range of adenylating enzymes from nonluminescent organisms, and the possibilities for application-focused design of enzyme-substrate pairs.

HIV-1 protease inhibitors with a P1 phosphonate modification maintain potency against drug-resistant variants by increased interactions with flap residues

Protease inhibitors are the most potent antivirals against HIV-1, but they still lose efficacy against resistant variants. Improving the resistance profile is key to developing more robust inhibitors, which may be promising candidates for simplified next-generation antiretroviral therapies. In this study, we explored analogs of darunavir with a P1 phosphonate modification in combination with increasing size of the P1' hydrophobic group and various P2' moieties to improve potency against resistant variants. The phosphonate moiety substantially improved potency against highly mutated and resistant HIV-1 protease variants, but only when combined with more hydrophobic moieties at the P1' and P2' positions. Phosphonate analogs with a larger hydrophobic P1' moiety maintained excellent antiviral potency against a panel of highly resistant HIV-1 variants, with significantly improved resistance profiles. The cocrystal structures indicate that the phosphonate moiety makes extensive hydrophobic interactions with the protease, especially with the flap residues. Many residues involved in these protease-inhibitor interactions are conserved, enabling the inhibitors to maintain potency against highly resistant variants. These results highlight the need to balance inhibitor physicochemical properties by simultaneous modification of chemical groups to further improve resistance profiles.

Non-canonical pattern recognition of a pathogen-derived metabolite by a nuclear hormone receptor identifies virulent bacteria in C. elegans

Distinguishing infectious pathogens from harmless microorganisms is essential for animal health. The mechanisms used to identify infectious microbes are not fully understood, particularly in metazoan hosts that eat bacteria as their food source. Here, we characterized a non-canonical pattern-recognition system in Caenorhabditis elegans (C. elegans) that assesses the relative threat of virulent Pseudomonas aeruginosa (P. aeruginosa) to activate innate immunity. We discovered that the innate immune response in C. elegans was triggered by phenazine-1-carboxamide (PCN), a toxic metabolite produced by pathogenic strains of P. aeruginosa. We identified the nuclear hormone receptor NHR-86/HNF4 as the PCN sensor in C. elegans and validated that PCN bound to the ligand-binding domain of NHR-86/HNF4. Activation of NHR-86/HNF4 by PCN directly engaged a transcriptional program in intestinal epithelial cells that protected against P. aeruginosa. Thus, a bacterial metabolite is a pattern of pathogenesis surveilled by nematodes to identify a pathogen in its bacterial diet.

Crystal Structures of Inhibitor-Bound Main Protease from Delta- and Gamma-Coronaviruses

With the spread of SARS-CoV-2 throughout the globe causing the COVID-19 pandemic, the threat of zoonotic transmissions of coronaviruses (CoV) has become even more evident. As human infections have been caused by alpha- and beta-CoVs, structural characterization and inhibitor design mostly focused on these two genera. However, viruses from the delta and gamma genera also infect mammals and pose a potential zoonotic transmission threat. Here, we determined the inhibitor-bound crystal structures of the main protease (Mpro) from the delta-CoV porcine HKU15 and gamma-CoV SW1 from the beluga whale. A comparison with the apo structure of SW1 Mpro, which is also presented here, enabled the identification of structural arrangements upon inhibitor binding at the active site. The cocrystal structures reveal binding modes and interactions of two covalent inhibitors, PF-00835231 (active form of lufotrelvir) bound to HKU15, and GC376 bound to SW1 Mpro. These structures may be leveraged to target diverse coronaviruses and toward the structure-based design of pan-CoV inhibitors.

Selection of HIV-1 for resistance to fifth generation protease inhibitors reveals two independent pathways to high-level resistance

Darunavir (DRV) is exceptional among potent HIV-1 protease inhibitors (PIs) in high drug concentrations that are achived in vivo. Little is known about the de novo resistance pathway for DRV. We selected for resistance to high drug concentrations against ten PIs and their structural precursor DRV. Mutations accumulated through two pathways (anchored by protease mutations I50V or I84V). Small changes in the inhibitor P1'-equivalent position led to preferential use of one pathway over the other. Changes in the inhbitor P2'-equivalent position determined differences in potency that were retained in the resistant viruses and that impacted the selected mutations. Viral variants from the two pathways showed differential selection of compensatory mutations in Gag cleavage sites. These results reveal the high level of selective pressure that is attainable with fifth generation PIs, and how features of the inhibitor affect both the resistance pathway and the residual potency in the face of resistance.

Allosteric quinoxaline-based inhibitors of the flavivirus NS2B/NS3 protease

Viruses from the Flavivirus genus infect millions of people worldwide and cause severe diseases, including recent epidemics of dengue virus (DENV), and Zika virus (ZIKV). There is currently no antiviral treatment against flavivirus infections, despite considerable efforts to develop inhibitors against essential viral enzymes including NS2B/NS3 protease. Targeting the flavivirus NS2B/NS3 protease proved to be challenging because of the conformational dynamics, topology, and electrostatic properties of the active site. Here, we report the identification of quinoxaline-based allosteric inhibitors by fragment-based drug discovery approach as a promising new drug-like scaffold to target the NS2B/NS3 protease. Enzymatic assays and mutational analysis of the allosteric site in ZIKV NS2B/NS3 protease support noncompetitive inhibition mechanism as well as engineered DENV protease construct indicating the compounds likely compete with the NS2B cofactor for binding to the protease domain. Furthermore, antiviral activity confirmed the therapeutic potential of this new inhibitor scaffold.

