Structural plasticity of SARS-CoV-2 3CL Mpro active site cavity revealed by room temperature X-ray crystallography

The study addresses how the SARS‑CoV‑2 main protease (3CL Mpro), an essential enzyme for viral polyprotein processing and a prime antiviral drug target, is structured at near-physiological temperature in its ligand-free state and how this relates to inhibitor binding. Given COVID‑19’s global impact and the lack of homologous human proteases, accurate structural information is critical for rational inhibitor design. The authors aim to determine a room-temperature (293 K) X-ray structure of unliganded 3CL Mpro to reveal the active-site conformation, compare it to low-temperature and inhibitor-bound structures, and assess implications for molecular docking and drug design.

Background literature notes that SARS‑CoV‑2 shares high sequence identity with SARS‑CoV and uses 3CL Mpro to cleave polyproteins at multiple conserved sites. Despite extensive efforts since SARS, no FDA-approved inhibitors exist for SARS-CoV 3CL protease. Prior crystallographic work established overall domain architecture and the catalytic Cys145–His41 dyad (rather than a canonical Ser/Cys-His-Asp/Glu triad). Numerous 100 K X-ray structures of ligand-free 3CL Mpro have been used for docking studies, and inhibitor-bound structures (e.g., with peptidomimetic N3) have informed binding subsites P1′–P5. However, potential temperature-dependent differences and active-site plasticity raise questions about the physiological relevance of low-temperature structures for docking.

- Protein engineering and cloning: SARS-CoV-2 3CL Mpro gene optimized for E. coli expression was synthesized and cloned into pET15b (MPro-PET15b), then inserted into PMCSG81-Delta238 to generate a MBP–self-cleavable fusion with an N-terminal viral autoprocessing site (SAVLQ↓SGFRK) and a C-terminal HRV 3C PreScission protease site linked to a His tag. Autoprocessing during expression produced an authentic N-terminus, and PreScission protease treatment generated an authentic C-terminus. - Expression and purification: Expressed in E. coli BL21(DE3) in LB with carbenicillin; induced at OD600 ~0.8 with 0.2 mM IPTG at 18 °C for 18 h. Lysis in 20 mM Tris pH 8, 40 mM imidazole, 150 mM NaCl, 1 mM TCEP; clarified and purified via HisTrap FF with gradient to 500 mM imidazole. PreScission protease added (500:1 molar ratio) and dialyzed (20 mM Tris pH 8, 150 mM NaCl, 1 mM TCEP, 4 °C, 18 h) to remove His tag. Reapplied to HisTrap to remove protease, tag, and uncleaved protein; authentic 3CL Mpro collected and concentrated to 4 mg/mL. - Crystallization: Initial screening (1536 conditions) via sitting-drop; best crystals in 0.1 M BIS-TRIS pH 6.5, 25% PEG3350. Microseeding used to grow single plate-like crystals in 10 µL sitting drops mixing protein (4 mg/mL) and reservoir (0.1 M BIS-TRIS pH 6.5, 20% PEG3350) 1:1 plus 0.2 µL seeds (1:100). Measured drop pH ~7.0. Crystals mounted in room-temperature capillaries (MiTeGen) for data collection. - X-ray data collection and structure determination: Room-temperature (293 K) X-ray data collected on a Rigaku High Flux HomeLab (MicroMax-007 HF, Osmic VariMax optics) with an Eiger 4M detector. Data integrated with CrysAlis Pro; scaled with Aimless (CCP4). Molecular replacement with Molrep using PDB 6M03. Refinement with Phenix.refine; model building in COOT; validation with MolProbity. Final resolution 2.30 Å. Coordinates and structure factors deposited as PDB 6WQF. - Comparative structural analysis: Least-squares superpositions (COOT) against a 100 K ligand-free structure (PDB 6Y2E) and an inhibitor-bound complex with N3 (PDB 6LU7) to assess temperature-dependent and ligand-induced conformational differences. - Molecular dynamics (MD): 1 µs classical MD of apo 3CL Mpro dimer (adapted from PDB 6Y84) using GROMACS 2020 with CHARMM36m force field. Solvated in a rhombic dodecahedron box (10 Å to edge) with TIP3P water and 8 Na+; PBC applied. Minimization (<1000 steps, steepest descent), equilibration to 300 K (V-rescale thermostat) and 1 bar (Berendsen barostat). Production with leap-frog integrator, Nose–Hoover thermostat and Parrinello–Rahman barostat. SHAKE constraints on bonds to H; coordinates saved every 10 ps. RMSF computed after RMSD convergence; radius of gyration stable at 26.2 ± 0.15 Å over 1 µs.

- Room-temperature (293 K) ligand-free structure at 2.30 Å reveals the catalytic Cys145 Sγ is 3.8 Å from His41 Nε2, consistent with both residues being neutral at pH ~7.0 and lacking a strong Cys–His hydrogen bond under these conditions. - A conserved water molecule (H2Ocat) forms strong hydrogen bonds with His41 and with Asp187 and His164 (distances ~2.9–3.0 Å), creating an interaction network stabilized by a salt bridge between Asp187 and Arg40. Authors propose H2Ocat may functionally complete a non-canonical catalytic triad by stabilizing positive charge on His41 via interaction with Asp187 during catalysis. Notably, some 100 K ligand-free structures (e.g., PDB 6M03) lack this water. - Comparison to 100 K ligand-free structure (PDB 6Y2E) shows overall similarity (Cα RMSD ~0.32 Å) but significant differences in residues 192–198 (P5 pocket). At room temperature, the Ala194 peptide bond flips inward toward the P5 pocket, resembling the N3-bound conformation (PDB 6LU7). Positions differ for Thr196 (backbone carbonyl O shifts by ~1.3 Å) and Asp197 (CG atoms shift ~1.9 Å; backbone carbonyl O shifts ~2.6 Å). - Ligand-induced conformational changes (comparison with N3 complex): The small helix (residues 46–50, near P2) and β-hairpin loop (residues 166–170, near P3–P4) move apart by ~2.4 Å; the P5 loop (residues 190–194) shifts closer toward the β-hairpin loop to accommodate the P5 substituent. Met49 and Met165 adopt alternate side-chain conformations to avoid steric clash with the inhibitor’s P2 leucine, propagating changes to Ser46 and Leu50. The C-terminal tail (Ser301–Gln306) swings ~180° from the ligand-free position to sit above domain III in the N3-bound form, disrupting several dimer-interface hydrogen bonds and potentially reducing dimer stability. - MD simulation of the apo dimer identifies the P2 helix (45–50), P5 loop (190–194), and C-terminal tail as the most dynamic regions (highest RMSF), supporting substantial active-site plasticity. The dimer remains compact and stable (Rg 26.2 ± 0.15 Å over 1 µs). - Implication: The room-temperature ligand-free structure captures physiologically relevant conformations and waters, providing a more suitable template for molecular docking and inhibitor design than low-temperature structures in some cases.

The room-temperature structure clarifies the protonation state and geometry of the catalytic dyad under near-physiological conditions and reveals a conserved catalytic water network that may act as an effective third component in catalysis. Temperature-dependent differences, especially in the P5 pocket (residues 192–198), and clear ligand-induced movements of secondary-structure elements highlight pronounced active-site plasticity and induced-fit behavior upon inhibitor binding. MD simulations corroborate the intrinsic flexibility of the P2–P5 subsites and the C-terminus. Collectively, these findings suggest that room-temperature ligand-free structures, capturing relevant conformational ensembles and hydration, may improve the accuracy of molecular docking and structure-based drug design against 3CL Mpro.

This work reports a 2.30 Å room-temperature X-ray structure of unliganded SARS‑CoV‑2 3CL Mpro (PDB 6WQF) and demonstrates significant active-site plasticity and a conserved catalytic water network likely important for catalysis. Comparative analyses with low-temperature and inhibitor-bound structures, together with MD simulations, indicate that the room-temperature structure better reflects physiologically relevant conformations for docking and inhibitor optimization. The structure provides a robust template for structure-assisted drug design targeting 3CL Mpro.