The rapid spread of SARS-CoV-2, the virus responsible for COVID-19, created an urgent need for effective therapeutic strategies. The lack of approved drugs prompted investigations into repurposing existing medications. Lopinavir and ritonavir, HIV protease inhibitors, emerged as potential candidates, based on their use in treating related coronaviruses and ongoing clinical trials in China. However, the molecular mechanisms of their action against SARS-CoV-2 remained unclear. The recent publication of the SARS-CoV-2 Mpro crystal structure (PDB ID: 6LU7) opened up opportunities for structure-based drug discovery (SBDD) techniques to elucidate these mechanisms. This study aims to investigate the binding mechanisms of lopinavir, ritonavir, and nelfinavir (included due to its promising in-vitro activity against SARS-CoV) to SARS-CoV-2 Mpro using supervised molecular dynamics (SuMD), an advanced computational approach that allows for atomic-level detail of the ligand recognition process. Understanding these mechanisms is crucial for optimizing existing therapies and guiding the development of novel antivirals against COVID-19.

Previous research highlighted the structural similarities between SARS-CoV-2 Mpro and the proteases of other coronaviruses, suggesting the potential for repurposing drugs effective against related viruses. Several studies explored the use of HIV protease inhibitors against SARS-CoV and MERS-CoV, with some demonstrating promising antiviral activity. The high sequence similarity between SARS-CoV and SARS-CoV-2 Mpro (96.1% sequence identity) further supported the rationale for investigating HIV protease inhibitors as potential COVID-19 treatments. While molecular docking studies have been employed to predict the binding of potential inhibitors to SARS-CoV-2 Mpro, this study utilized SuMD to provide a more dynamic and comprehensive understanding of the molecular recognition process, considering protein flexibility and solvent effects.

The study employed supervised molecular dynamics (SuMD) simulations to analyze the binding mechanisms of lopinavir, ritonavir, and nelfinavir to the SARS-CoV-2 main protease (Mpro). The three-dimensional structure of Mpro (PDB ID: 6LU7) was obtained from the RCSB Protein Data Bank. The ligands were prepared using MOE, and simulations were performed using AMBER14 force field. Each system consisted of Mpro, one of the three inhibitors, and a water box. The SuMD protocol involves a tab-like algorithm that monitors the distance between the ligand and the Mpro binding site, accepting only simulation steps that show ligand approach towards the binding site. This approach accelerates the sampling of the binding trajectory, reducing computational time. The simulations captured the full recognition process, including the initial contacts, conformational changes during binding, and the final stable bound state. Analysis of the simulations included calculations of interaction energies, identification of key interacting residues, and generation of interaction energy landscapes. The software used included MOE (Molecular Operating Environment), NAMD, and custom Python scripts with ProDy for trajectory analysis. Videos visualizing the SuMD trajectories were also generated using VMD.

SuMD simulations provided detailed insights into the binding pathways of each inhibitor. For Lopinavir, approximately 20 ns of simulation time was sufficient to sample a complete recognition trajectory. Key interactions were identified, including a double hydrogen bond with Glu166, a residue known to be crucial in SARS-CoV complexes. Ritonavir also exhibited binding within 20 ns but displayed lower energy stability compared to Lopinavir, with its urea moiety less effectively accommodated within the binding site. Nelfinavir required a slightly longer simulation (30 ns) to fully explore its binding pathway, revealing a metastable binding site before reaching the final, stable interaction. This final binding mode showed strong similarities to that of a covalently bound peptidomimetic compound observed in the crystal structure, involving interactions with residues His166, Glu166, Gln189, Thr190, and Gln216. In the last ns of the Nelfinavir simulation, a stabilizing salt bridge interaction was observed between Glu166 and the inhibitor. Interaction energy landscapes and dynamic interaction energy plots provided a quantitative description of the ligand-protein interactions throughout the binding process.

The SuMD simulations provided a detailed molecular-level understanding of the binding mechanisms of three HIV protease inhibitors to the SARS-CoV-2 main protease. The findings support the potential of lopinavir and nelfinavir as effective Mpro inhibitors, given the strong and stable interactions observed in the simulations. Ritonavir, though less stable, might still contribute synergistically with lopinavir through its pharmacokinetic enhancing effect. The identification of key interacting residues validates previous experimental findings and provides further insights into the structure-activity relationships of these inhibitors. The study highlights the power of SuMD in elucidating complex binding pathways and its potential application in drug discovery for other viral targets.

This study used SuMD simulations to provide novel insights into the binding mechanisms of lopinavir, ritonavir, and nelfinavir to SARS-CoV-2 Mpro. The results support the potential of these HIV protease inhibitors as effective COVID-19 therapeutics and provide valuable information for future drug design and optimization efforts. Future research could focus on exploring other potential inhibitors targeting Mpro and validating the computational findings experimentally. Further exploration of the synergistic effects of lopinavir and ritonavir, and the role of the metastable binding state of nelfinavir, would also be worthwhile.

The study is limited to computational simulations and does not include experimental validation of the predicted binding modes. The accuracy of the results depends on the quality of the force field and the limitations inherent in molecular dynamics simulations. The study focused only on the interaction with Mpro, neglecting other potential interactions with host factors or off-target effects.