Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Oct 8;65(20):11314–11325. doi: 10.1021/acs.jcim.5c01708

De Novo-Designed Miniprotein Inhibits the Enzymatic Activity of the SARS-CoV‑2 Main Protease

Tayná E Lima †, Emerson G Moreira †,‡, Danilo F Coêlho ‡, Carlos H B Cruz §, Rafael Dhalia †, Bruno H S Leite †, Lícya S Xavier †, Marta Perez Illana †, Gabriel L Wallau ∥,⊥, Isabelle F T Viana †,‡,⊥,#,*, Roberto D Lins †,⊥,*
PMCID: PMC12570131  PMID: 41060276

Abstract

Targeting viral proteases is a well-established antiviral strategy and a promising approach that has been actively explored against SARS-CoV-2. The SARS-CoV-2 main protease (Mpro) is essential for viral replication and functions as a homodimer, making its dimerization interface an attractive therapeutic target. In this study, we report the rational design of HB3-Core25, a miniprotein computationally engineered to disrupt Mpro dimerization and inhibit its catalytic activity. In vitro production followed by biophysical characterization showed that HB3-Core25 folds into a compact trimeric helical bundle, exhibiting high solubility and thermal stability. Biophysical assays confirmed binding to Mpro with a dissociation constant (K D) of 0.567 μM and the lowest IC50 reported to date for the dimer interface. Functional assays further demonstrated inhibition of Mpro catalytic activity, with 51.1%. These findings highlight HB3-Core25 as a stable inhibitor of Mpro activity by interfering with its dimerization, offering a complementary strategy to classical active-site inhibition in antiviral drug development.

graphic file with name ci5c01708_0007.jpg

graphic file with name ci5c01708_0005.jpg

Introduction

Viruses often rely on proteases to cleave large polyproteins into functional units, enabling the production of mature viral proteins essential for replication and assembly of new infectious particles. This mechanism not only ensures efficient replication but also allows viruses to maintain compact genomes. , Owing to their indispensable role in the viral life cycle, viral proteases have emerged as well-established therapeutic targets, with both competitive and allosteric inhibitors successfully deployed against pathogens such as the human immunodeficiency virus (HIV) and hepatitis C virus (HCV). In the wake of the COVID-19 pandemic, similar antiviral strategies have been redirected toward combating Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), the causative agent of the disease. ,

SARS-CoV-2 encodes two polyproteins, pp1a and pp1ab, which are cleaved into 16 nonstructural proteins (NSPs) by two viral proteases: the papain-like protease (PLpro) and the main chymotrypsin-like protease (Mpro). PLpro cleaves the junctions between NSP1 and NSP4, whereas Mpro processes 11 additional cleavage sites, spanning NSP4–NSP16, including autocleavage at its own N- and C-terminal ends. Because of its crucial role in viral replication, Mpro has become a primary target in the development of antiviral drugs. , It is a cysteine protease that operates as a homodimer with catalytic activity driven by a dyad composed of histidine 41 and cysteine 145. , While significant efforts have been applied toward the discovery of effective antivirals for SARS-CoV-2, the vast majority of molecules that have completed clinical trials were originally developed for other infectious and inflammatory diseases and were repurposed for COVID-19. Various approaches have been explored, including chemically modified and structure-guided designed peptides targeting the catalytic site or the dimerization interface of Mpro, although drug development has primarily focused on competitive inhibitors. ,,−

Meanwhile, the N-terminal region of Mpro, commonly referred to as the N finger, is particularly important for maintaining its homodimeric structure, contributing approximately 39% of the interaction interface. , Prior studies have attempted to interfere with this interaction using peptides that mimic the N-terminal sequence, e.g., SGFRKMAF (N8 peptide) for SARS-CoV-1 and SGFRKMAFPS for SARS-CoV-2. These peptides successfully reduced enzymatic activity by up to 46.2%; however, they required high concentrations (IC50 ≥ 500 μM), likely due to their conformational flexibility. , Although these peptides demonstrate a proof-of-concept for disrupting Mpro dimerization, their limited potency and poor stability diminish their therapeutic potential.

To address these challenges, computational protein design offers a compelling alternative for targeting Mpro’s dimerization interface. − Small molecules and peptides have played central roles in antiviral drug development due to their oral bioavailability and ease of synthesis. However, they often struggle to target flat protein–protein interfaces, such as that of Mpro dimerization. Peptides can engage larger surfaces but are limited by low stability and rapid degradation. In contrast, synthetic proteins, particularly miniproteins, offer improved stability, resistance to proteolysis, and customizable pharmacokinetics. Miniproteins are compact, de novo-designed scaffolds (<50 residues) displaying properties enabling high-affinity and selective targeting of challenging surfaces, including the well-defined secondary structure, cooperative folding, and sequestered hydrophobic cores.

In this study, we employed de novo design to generate a miniprotein that selectively binds to this critical region of SARS-CoV-2 Mpro. Through structural and thermal stability characterization, along with in vitro functional assays, we demonstrate that this miniprotein exhibits a high binding affinity and effectively inhibits Mpro catalytic activity. Our findings lay the ground for a new class of antiviral agents that may complement existing therapies in the ongoing fight against COVID-19.

Experimental Section

Computational Procedures

The computational workflow comprised four key steps: (1) evaluation of N8 peptide-Mpro interactions; (2) de novo design of miniprotein inhibitors; (3) stability assessment of designed inhibitors; and (4) analysis of inhibitor stability and interface dynamics.

Structure Preparation

The monomeric structure of SARS-CoV-2 Mpro and its N8-derived peptide complex (Mpro-N8) were modeled based on the crystallographic structure of the Mpro homodimer (PDB ID: 7ALI, resolution: 1.65 Å). Crystallographic water molecules, ligands, and ions were removed prior to energy minimization, which was performed using the FastRelax protocol in the Rosetta suite (v3.13) with the scoring function described elsewhere.

Alanine Scanning

Binding hotspots within the N8 peptide were identified using a consensus approach across three alanine scanning methodologies: (1) FoldX v5; (2) MutaBind2 web server; and (3) Rosetta’s Flex ddG protocol. The contribution of each residue to binding energy (ΔΔGbind) was calculated using the equation ΔΔG bind = ΔG ala – ΔG wt, where ΔG wt and ΔG ala refer to the binding free energies of the wild-type and alanine-mutated peptides, respectively. Hotspot criteria were defined according to the criteria of each method used: (1) Foldx: +0.92 kcal/mol < ΔΔG bind ≤ +1.84 kcal/mol; , (2) MutaBind2: ΔΔG bind ≥ +1.5 kcal/mol; and (3) Flex ddG: ΔΔG bind ≥ +1.7 kcal/mol. A residue was classified as a binding hotspot if it met the threshold in at least two of the three methods.

De Novo Design of Miniprotein

The de novo design strategy, adapted from previous studies, , involved four stages: (1) motif binding design; (2) backbone design; (3) sequence design (monomer and interface); and (4) sequence-structure compatibility prediction. A 16-residue α-helix scaffold (H16) composed of valines (V) was generated using RosettaRemodel as a scaffold due to its structural compatibility with the binding groove of Mpro. Structural alignment of H16 Cα atoms with those of N8 hotspot residues guided motif positioning (Figure A). Based on the alanine scanning results from N8 in complex with Mpro (Figure S1), valine residues in H16 at the corresponding positions of R4 and M6 were mutated to R and M, respectively, using Pymol’s mutagenesis tool. Optimal side chain rotamers for these residues were manually selected prior to starting the sequence design protocol. These residues were not mutated during the design.

