key: cord-1021421-dnppshnv authors: Hognon, Cécilia; Miclot, Tom; Iriepa, Cristina Garcia; Francés-Monerris, Antonio; Grandemange, Stephanie; Terenzi, Alessio; Marazzi, Marco; Barone, Giampaolo; Monari, Antonio title: Role of RNA Guanine Quadruplexes in Favoring the Dimerization of SARS Unique Domain in Coronaviruses date: 2020-05-27 journal: bioRxiv DOI: 10.1101/2020.04.07.029447 sha: 4bd359383066055b1658c45d962b494ee983b337 doc_id: 1021421 cord_uid: dnppshnv Coronaviruses may produce severe acute respiratory syndrome (SARS). As a matter of fact, a new SARS-type virus, SARS-CoV-2, is responsible of a global pandemic in 2020 with unprecedented sanitary and economic consequences for most countries. In the present contribution we study, by all-atom equilibrium and enhanced sampling molecular dynamics simulations, the interaction between the SARS Unique Domain and RNA guanine quadruplexes, a process involved in eluding the defensive response of the host thus favoring viral infection of human cells. Our results evidence two stable binding modes involving an interaction site spanning either the protein dimer interface or only one monomer. The free energy profile unequivocally points to the dimer mode as the thermodynamically favored one. The effect of these binding modes in stabilizing the protein dimer was also assessed, being related to its biological role in assisting SARS viruses to bypass the host protective response. This work also constitutes a first step of the possible rational design of efficient therapeutic agents aiming at perturbing the interaction between SARS Unique Domain and guanine quadruplexes, hence enhancing the host defenses against the virus. TOC GRAPHICS 4 different domains of Nsp3, 26 ,27 whose precise function of some of them has not been entirely clarified yet, the so-called SARS Unique Domain (SUD) deserves a special attention, since it is present only in SARS-type coronaviruses and hence it has been associated to the increased pathogenicity of this viral family. The structure of SUD (presumably a common domain of different SARS viruses) has been resolved experimentally, [28] [29] [30] and it has been proved that the macrodomain is indeed constituted by a dimer of two symmetric monomers. Furthermore, both experimental and molecular docking investigations have pointed out a possible favorable interaction of SUD with nucleic acids, and in particular with RNA in G-quadruplex (G4) conformation. 28 The presence of a high density of lysine residues at the interface between two SUD monomers, forming a positively charged pocket, also suggests that RNA may be instrumental in favoring SUD dimerization, due to the negative charge of the RNA backbone hence suggesting the occurrence of electrostatic attraction. This observation may have a highly important biological implication since the dimerization has also been connected to the SUD native function. Tan et al 28 have proposed that the ability of SUD to recognize and bind specific viral and/or host G4 sequences may have implications in regulating viral replication and/or hampering the host response to viral infection, as schematized in Figure 1 . The hypothesis is based on the identification of G4 sequences in key host mRNA that encode proteins involved in different signaling pathways such as apoptoting or survival signaling. [31] [32] [33] [34] [35] [36] [37] [38] These proteins could induce a controlled cellular death of infected cells slowing down or stopping the infection, or promote cell survival to delay apoptosis by producing antiviral cytokines. 37 However, the removal of the mRNA necessary to produce these signaling factors by viral SUD may impair the apoptosis/survival response pathways allowing massive cell infection. 28, 37 In this letter, we report an extended all-atom molecular dynamics (MD) study of the interactions produced between a dimeric SUD domain and a short RNA G4 sequence. The crystal structure of the protein (pdb 2W2G) and of the oligonucleotide (pdb 1J8G) 39 have been chosen coherently with the experimental work performed by Tan et al. 28 Even though the chosen SUD starting structure belongs to the 2009 SARS-CoV, the very high nucleotide affinity 40 and the global conservation of the Nsp protein suggest that the RNA binding spots should be globally preserved. This is also further justified by the fact that SARS-CoV-2 Nsp has also been recognized to suppress host gene expression and hence inhibit the immune response. 41 Equilibrium MD has allowed to assess some of the hypothesized complexation modes between G4 and SUD, while also highlighting the most important interactions patterns at an atomistic level, and the effects of G4 in maintaining the dimer stability. Furthermore, to better sample the multidimensional conformational space and to quantify the strength of the interactions coming into play, the free-energy surface has been explored using enhanced sampling methods. A two-dimensional (2D) free energy profile has been computed along two coordinates defining the distance between the centers of mass of G4 and one SUD domain (G4-SUD A ), and the two SUD domains (SUD A -SUD B ), respectively. The corresponding 2D potential of mean force (PMF) was obtained using a recently developed combination of extended adaptative biased force (eABF) 42 and metadynamics, 43 hereafter named meta-eABF. 44, 45 Both protein and RNA have been described with the amber force field 46 including the bsc1 corrections, 47, 48 and the MD simulations have been performed in the constant pressure and temperature ensemble (NPT) at 300K and 1 atm. All MD simulations have been performed using the NAMD code 49 and analyzed via VMD, 50 the G4 structure has also been analyzed with the 3DNA suite. 