Dual Inhibitors of Main Protease (MPro) and Cathepsin L as Potent Antivirals against SARS-CoV2

Given the current impact of SARS-CoV2 and COVID-19 on human health and the global economy, the development of direct acting antivirals is of paramount importance. Main protease (MPro), a cysteine protease that cleaves the viral polyprotein, is essential for viral replication. Therefore, MPro is a novel therapeutic target. We identified two novel MPro inhibitors, D-FFRCMKyne and D-FFCitCMKyne, that covalently modify the active site cysteine (C145) and determined cocrystal structures. Medicinal chemistry efforts led to SM141 and SM142, which adopt a unique binding mode within the MPro active site. Notably, these inhibitors do not inhibit the other cysteine protease, papain-like protease (PLPro), involved in the life cycle of SARS-CoV2. SM141 and SM142 block SARS-CoV2 replication in hACE2 expressing A549 cells with IC50 values of 8.2 and 14.7 nM. Detailed studies indicate that these compounds also inhibit cathepsin L (CatL), which cleaves the viral S protein to promote viral entry into host cells. Detailed biochemical, proteomic, and knockdown studies indicate that the antiviral activity of SM141 and SM142 results from the dual inhibition of MPro and CatL. Notably, intranasal and intraperitoneal administration of SM141 and SM142 lead to reduced viral replication, viral loads in the lung, and enhanced survival in SARS-CoV2 infected K18-ACE2 transgenic mice. In total, these data indicate that SM141 and SM142 represent promising scaffolds on which to develop antiviral drugs against SARS-CoV2.

Structure of the catalytically active APOBEC3G bound to a DNA oligonucleotide inhibitor reveals tetrahedral geometry of the transition state

APOBEC3 proteins (A3s) are enzymes that catalyze the deamination of cytidine to uridine in single-stranded DNA (ssDNA) substrates, thus playing a key role in innate antiviral immunity. However, the APOBEC3 family has also been linked to many mutational signatures in cancer cells, which has led to an intense interest to develop inhibitors of A3's catalytic activity as therapeutics as well as tools to study A3's biochemistry, structure, and cellular function. Recent studies have shown that ssDNA containing 2'-deoxy-zebularine (dZ-ssDNA) is an inhibitor of A3s such as A3A, A3B, and A3G, although the atomic determinants of this activity have remained unknown. To fill this knowledge gap, we determined a 1.5 Å resolution structure of a dZ-ssDNA inhibitor bound to active A3G. The crystal structure revealed that the activated dZ-H₂O mimics the transition state by coordinating the active site Zn²⁺ and engaging in additional stabilizing interactions, such as the one with the catalytic residue E259. Therefore, this structure allowed us to capture a snapshot of the A3's transition state and suggests that developing transition-state mimicking inhibitors may provide a new opportunity to design more targeted molecules for A3s in the future.

Mucosal nanobody IgA as inhalable and affordable prophylactic and therapeutic treatment against SARS-CoV-2 and emerging variants

Anti-COVID antibody therapeutics have been developed but not widely used due to their high cost and escape of neutralization from the emerging variants. Here, we describe the development of VHH-IgA1.1, a nanobody IgA fusion molecule as an inhalable, affordable and less invasive prophylactic and therapeutic treatment against SARS-CoV-2 Omicron variants. VHH-IgA1.1 recognizes a conserved epitope of SARS-CoV-2 spike protein Receptor Binding Domain (RBD) and potently neutralizes major global SARS-CoV-2 variants of concern (VOC) including the Omicron variant and its sub lineages BA.1.1, BA.2 and BA.2.12.1. VHH-IgA1.1 is also much more potent against Omicron variants as compared to an IgG Fc fusion construct, demonstrating the importance of IgA mediated mucosal protection for Omicron infection. Intranasal administration of VHH-IgA1.1 prior to or after challenge conferred significant protection from severe respiratory disease in K18-ACE2 transgenic mice infected with SARS-CoV-2 VOC. More importantly, for cost-effective production, VHH-IgA1.1 produced in Pichia pastoris had comparable potency to mammalian produced antibodies. Our study demonstrates that intranasal administration of affordably produced VHH-IgA fusion protein provides effective mucosal immunity against infection of SARS-CoV-2 including emerging variants.

Defining the substrate envelope of SARS-CoV-2 main protease to predict and avoid drug resistance

Coronaviruses can evolve and spread rapidly to cause severe disease morbidity and mortality, as exemplified by SARS-CoV-2 variants of the COVID-19 pandemic. Although currently available vaccines remain mostly effective against SARS-CoV-2 variants, additional treatment strategies are needed. Inhibitors that target essential viral enzymes, such as proteases and polymerases, represent key classes of antivirals. However, clinical use of antiviral therapies inevitably leads to emergence of drug resistance. In this study we implemented a strategy to pre-emptively address drug resistance to protease inhibitors targeting the main protease (M^pro) of SARS-CoV-2, an essential enzyme that promotes viral maturation. We solved nine high-resolution cocrystal structures of SARS-CoV-2 M^pro bound to substrate peptides and six structures with cleavage products. These structures enabled us to define the substrate envelope of M^pro, map the critical recognition elements, and identify evolutionarily vulnerable sites that may be susceptible to resistance mutations that would compromise binding of the newly developed M^pro inhibitors. Our results suggest strategies for developing robust inhibitors against SARS-CoV-2 that will retain longer-lasting efficacy against this evolving viral pathogen.

Comprehensive fitness landscape of SARS-CoV-2 Mpro reveals insights into viral resistance mechanisms

With the continual evolution of new strains of SARS-CoV-2 that are more virulent, transmissible, and able to evade current vaccines, there is an urgent need for effective anti-viral drugs SARS-CoV-2 main protease (M^pro) is a leading target for drug design due to its conserved and indispensable role in the viral life cycle. Drugs targeting M^pro appear promising but will elicit selection pressure for resistance. To understand resistance potential in M^pro, we performed a comprehensive mutational scan of the protease that analyzed the function of all possible single amino acid changes. We developed three separate high-throughput assays of M^pro function in yeast, based on either the ability of M^pro variants to cleave at a defined cut-site or on the toxicity of their expression to yeast. We used deep sequencing to quantify the functional effects of each variant in each screen. The protein fitness landscapes from all three screens were strongly correlated, indicating that they captured the biophysical properties critical to M^pro function. The fitness landscapes revealed a non-active site location on the surface that is extremely sensitive to mutation making it a favorable location to target with inhibitors. In addition, we found a network of critical amino acids that physically bridge the two active sites of the M^pro dimer. The clinical variants of M^pro were predominantly functional in our screens, indicating that M^pro is under strong selection pressure in the human population. Our results provide predictions of mutations that will be readily accessible to M^pro evolution and that are likely to contribute to drug resistance. This complete mutational guide of M^pro can be used in the design of inhibitors with reduced potential of evolving viral resistance.