1.

1

Open in a new tab

Computational de novo design of the Mpro-binding miniprotein. (A) Design protocol of the Mpro-binding protein backbone. Mpro surface is shown in light purple, while the N8 peptide is depicted in pink licorice mode. The H16 residue and the transplanted R and M residues are shown in green. The backbone model of the miniprotein mimetic derived from the N8 peptide is also shown in green cartoon mode. (B) Assessment of the prediction quality. Results of the ab initio relaxation simulations are shown in the plot (left). The horizontal axis represents the Rosetta energy per residue, while the vertical axis represents the root-mean-square deviation (RMSD) of the structures relative to the design model. Gray dots correspond to all generated structures; the light green dot indicates the lowest-energy structure. On the right side, low-scoring models from the folding simulations with ab initio relax (pink) are superimposed on the models predicted by AF2 (cyan) and the design model (light green) in cartoon mode. (C) RMSD and (D) root-mean-square fluctuation (RMSF) obtained by classical molecular dynamics simulations, having the designed coordinates as the reference structure, in three temperatures: 25 °C (green), 55 °C (blue), and 95 °C (dark blue).

Backbone design was carried out using Rosetta’s BluePrintBDR protocol via sequence-independent fragment assembly simulations. A three-helix bundle topology with 2–3 residue loops was selected, based on previous studies. Backbone torsion angles were defined using the ABEGO notation. For each blueprint file, 15 000 models were generated using the fldcen.wts weight set. Designs were selected based on (1) ≥20% of residues forming a hydrophobic core and (2) absence of steric clashes with Mpro.

The sequence design was performed in two steps: full monomer sequence design followed by interface optimization (residues within 10 Å of Mpro were refined for enhanced binding affinity). The FastDesign protocol generated 50 000 sequences. Additional scripts (RelaxScripts, MonomerDesign2019, and InterfaceDesign2019) were used to ensure incorporation of bulky residues at the hydrophobic core and interface. Rosetta's LayerDesign TaskOperation was applied to restrict the sequence space. Designs were filtered based on monomer and interface criteria. Monomer metrics included: energy per residue ≤ −3 kcal/mol, shape complementarity (SS-Sc) ≥ 0.8, buried nonpolar area residue per (nspa) ≥ 50 Å2, ≤5 alanines overall, ≤3 alanines in the hydrophobic core, and ≥20% hydrophobic core residues. Interface metrics included: ddG ≤ −40 kcal/mol, buried solvent-accessible surface area (SASA) ≤ 1200 Å2, interface energy density (dG separated × 100/dSASA), buried unsatisfied hydrogen bonds (ΔInsatLH) ≤ 10, shape complementarity (Sc) ≥ 0.6, molecular contact area ≥ 450, and PackStat ≥ 0.65. Structure prediction was performed to validate sequence-structure compatibility using both ab initio relaxation (Rosetta v3.13) and AlphaFold2 (AF2) via ColabFold. In the Rosetta protocol, 3- and 9-residue fragments generated with Robetta were used to build the protein structure for simulations (25 000–30 000 total). The generated structures were compared to the original design model. Folding success was defined by the emergence of a funnel in the energy vs RMSD plot. All Rosetta Scripts from backbone generations and sequence design can be found in the Supporting Information.

Molecular Dynamics (MD) Simulations

MD simulations were performed with GROMACS 2019.4 using the CHARMM36m force field to assess the structural stability of the HB3-Core25 miniprotein, its complex with Mpro, and the Mpro dimer. Protonation states were assigned with H++ server at pH 7.4 and 0.15 M ionic strength. Systems were solvated with TIP3P water in a cubic box and neutralized with counterions (achieving a final ionic concentration of 150 mM). Energy minimization was performed using 5000 steps of the Steepest Descent algorithm (maximum atomic force ≤ 100 kJ mol–1 nm–2). Systems were heated (NVT ensemble, 1 ns) and equilibrated in five 1 ns NPT steps, reducing positional restraints from 5000 kJ·mol–1 to 1000 kJ mol–1 nm–2. Production runs were performed by using the NPT ensemble: 0.5 μs for the monomer and 1.0 μs for complexes. Periodic boundary conditions were applied in all directions (x, y, and z) during all simulation stages. The V-rescale maintained a temperature constant with a coupling constant of 0.4 ps. Pressure was set to 1 bar using the Parrinello–Rahman barostat with a relaxation time of 2.0 ps and a compressibility constant of 4.6 × 10–5 bar–1. Long-range electrostatics were handled using the PME method, while van der Waals interactions and short-range electrostatics were truncated at 1.2 nm cutoff radii. The LINCS algorithm was used to constrain bonds involving hydrogen atoms. , The leapfrog algorithm, with a 2.0 fs integration time step, was used for numerical integration equations of motion. Simulation box sizes were 5.50 nm3 for the Mpro monomer; 9.98 nm3 for the Mpro homodimer, and 9.80 nm3 for the Mpro-HB-Core25 complex.

Recombinant Protein Expression

Synthetic expression plasmids (GenScript) encoding HB3-Core25 (pET-29b­(+)) and the GST-tagged variant (pGEX-4T-1) were transformed into Escherichia coli BL21 Star (Invitrogen). SARS-CoV-2 Mpro sequence was cloned into the same GST vector. Cultures were grown in a Luria–Bertani (LB) medium supplemented with 50 mg/mL ampicillin or 10 mg/mL kanamycin. At OD600 nm = 0.5–0.8, protein expression was induced with 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) at 22 °C, 225 rpm, for 20 h. Cells were pelleted (7000 g, 4 °C, for 30 min) and resuspended in lysis buffer (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 20 mM imidazole) supplemented with protease inhibitor (Roche).

Affinity Purification

Bacterial cells were disrupted by sonication (Vibracell VCX 750 Sonicator) on ice, and lysates were purified via metal ion affinity chromatography using a HisTrap HP column (Cytiva). Proteins were eluted with an imidazole gradient (up to 500 mM) in elution buffer (50 mM Tris-HCl, pH 8.0, 500 mM NaCl2) under a stepwise fashion. For SARS-CoV-2 Mpro purification, 10 mM dithiothreitol (DTT) was added to all of the buffers. The GST-tagged version of the miniprotein (GST-HB-Core25) was purified using a GSTrap FF column (Cytiva) through isocratic elution in 50 mM Tris-HCl, pH 8.0, 10 mM reduced glutathione. SDS-PAGE assessed purity (BlueSafe stain (NZYtech)), and the protein concentration was determined by spectrophotometry (Nanodrop Onec, Thermo) using extinction coefficients at 280 nm (Mpro) or 227 nm (HB3-Core25).

Size-Exclusion Chromatography (SEC)

Proteins were dialyzed overnight in 1× phosphate-buffered saline (PBS) at 4 °C and concentrated using Amicon Ultra 15 filters (Millipore). HB3-Core25 (100 μM) was incubated with Mpro (30 μM) at 22 °C for 30 min to form complexes. The resulting complexes were loaded onto a HiLoad 26/600 Superdex 75 pg column (Cytiva). Peaks were assigned by comparison and data interpolation to a standard curve. Partition coefficients (K av) were calculated for protein complex characterization.