51, 52 More details on the simulation protocol can be found in Supplementary Information (SI). To obtain starting conformations, the RNA was manually positioned in two different orientations close to the experimentally suggested SUD interaction area. 28 The equilibrium MD evolved yielding two distinct interaction modes, as reported in Figure 2 . In particular, we can easily distinguish between a first mode of binding in which the G4 mainly interacts with only one SUD monomer, called monomeric binding mode, and a second one in which the nucleic acid is firmly placed at the interface between the two protein monomeric subunits, referred as dimeric binding mode. Note that while for the monomeric mode we easily found a suitable starting point, two independent 200 ns MD trajectory were run to characterize the dimeric mode. The corresponding root mean square deviations (RMSD) with respect to the initial structure are also reported and globally show that both the RNA and the protein units are stable. As expected, a slightly larger value of the RMSD is observed for the protein, as a consequence of its larger flexibility compared to the rigid G4 structures (Figure 2 d, c) . Note also that the slight initial increase of the protein RMSD observed for the dimeric mode is due to the necessity of a slight structural rearrangement to accommodate the G4 in the interaction pocket. Both modes are globally stable all along the MD trajectory, and no spontaneous release of the G4 is observed. At phosphates. This finding is evidenced by the radial distribution function (RDF) between these positively charged lysine side chains and the negatively charged phosphate oxygen atoms of G4 (depicted in dark blue in Figure 3b ), which shows a very intense and sharp peak at around 2 Å ( Figure 3a ). Interestingly, a secondary peak in the RDF is also observed at 3.5 Å, probably defining a second layer of interaction patterns that should contribute to the overall stabilization of the binding. Conversely, the monomeric mode is driven by interactions mainly involving the terminal uracil moieties and the top guanine leaflet instead of the phosphate backbone of G4. As shown in interacting with the peripheral uracil nucleobases, in a mode strongly resembling the top-binding experienced by a number of G4 drugs. [53] [54] [55] This is nicely confirmed by the analysis of the time series of the distance between the a-carbon of these amino acids and the nearby guanine that readily drops at around 5 Å and stays remarkably stable all along the MD. Interestingly, the interaction is sufficiently strong to induce a partial deformation of the planarity of the G4 leaflets. Even though from considerations based on chemical intuition those interactions could be referred as mainly due to dispersion, the inherent parameterization of the amber force field does not allow to completely disentangle and decompose the polarization, dispersion and electrostatic contributions. The fact that the monomeric binding mode is driven by non-covalent interactions with one of the exposed G4 leaflets may also contribute explaining the fact that longer G4 sequences are preferentially recognized by the SUD interface region. Indeed, in this case, for obvious statistic reasons, the ratio between the interaction with the backbone or with the terminal leaflet clearly favors the former. On the other hand, this mode may also act efficiently in the process of recruitment of RNA, either viral or cellular, efficiently anchoring the oligomer that can subsequently be safely displaced through the interface area. Apart from the different positioning of the G4, other structural evidences can already be surmised from the visual inspection of the MD trajectory. In particular, the SUD dimers appear more compact and the interface region better conserved when the RNA G4 adopts the dimeric binding mode, as can also been appreciated in Figure 4 . These results clearly indicate that the dimeric mode leads to a greater stability of the G4-SUD complex. and Arg266, as reported in Figure 4 . Indeed, while in the case of the dimeric-like conformation a peaked distribution centered at relative close distances (6.7 Å) is observed, indicating a closed and quite rigid disposition, a much broader and bimodal distribution is found for the monomer-like conformation, presenting, most notably, a secondary maximum at about 11 Å, which confirms the partial destabilization of the SUD subdomains interface and the greater flexibility induced by this binding mode. To further examine the conformational space spanned by the G4/SUD complex, and in particular the role of the RNA in favoring the dimerization and the structure of the interface, we resorted to enhanced sampling MD simulations to obtain the 2D free energy profile along two relevant collective variables: first, the distance between G4 and SUD, and second, the separation between the two SUD subdomains. Our choice of collective variables does not allow to explore the binding between the two surfaces of the SUD domain, however from the results of Tan et al. 28 it is clear that the interaction with RNA takes place preferentially at the positively charged interface. On the other hand, our methodology is perfectly adapted to characterize the influence of the binding of RNA to the stabilization of the interface between the two SUD monomers since it allows the sliding of the G4 on the SUD surface. The PMF is reported in Figure 5 together with representative snapshots along the reaction coordinates. From the analysis of the PMF, one can evidence the presence of an evident minimum in the free energy profile corresponding to the situation in which the G4 is interacting through the dimer mode, in which the SUD dimer is compact (Figure 5b ). The free-energy stabilization, with respect to the situation in which the G4 is well separated from the protein, amounts to about 6 kcal/mol. Around the principal minimum a slightly less stable and extended region is also observed having a stabilization free energy of about 3-4 kcal/mol and corresponding to the sliding of G4 in the monomer conformation (Figure 5c ). The rest of the free energy surface appears instead rather flat, and no appreciable barriers are observed along the collective variable. The topology of the free energy surface hence accounts for the possibility to observe important conformational movements, leading to open conformations in which the SUD subdomain interface has been basically destroyed (Figure 5d) . However, such conformations are instead hampered by the dimer-like conformation of the RNA. The free energy map unambiguously shows that the dimer mode is the preferred one, and also confirms the role of the G4 binding in maintaining the dimeric SUD conformation, since no appreciable free energy barrier for the opening of the SUD dimer is observed when the RNA is unbound. Thus, the dimer mode binding site clearly constitutes a specific target that may help in the development of new efficient antiviral agents against coronaviruses. Crystal Structure of SARS-CoV-2 Main Protease Provides a Basis for Design of Improved α-Ketoamide Inhibitors Potential Treatments for COVID-19; a Predicting Commercially Available Antiviral Drugs That May Act on the Novel Coronavirus (2019-NCoV) Use of Antiviral Drugs to Reduce COVID-19 Transmission. The Lancet Global Health Molecular Dynamics Simulations Indicate the COVID-19 Mpro Is Not a Viable Target for Small-Molecule Inhibitors Design Changes in Nonstructural Protein 3 Are Associated with Attenuation in Avian Coronavirus Infectious Bronchitis Virus SARS Coronavirus Replicase Proteins in Pathogenesis The SARS-Unique Domain (SUD) of SARS Coronavirus Contains Two Macrodomains That Bind G-Quadruplexes Unique Domain: Three-Domain Molecular Architecture in Solution and RNA Binding Nuclear Magnetic Resonance Structure Shows That the Severe Acute Respiratory Syndrome Coronavirus-Unique Domain Contains a Macrodomain Fold An Overview of Their Replication and Pathogenesis BT -Coronaviruses: Methods and Protocols The Rich World of P53 DNA Binding Targets: The Role of DNA Structure Nsp3 of Coronaviruses: Structures and Functions of a Large Multi-Domain Protein Binding Macrodomain within the "SARS-Unique Domain" Is Essential for the Activity of the SARS-Coronavirus Replication-Transcription Complex BBC3 Mediates Fenretinide-Induced Cell Death in Neuroblastoma The Structure of Human Neuronal Rab6B in the Active and Inactive Form Signal Transduction in SARS-CoV-Infected Cells Identification of a Human NF-ΚB-Activating Protein Analysis of an RNA Tetraplex (UGGGGU)4 with Divalent Sr2+ Ions at Subatomic Resolution (0.61 Å) Network-Based Drug Repurposing for Novel Coronavirus 2019-NCoV/SARS-CoV-2 Structural Basis for Translational Shutdown and Immune Evasion by the Nsp1 Protein of SARS-CoV-2 The Extended Generalized Adaptive Biasing Force Algorithm for Multidimensional Free-Energy Calculations Wiley Interdisciplinary Reviews: Computational Molecular Science Zooming across the Free-Energy Landscape: Shaving Barriers, and Flooding Valleys Taming Rugged Free Energy Landscapes Using an Average Force Ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from Ff99SB Assessing the Current State of Amber Force Field Modifications for DNA Refinement of the Sugar-Phosphate Backbone Torsion Beta for AMBER Force Fields Improves the Description of Z-and B-DNA Scalable Molecular Dynamics with NAMD VMD: Visual Molecular Dynamics 3DNA: A Software Package for the Analysis, Rebuilding and Visualization of Three-Dimensional Nucleic Acid Structures 3DNA: A Versatile, Integrated Software System for the Analysis, Rebuilding and Visualization of Three-Dimensional Nucleic-Acid Structures Targeting G-Quadruplexes with Organic Dyes: Chelerythrine-DNA Binding Elucidated by Combining Molecular Modeling and Optical Spectroscopy Nickel(II), Copper(II) and Zinc(II) Metallo-Intercalators: Structural Details of the DNA-Binding by a Combined Experimental and Computational Investigation Selective G-Quadruplex Stabilizers: Schiff-Base Metal Complexes with Anticancer Activity