HIV-1 VIF and human APOBEC3G interaction directly observed through molecular specific labeling using a new dual promotor vector

Over the last few decades, protein NMR isotope labeling methods using E. coli based expression have revolutionized the information accessible from biomolecular NMR experiments. Selective labeling of a protein of interest in a multi-protein complex can significantly reduce the number of cross-peaks and allow for study of large protein complexes. However, limitations still remain since some proteins are not stable independently and cannot be separately labeled in either NMR active isotope enriched or unenriched media and reconstituted into a multimeric complex. To overcome this limitation, the LEGO NMR method was previously developed using protein expression plasmids containing T7 or araBAD promoters to separately express proteins in the same E. coli after changing between labeled and unlabeled media. Building on this, we developed a method to label the Human Immunodeficiency Virus type 1 viral infectivity factor (HIV-1 Vif), a monomerically unstable protein, in complex with CBFβ, it's host binding partner. We designed a dual promoter plasmid containing both T7 and araBAD promoters to independently control the expression of HIV-1 Vif in NMR active isotope enriched media and CBFβ in unenriched media. Using this method, we assigned the backbone resonance and directly observed the binding of HIV-1 Vif with APOBEC3G, a host restriction factor to HIV-1.

Quantitative structural analysis of influenza virus by cryo-electron tomography and convolutional neural networks

Influenza viruses pose severe public health threats globally. Influenza viruses are extensively pleomorphic, in shape, size, and organization of viral proteins. Analysis of influenza morphology and ultrastructure can help elucidate viral structure-function relationships and aid in therapeutics and vaccine development. While cryo-electron tomography (cryoET) can depict the 3D organization of pleomorphic influenza, the low signal-to-noise ratio inherent to cryoET and viral heterogeneity have precluded detailed characterization of influenza viruses. In this report, we leveraged convolutional neural networks and cryoET to characterize the morphological architecture of the A/Puerto Rico/8/34 (H1N1) influenza strain. Our pipeline improved the throughput of cryoET analysis and accurately identified viral components within tomograms. Using this approach, we successfully characterized influenza morphology, glycoprotein density, and conducted subtomogram averaging of influenza glycoproteins. Application of this processing pipeline can aid in the structural characterization of not only influenza viruses, but other pleomorphic viruses and infected cells.

Deciphering the Molecular Mechanism of HCV Protease Inhibitor Fluorination as a General Approach to Avoid Drug Resistance

Third generation Hepatitis C virus (HCV) NS3/4A protease inhibitors (PIs), glecaprevir and voxilaprevir, are highly effective across genotypes and against many resistant variants. Unlike earlier PIs, these compounds have fluorine substitutions on the P2-P4 macrocycle and P1 moieties. Fluorination has long been used in medicinal chemistry as a strategy to improve physicochemical properties and potency. However, the molecular basis by which fluorination improves potency and resistance profile of HCV NS3/4A PIs is not well understood. To systematically analyze the contribution of fluorine substitutions to inhibitor potency and resistance profile, we used a multi-disciplinary approach involving inhibitor design and synthesis, enzyme inhibition assays, co-crystallography, and structural analysis. A panel of inhibitors in matched pairs were designed with and without P4 cap fluorination, tested against WT protease and the D168A resistant variant, and a total of 22 high-resolution co-crystal structures were determined. While fluorination did not significantly improve potency against the WT protease, PIs with fluorinated P4 caps retained much better potency against the D168A protease variant. Detailed analysis of the co-crystal structures revealed that PIs with fluorinated P4 caps can sample alternate binding conformations that enable adapting to structural changes induced by the D168A substitution. Our results elucidate molecular mechanisms of fluorine-specific inhibitor interactions that can be leveraged in avoiding drug resistance.

Analyses of HIV proteases variants at the threshold of viability reveals relationships between processing efficiency and fitness

Investigating the relationships between protein function and fitness provides keys for understanding biochemical mechanisms that underly evolution. Mutations with partial fitness defects can delineate the threshold of biochemical function required for viability. We utilized a previous deep mutational scan of HIV-1 protease (PR) to identify variants with 15–45 per cent defects in replication and analysed the biochemical function of eight variants (L10M, L10S, V32C, V32I, A71V, A71S, Q92I, Q92N). We purified each variant and assessed the efficiency of peptide cleavage for three cut sites (MA-CA, TF-PR, and PR-RT) as well as gel-based analyses of processing of purified Gag. The cutting activity of at least one site was perturbed relative to WT protease for all variants, consistent with cutting activity being a primary determinant of fitness effects. We examined the correlation of fitness defects with cutting activity of different sites. MA-CA showed the weakest correlation (R² = 0.02) with fitness, suggesting relatively weak coupling with viral replication. In contrast, cutting of the TF-PR site showed the strongest correlation with fitness (R² = 0.53). Cutting at the TF-PR site creates a new PR protein with a free N-terminus that is critical for activity. Our findings indicate that increasing the pool of active PR is rate limiting for viral replication, making this an ideal step to target with inhibitors.