Circular Dichroism (CD) Spectroscopy

The secondary structure was analyzed on a J-1100 spectropolarimeter (Jasco) from 190 to 260 nm at 25 °C, with three accumulations and a scanning speed of 20 nm/min, using 1 mm quartz cuvettes. HB3-core25 (5 μM) and Mpro (10 μM) were diluted in 100 mM sodium phosphate buffer (pH 7.4). Thermal stability was assessed by monitoring ellipticity at 222 nm from 10 to 100 °C in 2 °C increments and 30 s equilibration steps, while reversibility of unfolding was assessed by cooling the same sample from 100 to 10 °C. Complex melting temperature (T m) was determined after incubation of equimolar protein conditions (5 μM, 22 °C, 30 min) during thermal titration (10–100 °C) and compared to the denaturation curve of Mpro alone at the same concentration (5 μM). All experiments were repeated three times.

Microscale Thermophoresis (MST)

The equilibrium dissociation constant (K D) of the Mpro-HB3-Core25 interaction was measured by MST using fluorescently labeled Mpro (RED-MALEIMIDE second Generation protein labeling kit, NanoTemper). Labeled Mpro (250 nM) was titrated with 16 serial dilutions of GST-HB3-Core25 (101–0.00308 μM). Following incubation (22 °C, 30 min, in the dark), the samples were loaded into standard MST capillaries and analyzed on a Monolith NT.115 instrument (NanoTemper). The K D values, along with noise and user-estimated errors, were calculated using the MO-affinity Analysis software.

SARS-CoV-2 Mpro Enzymatic Inhibition Assay

The inhibitory effect of HB3-Core25 on Mpro was assessed using a luminogenic substrate (Z-RLRGG-aminoluciferin, Promega), which mimics the N-terminal protease’s autocleavage sequence. The assay was conducted in 96-well plates with 40 μM substrate, 0.16 μg/μL Mpro, and serial dilutions of HB3-Core25 (1 nM–25 μM), incubated at 37 °C for 1 h. Luminescence was measured after adding 50 μL of detection reagent (Promega) to each well and a 20 min stabilization period using a Glomax luminometer (Promega). The results were compared to the positive control (Mpro in the absence of the inhibitor). Mean inhibitory concentrations (IC50) were calculated from dose–response curves using GraphPad Prism 6.

Results

De Novo Mpro Binding Protein Folds into the Desired Conformation and Exhibits High Stability in Solution

The N-terminal region of Mpro from both SARS-CoV-1 and SARS-Cov-2 has previously been shown to inhibit proteolytic activity by disrupting homodimer formation, which is essential for protease function. , To identify the key residues contributing to binding affinity in the Mpro-N8 complex, we first performed alanine scanning (AS) mutagenesis across the peptide sequence. This analysis revealed that substitutions R4A and M6A led to significant increases in the binding free energy (Figure S1), indicating that R4 mediates critical electrostatic interactions, particularly with E290 of Mpro, while M6 is essential for hydrophobic packing within the complex. These findings identified the R4 and M6 residues as key for rational inhibitor design.

Guided by this finding, we pursued a de novo design strategy to engineer a stable miniprotein inhibitor targeting Mpro. We began by grafting the key residues R4 and M6 onto a 16-residue α-helix, which served as the central motif of a three-helix bundle topology (Figure A). A total of 45 000 simulations were conducted to generate candidate scaffolds. Designs were filtered based on their ability to form well-packet hydrophobic cores, avoid steric clashes with Mpro, and incorporate loops as short as possible to enhance rigidity. These criteria ensured optimal side chain packing around the central helix, which housed the transplanted binding residues.

Following scaffold selection, we carried out extensive sequence design using Rosetta’s FastRelax protocol, generating 50 000 sequences optimized for structural compatibility and target binding. Three top candidates were initially selected based on the following metrics: total energy per residue (ranging from −3.53 to −3.41 kcal/mol), buried nonpolar surface area (nSPA, 60.72–68.93 Å2), and overall packing efficiency. However, only one exhibited the desired biological activity experimentally. For conciseness, we will only discuss this single candidate (Figure S2 and Table S1). Importantly, the designed interfaces showed improved predicted binding affinity (ΔΔG) compared to the original N8 peptide, along with strong shape complementarity (Sc > 0.66), excellent internal packing (Packstat >0.66), and a high contact molecular surface (CMS > 498). A comparison of the designed sequence with the N8 peptide reveals a significant sequence difference, where only 3 out 8 residues are identical. Besides the grafted R4 and M6, the miniprotein has an alanine at the A7 corresponding position in the N8 peptide. (The N8 peptide sequence is SGFRKMAF, while the corresponding region in the designed protein is HEARVMAM.)

To ensure that the designed sequence would indeed adopt the intended fold, we performed structural validation using both ab initio relax folding simulations and AlphaFold2-based predictions. Energy landscape analysis revealed a narrow and well-defined funnel, consistent with a strongly preferred global minimum and reduced conformational heterogeneity. The RMSD between the designed model and the lowest-energy folded structure was 0.0 Å (Figure B), suggesting a perfect structural agreement. Similarly, AlphaFold2 and AlphaFold3 predictions yielded high per-residue confidence scores, with per-residue measure of local confidence (pLDDT) values above 95%, reinforcing the reliability of the generated models. The predicted structure of the complex also exhibited remarkable agreement with the AlphaFold2 and AlphaFold3 models, with backbone RMSD values of ca. 0.6 Å further validating the reliability of the model. Altogether, these results demonstrate that the designed miniprotein adopts the desired three-helix bundle with high fidelity and stability, making it a promising scaffold for Mpro inhibition.

Mpro-HB3-Core25 Complex Forms a Stable Association in Solution

While Rosetta implicitly accounts for solvent effects during interface design, penalizing polar atoms with unsatisfied hydrogen bonds, explicit solvent dynamics is essential for accurately evaluating the stability of protein–protein interactions. An MD simulation of the Mpro-HB3-Core25 system was carried out to refine the key interactions of the structural complex. A simulation of the native Mpro homodimer was also performed for comparison purposes. Simulations were conducted at 25 °C, over a 1 μs trajectory. Both complexes remained structurally stable throughout the simulation time, exhibiting low RMSD fluctuations in the backbone atoms across all chains (Figures A,B and S3). The center-of-mass (COM) distances between binding partners also remained consistent, and fluctuations in the RMSD of interfacial heavy atoms showed minimal RMSD variations, indicating the structural integrity of the interfaces (Figure A,B).

2.

2

Open in a new tab

Structural dynamics and key interactions in the Mpro dimer and Mpro-HB3-Core25. (A) Distance between the centers of mass of the heavy atoms at the interface residues. (B) RMSD of the heavy chain atoms at the interface residues of Mpro (blue) and HB3-Core25 (green). (C) Protein–protein interface interaction represented by residue distances in Å. The Mpro interface is colored blue, and HB3-Core25 is shown in green. Dashed lines indicate possible interactions between the molecules: hydrogen bonds (orange), salt bridges (purple), and hydrophobic interactions (orange). Interface was defined as any atomic contacts that fell within a 6 Å distance. (D) RMSF of residue-averaged heavy chain atoms of Mpro in the presence (green) and absence of the miniprotein (blue). (The Mpro curve is shown as the resulting average of the two homodimer chains).

To understand the molecular basis of this stability, we characterized the key interfacial interactions maintained during the simulations. The Mpro-HB3-Core25 interface was primarily stabilized by hydrogen bonds and salt bridges, formed by residue pairs including R30A–E290B (100%), M26A–Y126B (99.7%), Q27A–K5B (97.5%), L36A–L286B (96%), E34A–K189B (93.4%), M24A–58B (90.6%), and E46A–R4B (89.5%) (Figure C). These interactions were present for the majority of the simulation time and closely mirrored those observed in the native Mpro homodimer (Table S2). Although hydrophobic contacts constitute a smaller portion of the total interactions, they remained stable over the course of the simulation, contributing additional support to the interface integrity. The overall dynamical fluctuation remained unchanged for Mpro upon miniprotein binding (Figure D). A very low RMSF is observed in both cases throughout the protein except for the C-terminal residues in both cases. The complex was further validated by calculating the binding affinity of the miniprotein to Mpro by PBEE, a recently developed ML-based software to estimate protein–protein binding free energies with unprecedent accuracy. PBEE predicts a ΔG bind of −10.72 kcal/mol, which corresponds to a K D of ca. 10-8 M. Collectively, these results suggest that the protein–protein interactions engineered by Rosetta form a robust and persistent interface under physiologically relevant conditions. These results suggest that HB3-Core25 exhibits sufficient structural and thermodynamic stability to engage Mpro in a dimer-blocking mode.

Miniproteins Are Thermostable and Maintain the Desired Conformation

To validate the in silico predictions and evaluate the biophysical properties of HB3-Core25, the miniprotein was recombinantly expressed in E. coli. Expression was successful and yielded 0.9 g/L (Figure S4A). The protein remained highly soluble throughout purification and handling, as evidenced by the absence of aggregation or viscosity changes. To assess whether the designed secondary structure was preserved after expression, we performed circular dichroism (CD) spectroscopy. The CD spectrum displayed hallmark features of α-helical structures, with characteristic minima at 208 and 222 nm, and a maximum at 193 nm (Figure A), confirming that HB3-Core25 adopts an α-helical conformation in solution. Moreover, the CD signal remained unchanged upon heating, indicating resistance to complete thermal denaturation. These findings are consistent with the MD simulations conducted at comparable temperatures (Figure D), further supporting the structural robustness of the miniprotein.

3.

3

Open in a new tab

Biophysical characterization of the HB3-Core25 miniprotein. (A) CD spectra of HB3-Core25 recorded over the wavelength range of 190–260 nm from 25 to 95 °C, at 10 °C intervals. (B) CD spectra of HB3-Core25 at a fixed wavelength of 222 nm during thermal titration from 10 to 100 °C (green line) and from 100 to 10 °C (black line). Shaded areas represent experimental error over three independent measurements.

To evaluate thermal stability more precisely, we conducted thermal denaturation and refolding capacity assays by monitoring the CD signal at 222 nm across a temperature range of 10–100 °C (upon sample heating and cooling). Remarkably, no significant structural changes were detected within this range, suggesting that the miniprotein remains fully folded, even at the highest tested temperature (Figure B). This result indicates a melting temperature (T m) exceeding 100 °C and highlights the exceptional thermostability of HB3-Core25. Noteworthy, the heating and cooling CD spectra showed an overlapping pattern that indicates not only the reversibility of unfolding but also that HB3-Core25 exhibits a robust folding pathway, as expected from a well-designed miniprotein.

HB3-Core25 Miniprotein Forms a Complex with SARS-CoV-2 Mpro and Inhibits Its Catalytic Activity

Although the HB3-Core25 miniprotein was shown to be structurally robust, its inhibitory potential against Mpro depends on its ability to form a functional complex with the protease. CD spectroscopy and microscale thermophoresis experiments were performed to characterize the interaction between the recombinantly produced SARS-CoV-2 Mpro (Figure S4C) and HB3-Core25. CD spectroscopy of the recombinant SARS-CoV-2 Mpro confirmed that the protease possesses a mixed secondary structure, displaying characteristic minima at 208 and 225 nm, and a maximum at 195 nm, consistent with both α-helical and β-sheet elements (Figure S5). Thermal denaturation assays demonstrated a complete loss of secondary structure above 55 °C, with a T m of 46.8 °C (Figure A).

4.

4

Open in a new tab

Analysis of Mpro binding and inhibition by HB3-Core25. (A) CD spectra of SARS-CoV-2 Mpro (blue line and dots) and the Mpro- HB3-Core25 complex (red line and dots) at a fixed wavelength of 222 nm, during thermal titration from 10 to 100 °C. The Tm of SARS-CoV-2 Mpro is 46.8 °C and it was kept unchanged upon binding to the inhibitor. Mpro concentration was kept constant at 5 μM for both experiments. Shaded areas represent experimental error over three independent measurements. (B) Size-exclusion chromatography of Mpro in the presence and absence of the HB3-Core25 miniprotein. The chromatogram shows two broad peaks with maxima values centered at 138 and 158 elution volumes. The first peak corresponds to the dimer population, while the second peak corresponds to the monomeric Mpro. A comparison shows a relatively larger population of the monomeric form of Mpro in the presence of the miniprotein. Integration of the area under the curve of the second peak reveals a 40% increase in the area for the complex. (C) MST binding assays. Each point on the dotted line represents the change in fluorescence of the labeled Mpro as the concentration of the HB3-Core25 inhibitor increases. The dissociation constant KD for the binding of the HB3-Core25 miniproteins to Mpro is 0.59 μM, 95% CI: 0.41–0.77 μM (dashed line). (D) Mpro enzymatic inhibition assay with HB3-Core25. Each point (blue) on the curve represents the concentration of the inhibitor. The inhibitory concentration (IC50) of HB3-Core25 corresponds to 0.27 μM, 95% CI: 0.12–0.35 μM. RLU stands for relative light units. The gray dotted line indicates the luminescence signal of the positive control. (E) Percentage of inhibition of Mpro enzymatic activity with increasing concentrations of HB3-Core25.

To determine whether HB3-Core25 forms a stable complex with Mpro, thermal shift assays were performed. Complex formation typically results in increased or unchanged thermostability due to cooperative folding and interaction. Upon incubation with HB3-Core25, the Tm of Mpro remained unaltered, indicating the formation of a thermodynamically stable complex and absence of structural changes in Mpro upon binding (Figure A). It is important to note that due to the monomer–dimer equilibrium of Mpro, the Tm is concentration-dependent. To ensure comparability, all denaturation assays were conducted using the same Mpro concentration. Binding affinity was further quantified by microscale thermophoresis (MST), using a fluorophore-labeled Mpro and a GST-tagged HB3-Core25 construct to enhance molecular weight and signal intensity (Figure S4B). The assays yielded a dissociation constant (K D) of 5.9 × 10–7 M, 95% confidence interval (CI): (4.1–0.7) × 10–7 M, confirming a direct and specific interaction between the miniprotein and Mpro (Figure B). The observed fluorescence changes were dose-dependent, further validating the binding interaction. Moreover, to confirm HB3-Core25 is effectively binding to the Mpro dimerization interface, MST assays were performed using an Mpro variant harboring two mutations previously demonstrated to abrogate dimerization (E290A and R298A) (Figure S6). While HB3-Core25 binds to wild-type Mpro, no binding was detected with the dimerization-deficient mutant, supporting our conclusion that HB3-Core25 specifically targets the dimerization interface of Mpro.

Finally, to assess the functional consequence of this interaction, enzymatic inhibition tests were conducted. HB3-Core25 effectively suppressed the catalytic activity of Mpro, with an inhibitory concentration (IC50) of 0.27 μM, 95% CI: 0.12–0.35 μM. At a concentration of 25 μM, the miniprotein inhibited 51.1% of substrate cleavage (Figure C,D). These results demonstrate that HB3-Core25 functions as an Mpro inhibitor, likely by interfering with the dimerization process required for enzymatic activation. It is supported by SEC experiments, where HB3-Core25 is shown to increase the Mpro monomer population by around 40% (Figure B).

Discussion

Viral proteases, such as the main protease (Mpro) of SARS-CoV-2, are attractive drug targets due to their essential role in viral replication and their high degree of conservation across coronavirus strains. ,− Since the functional unit of Mpro is a homodimer, blockers of protein dimerization should be in principle able to diminish its catalytic activity and may lead to potential broad-spectrum antiviral drugs. In this regard, eight small molecules targeting the Mpro dimerization region have been developed by other study groups to inhibit Mpro protease activity. − While effective to some extent, three of those molecules have shown no detectable half-maximal effective concentration (EC50) or half-maximal inhibitory concentration (IC50). The other class of inhibitors comprises peptides with high selectivity for Mpro and with low accumulation levels in the in vivo studies. Among these peptides, Ding et al. have described that the eight N-terminal residues (SGFRKMAF, named N8) of Mpro act as a dimerization inhibitor. , This strategy, along with the presence of well-defined Mpro allosteric sites, has been largely explored.

These studies highlighted that the N-terminus of Mpro plays a critical role in mediating dimerization. , Specifically, N-terminal residues S1, R4, and M6 are essential for stabilizing the dimer interface. S1 forms key interactions with F140 and E166 of the adjacent monomer, which are vital for maintaining the proper conformation of the catalytic site. , Additionally, R4 establishes a salt bridge with E290 from the opposite subunit, further contributing to dimer stability. ,− M6, in turn, projects into a hydrophobic pocket of the neighboring protomer, reinforcing the dimer through hydrophobic interactions. To effectively promote both Mpro dimer dissociation and inhibition of its catalytic activity, a molecule is required to form, at minimum, a hydrogen bond with E166. The critical role of this residue (conserved across all known human coronaviruses) has been highlighted in recent studies as essential for effective inhibition. Notably, Goyal and Goyal have identified E166 as one of the key residues that should be targeted to disrupt SARS-CoV Mpro dimerization.

In this context, we computationally designed HB3-Core25, a de novo miniprotein aimed to incorporate key traits of the binding interface of N8 and enhance interaction with the Mpro dimerization interface to inhibit its enzymatic activity. This strategy was motivated by the fact that, when compared to linear peptides, small proteins have greater functional diversity and higher structural stability through the precise design of noncovalent interactions, which results in superior biological function. ,− On the other hand, a key challenge in computational protein design has historically been the low rate of experimental successapproximately 2% at the time this study was initiated. , However, recent advances in artificial intelligence and deep learning-guided modeling have dramatically increased success rates to 30% or more. , Furthermore, when the design process incorporates chemical intuition in the manual supervision of interface geometry and residue–residue interactions, success rates can approach 50%. It is important to highlight that modern computational design strategies enable the prediction of high-confidence protein candidates. , Rather than requiring large-scale experimental screening, these approaches prioritize variants with the greatest likelihood of correct folding and target engagement. Previous studies have demonstrated that even a single rationally designed miniprotein can effectively validate the intended mechanism of action, particularly when guided by strong biophysical principles and structural intuition. , The development of HB3-Core25 reflects this approach. Among several computational candidates, HB3-Core25 was selected not only based on predicted stability and binding energy but also guided by visual inspection and chemically informed assessment of its interface geometry. Other studies have shown that conformational stabilization of miniproteins by substantial engineering is crucial to achieve potent inhibitory activity. − Adding modifications to the loops that connect the α-helices and redesigning new interactions to stabilize the coiled-coil structures are required to strongly enhance miniprotein’s propensity to fold into the intended compact three-helix bundle and to exert its proper biological function.

Validating this approach, the rational design of HB3-Core25 resulted in a miniprotein exhibiting a melting temperature exceeding 100 °C and a 24-fold higher affinity (K D = 0.59 μM, CI: 0.41–0.77 μM) by Mpro than the original N8 peptide used as a template (K D ≈ 14 μM). In addition, the HB3-Core25 miniprotein decreased Mpro in vitro substrate cleavage in half, with an IC50 of 0.27 μM, CI: 0.12–0.35 μM, about 2000 times less than the amount required of the N8 peptide to inhibit 46% of Mpro’s catalytic activity. When HB3-Core25 is compared with previously reported peptide inhibitors of Mpro, its inhibitory profile appears consistent with or superior to those targeting either the dimerization interface or the catalytic site. Stewart et al. reported peptide inhibitors optimized for Mpro dimer disruption, with affinities in the micromolar range but no evidence of thermal stability or structural rigidity comparable to de novo miniproteins. Moreover, peptides targeting the catalytic site, such as those described by Johansen-Leete et al. and macrocyclic peptides described by Harrison et al., , achieved IC50 values in the low micromolar range but showed limited stability in biochemical assays, posing a challenge for future in vivo applications. While other catalytic site inhibitors (such as dronedarone) can act directly at the catalytic site with reported potencies in the low micromolar to nanomolar range, inhibitors targeting Mpro dimerization interface typically require higher concentrations due to the broader and defined interaction surface. However, this strategy offers the advantage of acting on a highly conserved region of Mpro, which may reduce the likelihood of resistance development.

While the binding affinity of HB3-Core25 to Mpro is comparable to the well-characterized GC376, a prodrug (K D = 0.15 ± 0.03 μM), the clinically approved inhibitor nirmatrelvir displays a higher affinity (NMR/r; K D = 0.007 ± 0.004 μM). It is worth noting that these two compounds target Mpro’s active site, instead of the enzyme’s dimer interface, where high binding affinity does not seem to be a strict requirement for inhibitory activity. Silvestrini et al. investigated a range of compounds that inhibit Mpro’s catalytic activity by targeting either the monomer–monomer interface or the active site. Their results suggest a synergistic mechanism where compounds that do not strongly disrupt dimerization can still effectively inhibit enzymatic activity by acting on both sites. In line with these findings, HB3-Core25 miniprotein decreased Mpro in vitro substrate cleavage in half, even though the dissociation constant was also in the submicromolar range. This result can be partially explained by the fact that Mpro exists in a dynamic dimer–monomer equilibrium with dissociation constants reported between 1 and 7 μM. ,,,, Although this indicates a relatively weak affinity in absolute terms compared with nanomolar binders, it does not necessarily mean ineffective dimerization. Biologically, the intracellular concentration of Mpro favors dimer formation despite this, and therefore Mpro predominantly exists in its dimeric and catalytic active state. Therefore, at typical experimental concentrations, and even when a compound perturbs this equilibrium, a fraction of Mpro remains in its dimeric, active form and can still perform enzymatic activity in the presence of the substrate.

Altogether, our results demonstrate that HB3-Core25 is a thermostable inhibitor of SARS-CoV-2 Mpro. Designed to target the dimerization interface intracellularly, HB3-Core25 impairs Mpro activity likely by interfering in dimer formation (Figure B), offering an alternative mechanism of action to active-site inhibitors. In addition, it can also be synergistically used with Mpro’s active-site inhibitors to achieve higher levels of inhibition of the enzyme’s catalytic activity. While the current study focuses on the design and biophysical characterization of the miniprotein, immunogenicity and proteolytic stability analysis as well as strategies covering conjugation to cell-penetrating peptides, nanoparticle encapsulation, or mRNA-based delivery could be explored in future studies to ensure intracellular delivery, biocompatibility, and safety. Future in vitro and in vivo antiviral activity testing will be crucial to translate these findings into clinical applications. Nonetheless, this work underscores that computationally designed molecules, particularly when coupled with chemical intuition, can produce functional inhibitors with therapeutic potential.

Supplementary Material

ci5c01708_si_001.pdf (723.4KB, pdf)

Acknowledgments

We thank Camilla Adan for her fruitful comments and discussion. This work was supported by the Pernambuco Science and Technology Support Foundation (Fundação de Amparo à Ciência e Tecnologia do Estado de PernambucoFACEPE, Grant Numbers IBPG-0549-2-10/22, BFP-0010-2.11/22, and APQ-0346-2.09/19); the National Council for Scientific and Technological Development (Conselho Nacional de Desenvolvimento Científico e TecnológicoCNPq, Grant Numbers 165071/2018-4, 303833/2022-0, 425997/2018-9, INCT-FCx, and 307209/2023-7); the Oswaldo Cruz Foundation (Fundação Oswaldo CruzFiocruz, Grant Numbers VPPCB-007-FIO-18-2-134 and IAM-005-FIO-22-2-44); the Coordination for the Improvement of Higher Education Personnel Foundation (Coordenação de Aperfeiçoamento de Pessoal de Nível SuperiorCAPES, Grant Number 88887.814367/2023-00), the Fiocruz Genomic Surveillance Network (Grant Number VPGDI-050-FIO-20-2-13-36); and the Fundación Alfonso Martín Escudero. The LNCC Brazilian Supercomputing Center provided the partial allocation of computers. This work benefited from the infrastructure and expertise provided by the Protein Characterization Platform (RPT02I), Fiocruz, which supported the acquisition and analysis of protein biophysical data. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Glossary

Abbreviations

AS

alanine scanning

CD

circular dichroism

CMS

contact molecular surface

COM

center of mass

ΔInsatLH

buried unsatisfied hydrogen bonds

FES

free energy surface

HCV

hepatitis C virus

HIV

human immunodeficiency virus

IC50

inhibitory concentration

IPTG

isopropyl β-d-1-thiogalactopyranoside

K D

dissociation constant

LB

Luria–Bertani

MD

molecular dynamics

Mpro

main chymotrypsin-like protease

ML

machine learning

MST

microscale thermophoresis

NSP

nonstructural protein

nspa

buried nonpolar surface area

PLpro

papain-like protease

pLDDT

per-residue measure of local confidence

RMSD

root-mean-square deviation

RMSF

root-mean-square fluctuation

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

SASA

buried solvent-accessible surface area

SEC

size-exclusion chromatography

SS-Sc

shape complementarity

T m

melting temperature

Data supporting the findings of this study, including molecular data in PDB format and simulation scripts, are available at https://github.com/rlinslab/miniprotein-mpro-data. The Rosetta software is available at https://rosettacommons.org, GROMACS is available at https://www.gromacs.org, the PBEE software is available at https://github.com/chavesejf/PBEE, and Plumed is available at https://www.plumed.org.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.5c01708.

  • HB3-Core25 protein sequence and interaction profile between the N8 peptide and Mpro for SARS-CoV-2; metrics of monomers (A) and interfaces (B) for the selected sequences; structural stability analysis of protein–protein complexes via root-mean square deviation (RMSD); production of SARS-CoV-2 Mpro and HB3_CORE 25 recombinant proteins; biophysical characterization of SARS-CoV-2 Mpro; MST binding assay of double mutant E290A/R298A monomeric Mpro; metrics of the monomers and the interfaces between the miniproteins and Mpro, as well as the known dimerization inhibitors of Mpro from SARS-CoV-2; and percentages of interactions between amino acid residues in protein–protein complexes during molecular dynamics simulations (PDF)

∇.

T.E.L. and E.G.M. authors share first authorship. T.E.L. and E.G.M.: investigation, methodology, formal analysis, writingoriginal draft; D.F.C.: investigation, supervision, writingreview and editing; C.H.B.C.: investigation, formal analysis, supervision, writingreview and editing; R.D. and G.l.W.: resources, funding acquisition, writingreview and editing; B.H.S.L. and L.S.X.: investigation, methodology; M.P.I.: investigation, methodology, writingreview and editing; I.F.T.V.: investigation, funding acquisition, formal analysis, project administration, supervision, writingreview and editing; R.D.L.: conceptualization, investigation, funding acquisition, formal analysis, project administration, supervision, writingreview and editing.

The APC Funding Statement in DOTS is not present as a note. It should be "The Article Processing Charge for the publication of this research was funded by the Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES), Brazil (ROR identifier: 00x0ma614).

The authors declare no competing financial interest.

Published as part of Journal of Chemical Information and Modeling special issue “Computational Chemistry in the Global South: The Latin American Perspective”.

References

  1. Tong L.. Viral proteases. Chem. Rev. 2002;102:4609–4626. doi: 10.1021/cr010184f. [DOI] [PubMed] [Google Scholar]
  2. De Clercq E., Li G.. Approved Antiviral Drugs over the Past 50 Years. Clin. Microbiol. Rev. 2016;29:695–747. doi: 10.1128/CMR.00102-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Owen D. R., Allerton C. M. N., Anderson A. S., Aschenbrenner L., Avery M., Berritt S., Boras B., Cardin R. D., Carlo A., Coffman K. J., Dantonio A., Di L., Eng H., Ferre R., Gajiwala K. S., Gibson S. A., Greasley S. E., Hurst B. L., Kadar E. P., Kalgutkar A. S., Lee J. C., Lee J., Liu W., Mason S. W., Noell S., Novak J. J., Obach R. S., Ogilvie K., Patel N. C., Pettersson M., Rai D. K., Reese M. R., Sammons M. F., Sathish J. G., Singh R. S. P., Steppan C. M., Stewart A. E., Tuttle J. B., Updyke L., Verhoest P. R., Wei L., Yang Q., Zhu Y.. An oral SARS-CoV-2 M­(pro) inhibitor clinical candidate for the treatment of COVID-19. Science. 2021;374:1586–1593. doi: 10.1126/science.abl4784. [DOI] [PubMed] [Google Scholar]
  4. Goyal B., Goyal D.. Targeting the Dimerization of the Main Protease of Coronaviruses: A Potential Broad-Spectrum Therapeutic Strategy. ACS Comb. Sci. 2020;22:297–305. doi: 10.1021/acscombsci.0c00058. [DOI] [PubMed] [Google Scholar]
  5. Bai C., Zhong Q., Gao G. F.. Overview of SARS-CoV-2 genome-encoded proteins. Sci. China Life Sci. 2022;65:280–294. doi: 10.1007/s11427-021-1964-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Nashed N. T., Kneller D. W., Coates L., Ghirlando R., Aniana A., Kovalevsky A., Louis J. M.. Autoprocessing and oxyanion loop reorganization upon GC373 and nirmatrelvir binding of monomeric SARS-CoV-2 main protease catalytic domain. Commun. Biol. 2022;5:976. doi: 10.1038/s42003-022-03910-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Noske G. D., Song Y., Fernandes R. S., Chalk R., Elmassoudi H., Koekemoer L., Owen C. D., El-Baba T. J., Robinson C. V., Oliva G., Godoy A. S., The C. M. C.. An in-solution snapshot of SARS-COV-2 main protease maturation process and inhibition. Nat. Commun. 2023;14:1545. doi: 10.1038/s41467-023-37035-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Jin Z., Du X., Xu Y., Deng Y., Liu M., Zhao Y., Zhang B., Li X., Zhang L., Peng C., Duan Y., Yu J., Wang L., Yang K., Liu F., Jiang R., Yang X., You T., Liu X., Yang X., Bai F., Liu H., Liu X., Guddat L. W., Xu W., Xiao G., Qin C., Shi Z., Jiang H., Rao Z., Yang H.. Structure of M­(pro) from SARS-CoV-2 and discovery of its inhibitors. Nature. 2020;582:289–293. doi: 10.1038/s41586-020-2223-y. [DOI] [PubMed] [Google Scholar]
  9. Mengist H. M., Dilnessa T., Jin T.. Structural Basis of Potential Inhibitors Targeting SARS-CoV-2 Main Protease. Front. Chem. 2021;9:622898. doi: 10.3389/fchem.2021.622898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Johansen-Leete J., Ullrich S., Fry S. E., Frkic R., Bedding M. J., Aggarwal A., Ashhurst A. S., Ekanayake K. B., Mahawaththa M. C., Sasi V. M., Luedtke S., Ford D. J., O’Donoghue A. J., Passioura T., Larance M., Otting G., Turville S., Jackson C. J., Nitsche C., Payne R. J.. Antiviral cyclic peptides targeting the main protease of SARS-CoV-2. Chem. Sci. 2022;13:3826–3836. doi: 10.1039/D1SC06750H. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Harrison K., Carlos P. W., Ullrich S., Aggarwal A., Johansen-Leete J., Sasi V. M., Barter I., Maxwell J. W. C., Bedding M. J., Larance M., Turville S., Norman A., Jackson C. J., Nitsche C., Payne R. J.. Exploiting Hydrophobic Amino Acid Scanning to Develop Cyclic Peptide Inhibitors of the SARS-CoV-2 Main Protease with Antiviral Activity. Chem. - Eur. J. 2024;30:e202401606. doi: 10.1002/chem.202401606. [DOI] [PubMed] [Google Scholar]
  12. Stewart J., Shawon J., Ali M. A., Williams B., Shahinuzzaman A. D. A., Rupa S. A., Al-Adhami T., Jia R., Bourque C., Faddis R., Stone K., Sufian M. A., Islam R., McShan A. C., Rahman K. M., Halim M. A.. Antiviral peptides inhibiting the main protease of SARS-CoV-2 investigated by computational screening and in vitro protease assay. J. Pept. Sci. 2024;30:e3553. doi: 10.1002/psc.3553. [DOI] [PubMed] [Google Scholar]
  13. Wei P., Fan K., Chen H., Ma L., Huang C., Tan L., Xi D., Li C., Liu Y., Cao A., Lai L.. The N-terminal octapeptide acts as a dimerization inhibitor of SARS coronavirus 3C-like proteinase. Biochem. Biophys. Res. Commun. 2006;339:865–872. doi: 10.1016/j.bbrc.2005.11.102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Ding L., Zhang X. X., Wei P., Fan K., Lai L.. The interaction between severe acute respiratory syndrome coronavirus 3C-like proteinase and a dimeric inhibitor by capillary electrophoresis. Anal. Biochem. 2005;343:159–165. doi: 10.1016/j.ab.2005.04.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cao L., Goreshnik I., Coventry B., Case J. B., Miller L., Kozodoy L., Chen R. E., Carter L., Walls A. C., Park Y. J., Strauch E. M., Stewart L., Diamond M. S., Veesler D., Baker D.. De novo design of picomolar SARS-CoV-2 miniprotein inhibitors. Science. 2020;370:426–431. doi: 10.1126/science.abd9909. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chaves E. J. F., Coelho D. F., Cruz C. H. B., Moreira E. G., Simoes J. C. M., Nascimento-Filho M. J., Lins R. D.. Structure-based computational design of antibody mimetics: challenges and perspectives. FEBS Open Bio. 2025;15:223–235. doi: 10.1002/2211-5463.13855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Viana I. F., Cruz C. H., Athayde D., Adan W. C., Xavier L. S. S., Archer M., Lins R. D.. In vitro neutralisation of Zika virus by an engineered protein targeting the viral envelope fusion loop. Mol. Syst. Des. Eng. 2023;8:516–526. doi: 10.1039/D2ME00170E. [DOI] [Google Scholar]
  18. Baker E. G., Bartlett G. J., Porter Goff K. L., Woolfson D. N.. Miniprotein Design: Past, Present, and Prospects. Acc. Chem. Res. 2017;50:2085–2092. doi: 10.1021/acs.accounts.7b00186. [DOI] [PubMed] [Google Scholar]
  19. Leaver-Fay, A. ; Tyka, M. ; Lewis, S. M. ; Lange, O. F. ; Thompson, J. ; Jacak, R. ; Kaufman, K. W. ; Renfrew, P. D. ; Smith, C. A. ; Sheffler, W. ; Davis, I. W. ; Cooper, S. ; Treuille, A. ; Mandell, D. J. ; Richter, F. ; Ban, Y.-E. A. ; Fleishman, S. J. ; Corn, J. E. ; Kim, D. E. ; Lyskov, S. ; Berrondo, M. ; Mentzer, S. ; Popović, Z. ; Havranek, J. J. ; Karanicolas, J. ; Das, R. ; Meiler, J. ; Kortemme, T. ; Gray, J. J. ; Kuhlman, B. ; Baker, D. ; Bradley, P. . Chapter Nineteen - Rosetta3: An Object-Oriented Software Suite for the Simulation and Design of Macromolecules. In Methods in Enzymology; Johnson, M. L. ; Brand, L. , Eds.; Academic Press, 2011; Vol. 487, pp 545–574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Alford R. F., Leaver-Fay A., Jeliazkov J. R., O’Meara M. J., DiMaio F. P., Park H., Shapovalov M. V., Renfrew P. D., Mulligan V. K., Kappel K., Labonte J. W., Pacella M. S., Bonneau R., Bradley P., Dunbrack R. L. Jr., Das R., Baker D., Kuhlman B., Kortemme T., Gray J. J.. The Rosetta All-Atom Energy Function for Macromolecular Modeling and Design. J. Chem. Theory Comput. 2017;13:3031–3048. doi: 10.1021/acs.jctc.7b00125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Patick A. K., Potts K. E.. Protease inhibitors as antiviral agents. Clin. Microbiol. Rev. 1998;11:614–627. doi: 10.1128/CMR.11.4.614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Studer R. A., Christin P.-A., Williams M. A., Orengo C. A.. Stability-activity tradeoffs constrain the adaptive evolution of RubisCO. Proc. Natl. Acad. Sci. USA. 2014;111:2223–2228. doi: 10.1073/pnas.1310811111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Schymkowitz J., Borg J., Stricher F., Nys R., Rousseau F., Serrano L.. The FoldX web server: an online force field. Nucleic Acids Res. 2005;33:W382–W388. doi: 10.1093/nar/gki387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Zhang N., Chen Y., Lu H., Zhao F., Alvarez R. V., Goncearenco A., Panchenko A. R., Li M.. MutaBind2: Predicting the Impacts of Single and Multiple Mutations on Protein-Protein Interactions. IScience. 2020;23:100939. doi: 10.1016/j.isci.2020.100939. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kortemme T., Kim D. E., Baker D.. Computational Alanine Scanning of Protein-Protein Interfaces. Science’s STKE. 2004;2004:pl2. doi: 10.1126/stke.2192004pl2. [DOI] [PubMed] [Google Scholar]
  26. Chevalier A., Silva D. A., Rocklin G. J., Hicks D. R., Vergara R., Murapa P., Bernard S. M., Zhang L., Lam K. H., Yao G., Bahl C. D., Miyashita S. I., Goreshnik I., Fuller J. T., Koday M. T., Jenkins C. M., Colvin T., Carter L., Bohn A., Bryan C. M., Fernandez-Velasco D. A., Stewart L., Dong M., Huang X., Jin R., Wilson I. A., Fuller D. H., Baker D.. Massively parallel de novo protein design for targeted therapeutics. Nature. 2017;550:74–79. doi: 10.1038/nature23912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Sakuma K., Minami S.. Enumeration and comprehensive in-silico modeling of three-helix bundle structures composed of typical αα-hairpins. BMC Bioinf. 2021;22:465. doi: 10.1186/s12859-021-04380-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Huang P. S., Ban Y. E., Richter F., Andre I., Vernon R., Schief W. R., Baker D.. RosettaRemodel: a generalized framework for flexible backbone protein design. PLoS One. 2011;6:e24109. doi: 10.1371/journal.pone.0024109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Fleishman S. J., Leaver-Fay A., Corn J. E., Strauch E. M., Khare S. D., Koga N., Ashworth J., Murphy P., Richter F., Lemmon G., Meiler J., Baker D.. RosettaScripts: a scripting language interface to the Rosetta macromolecular modeling suite. PLoS One. 2011;6:e20161. doi: 10.1371/journal.pone.0020161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Rocklin G. J., Chidyausiku T. M., Goreshnik I., Ford A., Houliston S., Lemak A., Carter L., Ravichandran R., Mulligan V. K., Chevalier A., Arrowsmith C. H., Baker D.. Global analysis of protein folding using massively parallel design, synthesis, and testing. Science. 2017;357:168–175. doi: 10.1126/science.aan0693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Maguire J. B., Haddox H. K., Strickland D., Halabiya S. F., Coventry B., Griffin J. R., Pulavarti S. V. S. R. K., Cummins M., Thieker D. F., Klavins E., Szyperski T., DiMaio F., Baker D., Kuhlman B.. Perturbing the energy landscape for improved packing during computational protein design. Proteins: Struct., Funct., Bioinf. 2021;89:436–449. doi: 10.1002/prot.26030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mirdita M., Schütze K., Moriwaki Y., Heo L., Ovchinnikov S., Steinegger M.. ColabFold: making protein folding accessible to all. Nat. Methods. 2022;19:679–682. doi: 10.1038/s41592-022-01488-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kim D. E., Chivian D., Baker D.. Protein structure prediction and analysis using the Robetta server. Nucleic Acids Res. 2004;32:W526–31. doi: 10.1093/nar/gkh468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Spoel, L. A. H. v. d. GROMACS 2021.4 Manual 2021; Vol. 4, p 2021. [Google Scholar]
  35. Huang J., Rauscher S., Nawrocki G., Ran T., Feig M., de Groot B. L., Grubmüller H., MacKerell A. D. Jr.. CHARMM36m: an improved force field for folded and intrinsically disordered proteins. Nat. Methods. 2017;14:71–73. doi: 10.1038/nmeth.4067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Gordon J. C., Myers J. B., Folta T., Shoja V., Heath L. S., Onufriev A.. H++: a server for estimating pKas and adding missing hydrogens to macromolecules. Nucleic Acids Res. 2005;33:W368–W371. doi: 10.1093/nar/gki464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Bussi G., Donadio D., Parrinello M.. Canonical sampling through velocity rescaling. J. Chem. Phys. 2007;126:014101. doi: 10.1063/1.2408420. [DOI] [PubMed] [Google Scholar]
  38. Parrinello M., Rahman A.. Polymorphic transitions in single crystals: A new molecular dynamics method. J. Appl. Phys. 1981;52:7182–7190. doi: 10.1063/1.328693. [DOI] [Google Scholar]
  39. Hess B.. P-LINCS: A Parallel Linear Constraint Solver for Molecular Simulation. J. Chem. Theory Comput. 2008;4:116–122. doi: 10.1021/ct700200b. [DOI] [PubMed] [Google Scholar]
  40. Hess B., Bekker H., Berendsen H. J. C., Fraaije J. G. E. M.. LINCS: A linear constraint solver for molecular simulations. J. Comput. Chem. 1997;18:1463–1472. doi: 10.1002/(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.CO;2-H. [DOI] [Google Scholar]
  41. Van Gunsteren W. F., Berendsen H. J. C.. A leap-frog algorithm for stochastic dynamics. Mol. Simul. 1988;1:173–185. doi: 10.1080/08927028808080941. [DOI] [Google Scholar]
  42. Zhang L., Lin D., Sun X., Curth U., Drosten C., Sauerhering L., Becker S., Rox K., Hilgenfeld R.. Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved alpha-ketoamide inhibitors. Science. 2020;368:409–412. doi: 10.1126/science.abb3405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Chaves E. J. F., Sartori J., Santos W. M., Cruz C. H. B., Mhrous E. N., Nacimento-Filho M. F., Ferraz M. V. F., Lins R. D.. Estimating Absolute Protein–Protein Binding Free Energies by a Super Learner Model. J. Chem. Inf. Model. 2025;65:2602–2609. doi: 10.1021/acs.jcim.4c01641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nashed N. T., Aniana A., Ghirlando R., Chiliveri S. C., Louis J. M.. Modulation of the monomer-dimer equilibrium and catalytic activity of SARS-CoV-2 main protease by a transition-state analog inhibitor. Commun. Biol. 2022;5:160. doi: 10.1038/s42003-022-03084-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Dai W., Zhang B., Jiang X. M., Su H., Li J., Zhao Y., Xie X., Jin Z., Peng J., Liu F., Li C., Li Y., Bai F., Wang H., Cheng X., Cen X., Hu S., Yang X., Wang J., Liu X., Xiao G., Jiang H., Rao Z., Zhang L. K., Xu Y., Yang H., Liu H.. Structure-based design of antiviral drug candidates targeting the SARS-CoV-2 main protease. Science. 2020;368:1331–1335. doi: 10.1126/science.abb4489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Liu J., Xu L., Guo W., Li Z., Khan M. K. H., Ge W., Patterson T. A., Hong H.. Developing a SARS-CoV-2 main protease binding prediction random forest model for drug repurposing for COVID-19 treatment. Exp. Biol. Med. 2023;248:1927–1936. doi: 10.1177/15353702231209413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yang Y., Zhao Z., Li X., Chen Y., Liu L., Zhang S. L., Yang A.. Design and Evaluation of Peptide Inhibitors Targeting the Dimerization of SARS-CoV-2 Main Protease. ChemBioChem. 2025;26:e202400688. doi: 10.1002/cbic.202400688. [DOI] [PubMed] [Google Scholar]
  48. Günther S., Reinke P. Y. A., Fernandez-Garcia Y., Lieske J., Lane T. J., Ginn H. M., Koua F. H. M., Ehrt C., Ewert W., Oberthuer D., Yefanov O., Meier S., Lorenzen K., Krichel B., Kopicki J. D., Gelisio L., Brehm W., Dunkel I., Seychell B., Gieseler H., Norton-Baker B., Escudero-Perez B., Domaracky M., Saouane S., Tolstikova A., White T. A., Hanle A., Groessler M., Fleckenstein H., Trost F., Galchenkova M., Gevorkov Y., Li C., Awel S., Peck A., Barthelmess M., Schlunzen F., Lourdu Xavier P., Werner N., Andaleeb H., Ullah N., Falke S., Srinivasan V., Franca B. A., Schwinzer M., Brognaro H., Rogers C., Melo D., Zaitseva-Kinneberg J. I., Knoska J., Pena-Murillo G. E., Mashhour A. R., Hennicke V., Fischer P., Hakanpaa J., Meyer J., Gribbon P., Ellinger B., Kuzikov M., Wolf M., Beccari A. R., Bourenkov G., von Stetten D., Pompidor G., