AI-guided pipeline for protein-protein interaction drug discovery identifies a SARS-CoV-2 inhibitor.

Protein-protein interactions (PPIs) offer great opportunities to expand the druggable proteome and therapeutically tackle various diseases, but remain challenging targets for drug discovery. Here, we provide a comprehensive pipeline that combines experimental and computational tools to identify and validate PPI targets and perform early-stage drug discovery. We have developed a machine learning approach that prioritizes interactions by analyzing quantitative data from binary PPI assays or AlphaFold-Multimer predictions. Using the quantitative assay LuTHy together with our machine learning algorithm, we identified high-confidence interactions among SARS-CoV-2 proteins for which we predicted three-dimensional structures using AlphaFold-Multimer. We employed VirtualFlow to target the contact interface of the NSP10-NSP16 SARS-CoV-2 methyltransferase complex by ultra-large virtual drug screening. Thereby, we identified a compound that binds to NSP10 and inhibits its interaction with NSP16, while also disrupting the methyltransferase activity of the complex, and SARS-CoV-2 replication. Overall, this pipeline will help to prioritize PPI targets to accelerate the discovery of early-stage drug candidates targeting protein complexes and pathways.

AI-guided pipeline for protein-protein interaction drug discovery identifies a SARS-CoV-2 inhibitor.

Protein-protein interactions (PPIs) offer great opportunities to expand the druggable proteome and therapeutically tackle various diseases, but remain challenging targets for drug discovery. Here, we provide a comprehensive pipeline that combines experimental and computational tools to identify and validate PPI targets and perform early-stage drug discovery. We have developed a machine learning approach that prioritizes interactions by analyzing quantitative data from binary PPI assays and AlphaFold-Multimer predictions. Using the quantitative assay LuTHy together with our machine learning algorithm, we identified high-confidence interactions among SARS-CoV-2 proteins for which we predicted three-dimensional structures using AlphaFold Multimer. We employed VirtualFlow to target the contact interface of the NSP10-NSP16 SARS-CoV-2 methyltransferase complex by ultra-large virtual drug screening. Thereby, we identified a compound that binds to NSP10 and inhibits its interaction with NSP16, while also disrupting the methyltransferase activity of the complex, and SARS-CoV-2 replication. Overall, this pipeline will help to prioritize PPI targets to accelerate the discovery of early-stage drug candidates targeting protein complexes and pathways.

A proteome-scale map of the SARS-CoV-2-human contactome.

Understanding the mechanisms of coronavirus disease 2019 (COVID-19) disease severity to efficiently design therapies for emerging virus variants remains an urgent challenge of the ongoing pandemic. Infection and immune reactions are mediated by direct contacts between viral molecules and the host proteome, and the vast majority of these virus-host contacts (the 'contactome') have not been identified. Here, we present a systematic contactome map of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) with the human host encompassing more than 200 binary virus-host and intraviral protein-protein interactions. We find that host proteins genetically associated with comorbidities of severe illness and long COVID are enriched in SARS-CoV-2 targeted network communities. Evaluating contactome-derived hypotheses, we demonstrate that viral NSP14 activates nuclear factor κB (NF-κB)-dependent transcription, even in the presence of cytokine signaling. Moreover, for several tested host proteins, genetic knock-down substantially reduces viral replication. Additionally, we show for USP25 that this effect is phenocopied by the small-molecule inhibitor AZ1. Our results connect viral proteins to human genetic architecture for COVID-19 severity and offer potential therapeutic targets.

SKAP2 Modular Organization Differently Recognizes SRC Kinases Depending on Their Activation Status and Localization.