Viral proteases: Structure, mechanism and inhibition

Viral proteases are diverse in structure, oligomeric state, catalytic mechanism, and substrate specificity. This chapter focuses on proteases from viruses that are relevant to human health: human immunodeficiency virus subtype 1 (HIV-1), hepatitis C (HCV), human T-cell leukemia virus type 1 (HTLV-1), flaviviruses, enteroviruses, and coronaviruses. The proteases of HIV-1 and HCV have been successfully targeted for therapeutics, with picomolar FDA-approved drugs currently used in the clinic. The proteases of HTLV-1 and the other virus families remain emerging therapeutic targets at different stages of the drug development process. This chapter provides an overview of the current knowledge on viral protease structure, mechanism, substrate recognition, and inhibition. Particular focus is placed on recent advances in understanding the molecular basis of diverse substrate recognition and resistance, which is essential toward designing novel protease inhibitors as antivirals.

Pan-3C Protease Inhibitor Rupintrivir Binds SARS-CoV-2 Main Protease in a Unique Binding Mode

Rupintrivir targets the 3C cysteine proteases of the picornaviridae family, which includes rhinoviruses and enteroviruses that cause a range of human diseases. Despite being a pan-3C protease inhibitor, rupintrivir activity is extremely weak against the homologous 3C-like protease of SARS-CoV-2. In this study, the crystal structures of rupintrivir were determined bound to enterovirus 68 (EV68) 3C protease and the 3C-like main protease (Mpro) from SARS-CoV-2. While the EV68 3C protease-rupintrivir structure was similar to previously determined complexes with other picornavirus 3C proteases, rupintrivir bound in a unique conformation to the active site of SARS-CoV-2 Mpro splitting the catalytic cysteine and histidine residues. This bifurcation of the catalytic dyad may provide a novel approach for inhibiting cysteine proteases.

Discovery of Quinoxaline-Based P1-P3 Macrocyclic NS3/4A Protease Inhibitors with Potent Activity against Drug-Resistant Hepatitis C Virus Variants

The three pan-genotypic HCV NS3/4A protease inhibitors (PIs) currently in clinical use-grazoprevir, glecaprevir, and voxilaprevir-are quinoxaline-based P2-P4 macrocycles and thus exhibit similar resistance profiles. Using our quinoxaline-based P1-P3 macrocyclic lead compounds as an alternative chemical scaffold, we explored structure-activity relationships (SARs) at the P2 and P4 positions to develop pan-genotypic PIs that avoid drug resistance. A structure-guided strategy was used to design and synthesize two series of compounds with different P2 quinoxalines in combination with diverse P4 groups of varying sizes and shapes, with and without fluorine substitutions. Our SAR data and cocrystal structures revealed the interplay between the P2 and P4 groups, which influenced inhibitor binding and the overall resistance profile. Optimizing inhibitor interactions in the S4 pocket led to PIs with excellent antiviral activity against clinically relevant PI-resistant HCV variants and genotype 3, providing potential pan-genotypic inhibitors with improved resistance profiles.

Affinity maturation of SARS-CoV-2 neutralizing antibodies confers potency, breadth, and resilience to viral escape mutations

Antibodies elicited by infection accumulate somatic mutations in germinal centers that can increase affinity for cognate antigens. We analyzed 6 independent groups of clonally related severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) Spike receptor-binding domain (RBD)-specific antibodies from 5 individuals shortly after infection and later in convalescence to determine the impact of maturation over months. In addition to increased affinity and neutralization potency, antibody evolution changed the mutational pathways for the acquisition of viral resistance and restricted neutralization escape options. For some antibodies, maturation imposed a requirement for multiple substitutions to enable escape. For certain antibodies, affinity maturation enabled the neutralization of circulating SARS-CoV-2 variants of concern and heterologous sarbecoviruses. Antibody-antigen structures revealed that these properties resulted from substitutions that allowed additional variability at the interface with the RBD. These findings suggest that increasing antibody diversity through prolonged or repeated antigen exposure may improve protection against diversifying SARS-CoV-2 populations, and perhaps against other pandemic threat coronaviruses.

Structural basis of substrate specificity in human cytidine deaminase family APOBEC3s

The human cytidine deaminase family of APOBEC3s (A3s) plays critical roles in both innate immunity and the development of cancers. A3s comprise seven functionally overlapping but distinct members that can be exploited as nucleotide base editors for treating genetic diseases. Although overall structurally similar, A3s have vastly varying deamination activity and substrate preferences. Recent crystal structures of ssDNA-bound A3s together with experimental studies have provided some insights into distinct substrate specificities among the family members. However, the molecular interactions responsible for their distinct biological functions and how structure regulates substrate specificity are not clear. In this study, we identified the structural basis of substrate specificities in three catalytically active A3 domains whose crystal structures have been previously characterized: A3A, A3B- CTD, and A3G-CTD. Through molecular modeling and dynamic simulations, we found an interdependency between ssDNA substrate binding conformation and nucleotide sequence specificity. In addition to the U-shaped conformation seen in the crystal structure with the CTC0 motif, A3A can accommodate the CCC0 motif when ssDNA is in a more linear (L) conformation. A3B can also bind both U- and L-shaped ssDNA, unlike A3G, which can stably recognize only linear ssDNA. These varied conformations are stabilized by sequence-specific interactions with active site loops 1 and 7, which are highly variable among A3s. Our results explain the molecular basis of previously observed substrate specificities in A3s and have implications for designing A3-specific inhibitors for cancer therapy as well as engineering base-editing systems for gene therapy.

Report of the National Institutes of Health SARS-CoV-2 Antiviral Therapeutics Summit

The NIH Virtual SARS-CoV-2 Antiviral Summit, held on 6 November 2020, was organized to provide an overview on the status and challenges in developing antiviral therapeutics for coronavirus disease 2019 (COVID-19), including combinations of antivirals. Scientific experts from the public and private sectors convened virtually during a live videocast to discuss severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) targets for drug discovery as well as the preclinical tools needed to develop and evaluate effective small-molecule antivirals. The goals of the Summit were to review the current state of the science, identify unmet research needs, share insights and lessons learned from treating other infectious diseases, identify opportunities for public-private partnerships, and assist the research community in designing and developing antiviral therapeutics. This report includes an overview of therapeutic approaches, individual panel summaries, and a summary of the discussions and perspectives on the challenges ahead for antiviral development.

Deciphering Antifungal Drug Resistance in Pneumocystis jirovecii DHFR with Molecular Dynamics and Machine Learning

Drug resistance impacts the effectiveness of many new therapeutics. Mutations in the therapeutic target confer resistance; however, deciphering which mutations, often remote from the enzyme active site, drive resistance is challenging. In a series of Pneumocystis jirovecii dihydrofolate reductase variants, we elucidate which interactions are key bellwethers to confer resistance to trimethoprim using homology modeling, molecular dynamics, and machine learning. Six molecular features involving mainly residues that did not vary were the best indicators of resistance.

Deciphering Complex Mechanisms of Resistance and Loss of Potency through Coupled Molecular Dynamics and Machine Learning

Drug resistance threatens many critical therapeutics through mutations in the drug target. The molecular mechanisms by which combinations of mutations, especially those remote from the active site, alter drug binding to confer resistance are poorly understood and thus difficult to counteract. A machine learning strategy was developed that coupled parallel molecular dynamics simulations with experimental potency to identify specific conserved mechanisms underlying resistance. Physical features were extracted from the simulations, analyzed, and integrated into one consistent and interpretable elastic network model. To rigorously test this strategy, HIV-1 protease variants with diverse mutations were used, with potencies ranging from picomolar to micromolar to the drug darunavir. Feature reduction resulted in a model with four specific features that predicts for both the training and test sets inhibitor binding free energy within 1 kcal/mol of the experimental value over this entire range of potency. These predictive features are physically interpretable, as they vary specifically with affinity and diagonally transverse across the protease homodimer. This physics-based strategy of parallel molecular dynamics and machine learning captures mechanisms by which complex combinations of mutations confer resistance and identify critical features that serve as bellwethers of affinity, which will be critical in future drug design.

Drug Design Strategies to Avoid Resistance in Direct-Acting Antivirals and Beyond

Drug resistance is prevalent across many diseases, rendering therapies ineffective with severe financial and health consequences. Rather than accepting resistance after the fact, proactive strategies need to be incorporated into the drug design and development process to minimize the impact of drug resistance. These strategies can be derived from our experience with viral disease targets where multiple generations of drugs had to be developed to combat resistance and avoid antiviral failure. Significant efforts including experimental and computational structural biology, medicinal chemistry, and machine learning have focused on understanding the mechanisms and structural basis of resistance against direct-acting antiviral (DAA) drugs. Integrated methods show promise for being predictive of resistance and potency. In this review, we give an overview of this research for human immunodeficiency virus type 1, hepatitis C virus, and influenza virus and the lessons learned from resistance mechanisms of DAAs. These lessons translate into rational strategies to avoid resistance in drug design, which can be generalized and applied beyond viral targets. While resistance may not be completely avoidable, rational drug design can and should incorporate strategies at the outset of drug development to decrease the prevalence of drug resistance.

Inhibiting HTLV-1 Protease: A Viable Antiviral Target

Human T-cell lymphotropic virus type 1 (HTLV-1) is a retrovirus that can cause severe paralytic neurologic disease and immune disorders as well as cancer. An estimated 20 million people worldwide are infected with HTLV-1, with prevalence reaching 30% in some parts of the world. In stark contrast to HIV-1, no direct acting antivirals (DAAs) exist against HTLV-1. The aspartyl protease of HTLV-1 is a dimer similar to that of HIV-1 and processes the viral polyprotein to permit viral maturation. We report that the FDA-approved HIV-1 protease inhibitor darunavir (DRV) inhibits the enzyme with 0.8 μM potency and provides a scaffold for drug design against HTLV-1. Analogs of DRV that we designed and synthesized achieved submicromolar inhibition against HTLV-1 protease and inhibited Gag processing in viral maturation assays and in a chronically HTLV-1 infected cell line. Cocrystal structures of these inhibitors with HTLV-1 protease highlight opportunities for future inhibitor design. Our results show promise toward developing highly potent HTLV-1 protease inhibitors as therapeutic agents against HTLV-1 infections.

Development of potency, breadth and resilience to viral escape mutations in SARS-CoV-2 neutralizing antibodies

Antibodies elicited in response to infection undergo somatic mutation in germinal centers that can result in higher affinity for the cognate antigen. To determine the effects of somatic mutation on the properties of SARS-CoV-2 spike receptor-binding domain (RBD)-specific antibodies, we analyzed six independent antibody lineages. As well as increased neutralization potency, antibody evolution changed pathways for acquisition of resistance and, in some cases, restricted the range of neutralization escape options. For some antibodies, maturation apparently imposed a requirement for multiple spike mutations to enable escape. For certain antibody lineages, maturation enabled neutralization of circulating SARS-CoV-2 variants of concern and heterologous sarbecoviruses. Antibody-antigen structures revealed that these properties resulted from substitutions that allowed additional variability at the interface with the RBD. These findings suggest that increasing antibody diversity through prolonged or repeated antigen exposure may improve protection against diversifying SARS-CoV-2 populations, and perhaps against other pandemic threat coronaviruses.

NAD(H) phosphates mediate tetramer assembly of human C-terminal binding protein (CtBP)

C-terminal binding proteins (CtBPs) are cotranscriptional factors that play key roles in cell fate. We have previously shown that NAD(H) promotes the assembly of similar tetramers from either human CtBP1 and CtBP2 and that CtBP2 tetramer destabilizing mutants are defective for oncogenic activity. To assist structure-based design efforts for compounds that disrupt CtBP tetramerization, it is essential to understand how NAD(H) triggers tetramer assembly. Here, we investigate the moieties within NAD(H) that are responsible for triggering tetramer formation. Using multiangle light scattering (MALS), we show that ADP is able to promote tetramer formation of both CtBP1 and CtBP2, whereas AMP promotes tetramer assembly of CtBP1, but not CtBP2. Other NAD(H) moieties that lack the adenosine phosphate, including adenosine and those incorporating nicotinamide, all fail to promote tetramer assembly. Our crystal structures of CtBP1 with AMP reveal participation of the adenosine phosphate in the tetrameric interface, pinpointing its central role in NAD(H)-linked assembly. CtBP1 and CtBP2 have overlapping but unique roles, suggesting that a detailed understanding of their unique structural properties might have utility in the design of paralog-specific inhibitors. We investigated the different responses to AMP through a series of site-directed mutants at 13 positions. These mutations reveal a central role for a hinge segment, which we term the 120s hinge that connects the substrate with coenzyme-binding domains and influences nucleotide binding and tetramer assembly. Our results provide insight into suitable pockets to explore in structure-based drug design to interfere with cotranscriptional activity of CtBP in cancer.

Crystal Structure of SARS-CoV-2 Main Protease in Complex with the Non-Covalent Inhibitor ML188

Viral proteases are critical enzymes for the maturation of many human pathogenic viruses and thus are key targets for direct acting antivirals (DAAs). The current viral pandemic caused by SARS-CoV-2 is in dire need of DAAs. The Main protease (Mpro) is the focus of extensive structure-based drug design efforts which are mostly covalent inhibitors targeting the catalytic cysteine. ML188 is a non-covalent inhibitor designed to target SARS-CoV-1 Mpro, and provides an initial scaffold for the creation of effective pan-coronavirus inhibitors. In the current study, we found that ML188 inhibits SARS-CoV-2 Mpro at 2.5 µM, which is more potent than against SAR-CoV-1 Mpro. We determined the crystal structure of ML188 in complex with SARS-CoV-2 Mpro to 2.39 Å resolution. Sharing 96% sequence identity, structural comparison of the two complexes only shows subtle differences. Non-covalent protease inhibitors complement the design of covalent inhibitors against SARS-CoV-2 main protease and are critical initial steps in the design of DAAs to treat CoVID 19.

Unique structural solution from a VH3-30 antibody targeting the hemagglutinin stem of influenza A viruses

Broadly neutralizing antibodies (bnAbs) targeting conserved influenza A virus (IAV) hemagglutinin (HA) epitopes can provide valuable information for accelerating universal vaccine designs. Here, we report structural details for heterosubtypic recognition of HA from circulating and emerging IAVs by the human antibody 3I14. Somatic hypermutations play a critical role in shaping the HCDR3, which alone and uniquely among VH3-30 derived antibodies, forms contacts with five sub-pockets within the HA-stem hydrophobic groove. 3I14 light-chain interactions are also key for binding HA and contribute a large buried surface area spanning two HA protomers. Comparison of 3I14 to bnAbs from several defined classes provide insights to the bias selection of VH3-30 antibodies and reveals that 3I14 represents a novel structural solution within the VH3-30 repertoire. The structures reported here improve our understanding of cross-group heterosubtypic binding activity, providing the basis for advancing immunogen designs aimed at eliciting a broadly protective response to IAV.

Interactions of APOBEC3s with DNA and RNA

APOBEC3 enzymes are key enzymes in our innate immune system regulating antiviral response in HIV and unfortunately adding diversity in cancer as they deaminate cytosine. Seven unique single and double domain APOBEC3s provide them with unique activity and specificity profiles for this deamination. Recent crystal and NMR structures of APOBEC3 complexes are unraveling the variety of epitopes involved in binding nucleic acids, including at the catalytic site, elsewhere on the catalytic domain and in the inactive N-terminal domain. The interplay between these diverse interactions is critical to uncovering the mechanisms by which APOBEC3s recognize and process their substrates.

Cryo-EM structure of CtBP2 confirms tetrameric architecture

C-terminal binding proteins 1 and 2 (CtBP1 and CtBP2) are transcriptional regulators that activate or repress many genes involved in cellular development, apoptosis, and metastasis. NADH-dependent CtBP activation has been implicated in multiple types of cancer and poor patient prognosis. Central to understanding activation of CtBP in oncogenesis is uncovering how NADH triggers protein assembly, what level of assembly occurs, and if oncogenic activity depends upon such assembly. Here, we present the cryoelectron microscopic structures of two different constructs of CtBP2 corroborating that the native state of CtBP2 in the presence of NADH is tetrameric. The physiological relevance of the observed tetramer was demonstrated in cell culture, showing that CtBP tetramer-destabilizing mutants are defective for cell migration, transcriptional repression of E-cadherin, and activation of TIAM1. Together with our cryoelectron microscopy studies, these results highlight the tetramer as the functional oligomeric form of CtBP2.

Crystal Structure of a Soluble APOBEC3G Variant Suggests ssDNA to Bind in a Channel that Extends between the Two Domains

APOBEC3G (A3G) is a single-stranded DNA (ssDNA) cytosine deaminase that can restrict HIV-1 infection by mutating the viral genome. A3G consists of a non-catalytic N-terminal domain (NTD) and a catalytic C-terminal domain (CTD) connected by a short linker. While the CTD catalyzes cytosine deamination, the NTD is believed to provide additional affinity for ssDNA. Structures of both A3G domains have been solved individually; however, a full-length A3G structure has been challenging. Recently, crystal structures of full-length rhesus macaque A3G variants were solved which suggested dimerization mechanisms and RNA binding surfaces, whereas the dimerization appeared to compromise catalytic activity. We determined the crystal structure of a soluble variant of human A3G (sA3G) at 2.5 Å and from these data generated a model structure of wild-type A3G. This model demonstrated that the NTD was rotated 90° relative to the CTD along the major axis of the molecule, an orientation that forms a positively charged channel connected to the CTD catalytic site, consisting of NTD loop-1 and CTD loop-3. Structure-based mutations, in vitro deamination and DNA binding assays, and HIV-1 restriction assays identify R24, located in the NTD loop-1, as essential to a critical interaction with ssDNA. Furthermore, sA3G was shown to bind a deoxy-cytidine dinucleotide near the catalytic Zn²⁺, yet not in the catalytic position, where the interactions between deoxy-cytidines and CTD loop-1 and loop-7 residues were different from those formed with substrate. These new interactions suggest a mechanism explaining why A3G exhibits a 3' to 5' directional preference in processive deamination.

Genome-scale in vivo CRISPR screen identifies RNLS as a target for beta cell protection in type 1 diabetes

Type 1 diabetes (T1D) is caused by the autoimmune destruction of pancreatic beta cells. Pluripotent stem cells can now be differentiated into beta cells, thus raising the prospect of a cell replacement therapy for T1D. However, autoimmunity would rapidly destroy newly transplanted beta cells. Using a genome-scale CRISPR screen in a mouse model for T1D, we show that deleting RNLS, a genome-wide association study candidate gene for T1D, made beta cells resistant to autoimmune killing. Structure-based modelling identified the U.S. Food and Drug Administration-approved drug pargyline as a potential RNLS inhibitor. Oral pargyline treatment protected transplanted beta cells in diabetic mice, thus leading to disease reversal. Furthermore, pargyline prevented or delayed diabetes onset in several mouse models for T1D. Our results identify RNLS as a modifier of beta cell vulnerability and as a potential therapeutic target to avert beta cell loss in T1D.

A cross-reactive human IgA monoclonal antibody blocks SARS-CoV-2 spike-ACE2 interaction

COVID-19 caused by SARS-CoV-2 has become a global pandemic requiring the development of interventions for the prevention or treatment to curtail mortality and morbidity. No vaccine to boost mucosal immunity, or as a therapeutic, has yet been developed to SARS-CoV-2. In this study, we discover and characterize a cross-reactive human IgA monoclonal antibody, MAb362. MAb362 binds to both SARS-CoV and SARS-CoV-2 spike proteins and competitively blocks ACE2 receptor binding, by overlapping the ACE2 structural binding epitope. Furthermore, MAb362 IgA neutralizes both pseudotyped SARS-CoV and SARS-CoV-2 in 293 cells expressing ACE2. When converted to secretory IgA, MAb326 also neutralizes authentic SARS-CoV-2 virus while the IgG isotype shows no neutralization. Our results suggest that SARS-CoV-2 specific IgA antibodies, such as MAb362, may provide effective immunity against SARS-CoV-2 by inducing mucosal immunity within the respiratory system, a potentially critical feature of an effective vaccine.

Here, Ejemel et al. report the identification and characterization of a cross-neutralizing human IgA monoclonal antibody, named MAb362, that binds the receptor-binding domain of SARS-CoV-2 Spike, blocking its interaction with the ACE2 host receptor.

Structural Analysis of Potent Hybrid HIV-1 Protease Inhibitors Containing Bis-tetrahydrofuran in a Pseudosymmetric Dipeptide Isostere

The design, synthesis, and X-ray structural analysis of hybrid HIV-1 protease inhibitors (PIs) containing bis-tetrahydrofuran (bis-THF) in a pseudo-C₂-symmetric dipeptide isostere are described. A series of PIs were synthesized by incorporating bis-THF of darunavir on either side of the Phe-Phe isostere of lopinavir in combination with hydrophobic amino acids on the opposite P2/P2' position. Structure-activity relationship studies indicated that the bis-THF moiety can be attached at either the P2 or P2' position without significantly affecting potency. However, the group on the opposite P2/P2' position had a dramatic effect on potency depending on the size and shape of the side chain. Cocrystal structures of inhibitors with wild-type HIV-1 protease revealed that the bis-THF moiety retained similar interactions as observed in the darunavir-protease complex regardless of the position on the Phe-Phe isostere. Analyses of cocrystal structures and molecular dynamics simulations provide insights into optimizing HIV-1 PIs containing bis-THF in non-sulfonamide dipeptide isosteres.

IgA MAb blocks SARS-CoV-2 Spike-ACE2 interaction providing mucosal immunity

COVID-19 caused by SARS-CoV-2 has become a global pandemic requiring the development of interventions for the prevention or treatment to curtail mortality and morbidity. No vaccine to boost mucosal immunity or as a therapeutic has yet been developed to SARS-CoV-2. In this study we discover and characterize a cross-reactive human IgA monoclonal antibody, MAb362. MAb362 binds to both SARS-CoV and SARS-CoV-2 spike proteins and competitively blocks hACE2 receptor binding, by completely overlapping the hACE2 structural binding epitope. Furthermore, MAb362 IgA neutralizes both pseudotyped SARS-CoV and SARS-CoV-2 in human epithelial cells expressing hACE2. SARS-CoV-2 specific IgA antibodies, such as MAb362, may provide effective immunity against SARS-CoV-2 by inducing mucosal immunity within the respiratory system, a potentially critical feature of an effective vaccine.

Avoiding Drug Resistance by Substrate Envelope-Guided Design: Toward Potent and Robust HCV NS3/4A Protease Inhibitors

Hepatitis C virus (HCV) infects millions of people worldwide, causing chronic liver disease that can lead to cirrhosis, hepatocellular carcinoma, and liver transplant. In the last several years, the advent of direct-acting antivirals, including NS3/4A protease inhibitors (PIs), has remarkably improved treatment outcomes of HCV-infected patients. However, selection of resistance-associated substitutions and polymorphisms among genotypes can lead to drug resistance and in some cases treatment failure. A proactive strategy to combat resistance is to constrain PIs within evolutionarily conserved regions in the protease active site. Designing PIs using the substrate envelope is a rational strategy to decrease the susceptibility to resistance by using the constraints of substrate recognition. We successfully designed two series of HCV NS3/4A PIs to leverage unexploited areas in the substrate envelope to improve potency, specifically against resistance-associated substitutions at D168. Our design strategy achieved better resistance profiles over both the FDA-approved NS3/4A PI grazoprevir and the parent compound against the clinically relevant D168A substitution. Crystallographic structural analysis and inhibition assays confirmed that optimally filling the substrate envelope is critical to improve inhibitor potency while avoiding resistance. Specifically, inhibitors that enhanced hydrophobic packing in the S4 pocket and avoided an energetically frustrated pocket performed the best. Thus, the HCV substrate envelope proved to be a powerful tool to design robust PIs, offering a strategy that can be translated to other targets for rational design of inhibitors with improved potency and resistance profiles.IMPORTANCE Despite significant progress, hepatitis C virus (HCV) continues to be a major health problem with millions of people infected worldwide and thousands dying annually due to resulting complications. Recent antiviral combinations can achieve >95% cure, but late diagnosis, low access to treatment, and treatment failure due to drug resistance continue to be roadblocks against eradication of the virus. We report the rational design of two series of HCV NS3/4A protease inhibitors with improved resistance profiles by exploiting evolutionarily constrained regions of the active site using the substrate envelope model. Optimally filling the S4 pocket is critical to avoid resistance and improve potency. Our results provide drug design strategies to avoid resistance that are applicable to other quickly evolving viral drug targets.

Optimizing the refinement of merohedrally twinned P61 HIV-1 protease-inhibitor cocrystal structures

Twinning is a crystal-growth anomaly in which protein monomers exist in different orientations but are related in a specific way, causing diffraction reflections to overlap. Twinning imposes additional symmetry on the data, often leading to the assignment of a higher symmetry space group. Specifically, in merohedral twinning, reflections from each monomer overlap and require a twin law to model unique structural data from overlapping reflections. Neglecting twinning in the crystallographic analysis of quasi-rotationally symmetric homo-oligomeric protein structures can mask the degree of structural non-identity between monomers. In particular, any deviations from perfect symmetry will be lost if higher than appropriate symmetry is applied during crystallographic analysis. Such cases warrant choosing between the highest symmetry space group possible or determining whether the monomers have distinguishable structural asymmetries and thus require a lower symmetry space group and a twin law. Using hexagonal cocrystals of HIV-1 protease, a C2-symmetric homodimer whose symmetry is broken by bound ligand, it is shown that both assigning a lower symmetry space group and applying a twin law during refinement are critical to achieving a structural model that more accurately fits the electron density. By re-analyzing three recently published HIV-1 protease structures, improvements in nearly every crystallographic metric are demonstrated. Most importantly, a procedure is demonstrated where the inhibitor can be reliably modeled in a single orientation. This protocol may be applicable to many other homo-oligomers in the PDB.

Molecular and Structural Mechanism of Pan-Genotypic HCV NS3/4A Protease Inhibition by Glecaprevir

Hepatitis C virus, causative agent of chronic viral hepatitis, infects 71 million people worldwide and is divided into seven genotypes and multiple subtypes with sequence identities between 68 to 82%. While older generation direct-acting antivirals had varying effectiveness against different genotypes, the newest NS3/4A protease inhibitors including glecaprevir (GLE) have pan-genotypic activity. The structural basis for pan-genotypic inhibition and effects of polymorphisms on inhibitor potency were not well-known due to lack of crystal structures of GLE-bound NS3/4A or genotypes other than 1. In this study, we determined the crystal structures of NS3/4A from genotypes 1a, 3a, 4a, and 5a in complex with GLE. Comparison with the highly similar grazoprevir indicated the mechanism of GLE's drastic improvement in potency. We found that, while GLE is highly potent against wild-type NS3/4A of all genotypes, specific resistance-associated substitutions (RASs) confer orders of magnitude loss in inhibition. Our crystal structures reveal molecular mechanisms behind pan-genotypic activity of GLE, including potency loss due to RASs at D168. Our structures permit for the first time analysis of changes due to polymorphisms among genotypes, providing insights into design principles that can aid future drug development and potentially can be extended to other proteins.

Characterizing Protein-Ligand Binding Using Atomistic Simulation and Machine Learning: Application to Drug Resistance in HIV-1 Protease

Over the past several decades, atomistic simulations of biomolecules, whether carried out using molecular dynamics or Monte Carlo techniques, have provided detailed insights into their function. Comparing the results of such simulations for a few closely related systems has guided our understanding of the mechanisms by which changes such as ligand binding or mutation can alter the function. The general problem of detecting and interpreting such mechanisms from simulations of many related systems, however, remains a challenge. This problem is addressed here by applying supervised and unsupervised machine learning techniques to a variety of thermodynamic observables extracted from molecular dynamics simulations of different systems. As an important test case, these methods are applied to understand the evasion by human immunodeficiency virus type-1 (HIV-1) protease of darunavir, a potent inhibitor to which resistance can develop via the simultaneous mutation of multiple amino acids. Complex mutational patterns have been observed among resistant strains, presenting a challenge to developing a mechanistic picture of resistance in the protease. In order to dissect these patterns and gain mechanistic insight into the role of specific mutations, molecular dynamics simulations were carried out on a collection of HIV-1 protease variants, chosen to include highly resistant strains and susceptible controls, in complex with darunavir. Using a machine learning approach that takes advantage of the hierarchical nature in the relationships among the sequence, structure, and function, an integrative analysis of these trajectories reveals key details of the resistance mechanism, including changes in the protein structure, hydrogen bonding, and protein-ligand contacts.