Dimerization of SRC kinase adaptor phosphoprotein 2 (SKAP2) induces an increase of binding for most SRC kinases suggesting a fine-tuning with transphosphorylation for kinase activation. This work addresses the molecular basis of SKAP2-mediated SRC kinase regulation through the lens of their interaction capacities. By combining a luciferase complementation assay and extensive site-directed mutagenesis, we demonstrated that SKAP2 interacts with SRC kinases through a modular organization depending both on their phosphorylation-dependent activation and subcellular localization. SKAP2 contains three interacting modules consisting in the dimerization domain, the SRC homology 3 (SH3) domain, and the second interdomain located between the Pleckstrin homology and the SH3 domains. Functionally, the dimerization domain is necessary and sufficient to bind to most activated and myristyl SRC kinases. In contrast, the three modules are necessary to bind SRC kinases at their steady state. The Pleckstrin homology and SH3 domains of SKAP2 as well as tyrosines located in the interdomains modulate these interactions. Analysis of mutants of the SRC kinase family member hematopoietic cell kinase supports this model and shows the role of two residues, Y390 and K7, on its degradation following activation. In this article, we show that a modular architecture of SKAP2 drives its interaction with SRC kinases, with the binding capacity of each module depending on both their localization and phosphorylation state activation. This work opens new perspectives on the molecular mechanisms of SRC kinases activation, which could have significant therapeutic impact.

The cargo adapter protein CLINT1 is phosphorylated by the Numb-associated kinase BIKE and mediates dengue virus infection.

The signaling pathways and cellular functions regulated by the four Numb-associated kinases are largely unknown. We reported that AAK1 and GAK control intracellular trafficking of RNA viruses and revealed a requirement for BIKE in early and late stages of dengue virus (DENV) infection. However, the downstream targets phosphorylated by BIKE have not yet been identified. Here, to identify BIKE substrates, we conducted a barcode fusion genetics-yeast two-hybrid screen and retrieved publicly available data generated via affinity-purification mass spectrometry. We subsequently validated 19 of 47 putative BIKE interactors using mammalian cell-based protein-protein interaction assays. We found that CLINT1, a cargo-specific adapter implicated in bidirectional Golgi-to-endosome trafficking, emerged as a predominant hit in both screens. Our experiments indicated that BIKE catalyzes phosphorylation of a threonine 294 CLINT1 residue both in vitro and in cell culture. Our findings revealed that CLINT1 phosphorylation mediates its binding to the DENV nonstructural 3 protein and subsequently promotes DENV assembly and egress. Additionally, using live-cell imaging we revealed that CLINT1 cotraffics with DENV particles and is involved in mediating BIKE's role in DENV infection. Finally, our data suggest that additional cellular BIKE interactors implicated in the host immune and stress responses and the ubiquitin proteasome system might also be candidate phosphorylation substrates of BIKE. In conclusion, these findings reveal cellular substrates and pathways regulated by the understudied Numb-associated kinase enzyme BIKE, a mechanism for CLINT1 regulation, and control of DENV infection via BIKE signaling, with potential implications for cell biology, virology, and host-targeted antiviral design.

A Comprehensive, Flexible Collection of SARS-CoV-2 Coding Regions.

The world is facing a global pandemic of COVID-19 caused by the SARS-CoV-2 coronavirus. Here we describe a collection of codon-optimized coding sequences for SARS-CoV-2 cloned into Gateway-compatible entry vectors, which enable rapid transfer into a variety of expression and tagging vectors. The collection is freely available. We hope that widespread availability of this SARS-CoV-2 resource will enable many subsequent molecular studies to better understand the viral life cycle and how to block it.

Maximizing binary interactome mapping with a minimal number of assays.

Complementary assays are required to comprehensively map complex biological entities such as genomes, proteomes and interactome networks. However, how various assays can be optimally combined to approach completeness while maintaining high precision often remains unclear. Here, we propose a framework for binary protein-protein interaction (PPI) mapping based on optimally combining assays and/or assay versions to maximize detection of true positive interactions, while avoiding detection of random protein pairs. We have engineered a novel NanoLuc two-hybrid (N2H) system that integrates 12 different versions, differing by protein expression systems and tagging configurations. The resulting union of N2H versions recovers as many PPIs as 10 distinct assays combined. Thus, to further improve PPI mapping, developing alternative versions of existing assays might be as productive as designing completely new assays. Our findings should be applicable to systematic mapping of other biological landscapes.

The SRC-family tyrosine kinase HCK shapes the landscape of SKAP2 interactome.

The SRC Kinase Adaptor Phosphoprotein 2 (SKAP2) is a broadly expressed adaptor associated with the control of actin-polymerization, cell migration, and oncogenesis. After activation of different receptors at the cell surface, this dimeric protein serves as a platform for assembling other adaptors such as FYB and some SRC family kinase members, although these mechanisms are still poorly understood. The goal of this study is to map the SKAP2 interactome and characterize which domains or binding motifs are involved in these interactions. This is a prerequisite to finely analyze how these pathways are integrated in the cell machinery and to study their role in cancer and other human diseases when this network of interactions is perturbed. In this work, the domain and the binding motif of fourteen proteins interacting with SKAP2 were precisely defined and a new interactor, FAM102A was discovered. Herein, a fine-tuning between the binding of SRC kinases and their activation was identified. This last process, which depends on SKAP2 dimerization, indirectly affects the binding of FYB protein. Analysis of conformational changes associated with activation/inhibition of SRC family members, presently limited to their effect on kinase activity, is extended to their interactive network, which paves the way for therapeutic development.

Comparative Profiling of Ubiquitin Proteasome System Interplay with Influenza A Virus PB2 Polymerase Protein Recapitulating Virus Evolution in Humans.

The optimized exploitation of cell resources is one cornerstone of a successful infection. Differential mapping of host-pathogen protein-protein interactions (PPIs) on the basis of comparative interactomics of multiple strains is an effective strategy to highlight correlations between host proteome hijacking and biological or pathogenic traits. Here, we developed an interactomic pipeline to deliver high-confidence comparative maps of PPIs between a given pathogen and the human ubiquitin proteasome system (UPS). This subarray of the human proteome represents a range of essential cellular functions and promiscuous targets for many viruses. The screening pipeline was applied to the influenza A virus (IAV) PB2 polymerase proteins of five strains representing different levels of virulence in humans. An extensive PB2-UPS interplay has been detected that recapitulates the evolution of IAVs in humans. Functional validation with several IAV strains, including the seasonal H1N1pdm09 and H3N2 viruses, confirmed the biological relevance of most identified UPS factors and revealed strain-independent and strain-specific effects of UPS factor invalidation on IAV infection. This strategy is applicable to proteins from any other virus or pathogen, providing a valuable resource with which to explore the UPS-pathogen interplay and its relationship with pathogenicity. IMPORTANCE Influenza A viruses (IAVs) are responsible for mild-to-severe seasonal respiratory illness of public health concern worldwide, and the risk of avian strain outbreaks in humans is a constant threat. Elucidating the requisites of IAV adaptation to humans is thus of prime importance. In this study, we explored how PB2 replication proteins of IAV strains with different levels of virulence in humans hijack a major protein modification pathway of the human host cell, the ubiquitin proteasome system (UPS). We found that the PB2 protein engages in an extended interplay with the UPS that evolved along with the virus's adaptation to humans. This suggests that UPS hijacking underlies the efficient infection of humans and can be used as an indicator for evaluation of the potential of avian IAVs to infect humans. Several UPS factors were found to be necessary for infection with circulating IAV strains, pointing to potential targets for therapeutic approaches.

Inhibition of the inflammatory response to stress by targeting interaction between PKR and its cellular activator PACT.

PKR is a cellular kinase involved in the regulation of the integrative stress response (ISR) and pro-inflammatory pathways. Two N-terminal dsRNA Binding Domains (DRBD) are required for activation of PKR, by interaction with either dsRNA or PACT, another cellular DRBD-containing protein. A role for PKR and PACT in inflammatory processes linked to neurodegenerative diseases has been proposed and raised interest for pharmacological PKR inhibitors. However, the role of PKR in inflammation is subject to controversy. We identified the flavonoid luteolin as an inhibitor of the PKR/PACT interaction at the level of their DRBDs using high-throughput screening of chemical libraries by homogeneous time-resolved fluorescence. This was further validated using NanoLuc-Based Protein Complementation Assay. Luteolin inhibits PKR phosphorylation, the ISR and the induction of pro-inflammatory cytokines in human THP1 macrophages submitted to oxidative stress and toll-like receptor (TLR) agonist. Similarly, luteolin inhibits induction of pro-inflammatory cytokines in murine microglial macrophages. In contrast, luteolin increased activation of the inflammasome, in a PKR-independent manner. Collectively, these data delineate the importance of PKR in the inflammation process to the ISR and induction of pro-inflammatory cytokines. Pharmacological inhibitors of PKR should be used in combination with drugs targeting directly the inflammasome.

Mapping the interactome of HPV E6 and E7 oncoproteins with the ubiquitin-proteasome system.

Protein ubiquitination and its reverse reaction, deubiquitination, regulate protein stability, protein binding activity, and their subcellular localization. These reactions are catalyzed by the enzymes E1, E2, and E3 ubiquitin (Ub) ligases and deubiquitinases (DUBs). The Ub-proteasome system (UPS) is targeted by viruses for the sake of their replication and to escape host immune response. To identify novel partners of human papillomavirus 16 (HPV16) E6 and E7 proteins, we assembled and screened a library of 590 cDNAs related to the UPS by using the Gaussia princeps luciferase protein complementation assay. HPV16 E6 was found to bind to the homology to E6AP C terminus-type Ub ligase (E6AP), three really interesting new gene (RING)-type Ub ligases (MGRN1, LNX3, LNX4), and the DUB Ub-specific protease 15 (USP15). Except for E6AP, the binding of UPS factors did not require the LxxLL-binding pocket of HPV16 E6. LNX3 bound preferentially to all high-risk mucosal HPV E6 tested, whereas LNX4 bound specifically to HPV16 E6. HPV16 E7 was found to bind to several broad-complex tramtrack and bric-a-brac domain-containing proteins (such as TNFAIP1/KCTD13) that are potential substrate adaptors of Cullin 3-RING Ub ligases, to RING-type Ub ligases implicated in innate immunity (RNF135, TRIM32, TRAF2, TRAF5), to the substrate adaptor DCAF15 of Cullin 4-RING Ub ligase and to some DUBs (USP29, USP33). The binding to UPS factors did not require the LxCxE motif but rather the C-terminal region of HPV16 E7 protein. The identified UPS factors interacted with most of E7 proteins across different HPV types. This study establishes a strategy for the rapid identification of interactions between host or pathogen proteins and the human ubiquitination system.

Quantifying domain-ligand affinities and specificities by high-throughput holdup assay.

Many protein interactions are mediated by small linear motifs interacting specifically with defined families of globular domains. Quantifying the specificity of a motif requires measuring and comparing its binding affinities to all its putative target domains. To this end, we developed the high-throughput holdup assay, a chromatographic approach that can measure up to 1,000 domain-motif equilibrium binding affinities per day. After benchmarking the approach on 210 PDZ-peptide pairs with known affinities, we determined the affinities of two viral PDZ-binding motifs derived from human papillomavirus E6 oncoproteins for 209 PDZ domains covering 79% of the human 'PDZome'. We obtained sharply sequence-dependent binding profiles that quantitatively describe the PDZome recognition specificity of each motif. This approach, applicable to many categories of domain-ligand interactions, has wide potential for quantifying the specificities of interactomes.

A comparative approach to characterize the landscape of host-pathogen protein-protein interactions.

Significant efforts were gathered to generate large-scale comprehensive protein-protein interaction network maps. This is instrumental to understand the pathogen-host relationships and was essentially performed by genetic screenings in yeast two-hybrid systems. The recent improvement of protein-protein interaction detection by a Gaussia luciferase-based fragment complementation assay now offers the opportunity to develop integrative comparative interactomic approaches necessary to rigorously compare interaction profiles of proteins from different pathogen strain variants against a common set of cellular factors. This paper specifically focuses on the utility of combining two orthogonal methods to generate protein-protein interaction datasets: yeast two-hybrid (Y2H) and a new assay, high-throughput Gaussia princeps protein complementation assay (HT-GPCA) performed in mammalian cells. A large-scale identification of cellular partners of a pathogen protein is performed by mating-based yeast two-hybrid screenings of cDNA libraries using multiple pathogen strain variants. A subset of interacting partners selected on a high-confidence statistical scoring is further validated in mammalian cells for pair-wise interactions with the whole set of pathogen variants proteins using HT-GPCA. This combination of two complementary methods improves the robustness of the interaction dataset, and allows the performance of a stringent comparative interaction analysis. Such comparative interactomics constitute a reliable and powerful strategy to decipher any pathogen-host interplays.

Comparative analysis of virus-host interactomes with a mammalian high-throughput protein complementation assay based on Gaussia princeps luciferase.

Comparative interactomics is a strategy for inferring potential interactions among orthologous proteins or "interologs". Herein we focus, in contrast to standard homology-based inference, on the divergence of protein interaction profiles among closely related organisms, showing that the approach can correlate specific traits to phenotypic differences. As a model, this new comparative interactomic approach was applied at a large scale to human papillomaviruses (HPVs) proteins. The oncogenic potential of HPVs is mainly determined by the E6 and E7 early proteins. We have mapped and overlapped the virus-host protein interaction networks of E6 and E7 proteins from 11 distinct HPV genotypes, selected for their different tropisms and pathologies. We generated robust and comprehensive datasets by combining two orthogonal protein interaction assays: yeast two-hybrid (Y2H), and our recently described "high-throughput Gaussia princeps protein complementation assay" (HT-GPCA). HT-GPCA detects protein interaction by measuring the interaction-mediated reconstitution of activity of a split G. princeps luciferase. Hierarchical clustering of interaction profiles recapitulated HPV phylogeny and was used to correlate specific virus-host interaction profiles with pathological traits, reflecting the distinct carcinogenic potentials of different HPVs. This comparative interactomics constitutes a reliable and powerful strategy to decipher molecular relationships in virtually any combination of microorganism-host interactions.

Mapping of Chikungunya virus interactions with host proteins identified nsP2 as a highly connected viral component.

Chikungunya virus (CHIKV) is a mosquito-transmitted alphavirus that has been responsible for an epidemic outbreak of unprecedented magnitude in recent years. Since then, significant efforts have been made to better understand the biology of this virus, but we still have poor knowledge of CHIKV interactions with host cell components at the molecular level. Here we describe the extensive use of high-throughput yeast two-hybrid (HT-Y2H) assays to characterize interactions between CHIKV and human proteins. A total of 22 high-confidence interactions, which essentially involved the viral nonstructural protein nsP2, were identified and further validated in protein complementation assay (PCA). These results were integrated to a larger network obtained by extensive mining of the literature for reports on alphavirus-host interactions. To investigate the role of cellular proteins interacting with nsP2, gene silencing experiments were performed in cells infected by a recombinant CHIKV expressing Renilla luciferase as a reporter. Collected data showed that heterogeneous nuclear ribonucleoprotein K (hnRNP-K) and ubiquilin 4 (UBQLN4) participate in CHIKV replication in vitro. In addition, we showed that CHIKV nsP2 induces a cellular shutoff, as previously reported for other Old World alphaviruses, and determined that among binding partners identified by yeast two-hybrid methods, the tetratricopeptide repeat protein 7B (TTC7B) plays a significant role in this activity. Altogether, this report provides the first interaction map between CHIKV and human proteins and describes new host cell proteins involved in the replication cycle of this virus.

The EVER proteins as a natural barrier against papillomaviruses: a new insight into the pathogenesis of human papillomavirus infections.

Infections by human papillomaviruses (HPVs) are the most frequently occurring sexually transmitted diseases. The crucial role of genital oncogenic HPV in cervical carcinoma development is now well established. In contrast, the role of cutaneous HPV in skin cancer development remains a matter of debate. Cutaneous beta-HPV strains show an amazing ubiquity. The fact that a few oncogenic genotypes cause cancers in patients suffering from epidermodysplasia verruciformis is in sharp contrast to the unapparent course of infection in the general population. Our recent investigations revealed that a natural barrier exists in humans, which protects them against infection with these papillomaviruses. A central role in the function of this HPV-specific barrier is played by a complex of the zinc-transporting proteins EVER1, EVER2, and ZnT-1, which maintain cellular zinc homeostasis. Apparently, the deregulation of the cellular zinc balance emerges as an important step in the life cycles not only of cutaneous but also of genital HPVs, although the latter viruses have developed a mechanism by which they can break the barrier and impose a zinc imbalance. Herein, we present a previously unpublished list of the cellular partners of EVER proteins, which points to future directions concerning investigations of the mechanisms of action of the EVER/ZnT-1 complex. We also present a general overview of the pathogenesis of HPV infections, taking into account the latest discoveries regarding the role of cellular zinc homeostasis in the HPV life cycle. We propose a potential model for the mechanism of function of the anti-HPV barrier.

Regulation of cellular zinc balance as a potential mechanism of EVER-mediated protection against pathogenesis by cutaneous oncogenic human papillomaviruses.

Epidermodysplasia verruciformis (EV) is a genodermatosis associated with skin cancers that results from a selective susceptibility to related human papillomaviruses (EV HPV). Invalidating mutations in either of two genes (EVER1 and EVER2) with unknown functions cause most EV cases. We report that EVER1 and EVER2 proteins form a complex and interact with the zinc transporter 1 (ZnT-1), as shown by yeast two-hybrid screening, GST pull-down, and immunoprecipitation experiments. In keratinocytes, EVER and ZnT-1 proteins do not influence intracellular zinc concentration, but do affect intracellular zinc distribution. EVER2 was found to inhibit free zinc influx to nucleoli. Keratinocytes with a mutated EVER2 grew faster than wild-type keratinocytes. In transiently and stably transfected HaCaT cells, EVER and ZnT-1 down-regulated transcription factors stimulated by zinc (MTF-1) or cytokines (c-Jun and Elk), as detected with luciferase assays. To get some insight into the control of EV HPV infection, we searched for interaction between EVER and ZnT-1 and oncoproteins of cutaneous (HPV5) and genital (HPV16) genotypes. HPV16 E5 protein binds to EVER and ZnT-1 and blocks their negative regulation. The lack of a functional E5 protein encoded by EV HPV genome may account for host restriction of these viruses.

Genital warts in Burmeister's porpoises: characterization of Phocoena spinipinnis papillomavirus type 1 (PsPV-1) and evidence for a second, distantly related PsPV.

We identified sequences from two distantly related papillomaviruses in genital warts from two Burmeister's porpoises, including a PV antigen-positive specimen, and characterized Phocoena spinipinnis papillomavirus type 1 (PsPV-1). The PsPV-1 genome comprises 7879 nt and presents unusual features. It lacks an E7, an E8 and a bona fide E5 open reading frame (ORF) and has a large E6 ORF. PsPV-1 L1 ORF showed the highest percentage of nucleotide identity (54-55 %) with human papillomavirus type 5, bovine papillomavirus type 3 (BPV-3) and Tursiops truncatus papillomavirus type 2 (TtPV-2). This warrants the classification of PsPV-1 as the prototype of the genus Omikronpapillomavirus. PsPV-1 clustered with TtPV-2 in the E6 and E1E2 phylogenetic trees and with TtPV-2 and BPV-3 in the L2L1 tree. This supports the hypothesis that PV evolution may not be monophyletic across all genes.

Human papillomavirus: cause of epithelial lacrimal sac neoplasia?

PURPOSE: Epithelial tumours of the lacrimal sac are rare but important entities that may carry grave prognoses. In this study the prevalence and possible role of human papillomavirus (HPV) infection in epithelial tumours of the lacrimal sac were evaluated. METHODS: Five papillomas and six carcinomas of the lacrimal sac were investigated for the presence of HPV using the polymerase chain reaction (PCR) technique. Fifteen specimens of dacryocystitis were included in the PCR reactions as controls. Furthermore, DNA in situ hybridization (ISH) and RNA ISH were performed. RESULTS: Low-risk HPV types 6 or 11 were identified in all four lacrimal sac papillomas suitable for PCR analysis and in situ hybridization. Four of six lacrimal sac carcinomas harboured HPV. One carcinoma was positive for HPV 11 only, two carcinomas had concomitant infection with HPV 6 or 11 and high-risk HPV 16, and the remaining carcinoma was positive for HPV 16. All specimens of dacryocystitis were betaglobin-positive and HPV-negative. Using DNA ISH, two papillomas and a single carcinoma showed evidence for vegetative HPV 11 DNA replication, whereas no HPV 16 DNA replication was found in the five carcinomas tested. HPV 11 RNA was demonstrated in two papillomas. CONCLUSIONS: By analysing 11 epithelial lacrimal sac papillomas and carcinomas using PCR, DNA ISH and RNA ISH, we found HPV DNA in all investigated transitional epithelium tumours of the lacrimal sac. HPV RNA was present in two of eight epithelial lacrimal sac tumours positive for HPV DNA. As RNA degrades fast in paraffin-embedded tissue, only a small fraction of DNA-positive tumours can be expected to be RNA-positive. We therefore suggest that HPV infection is associated with the development of lacrimal sac papillomas and carcinomas.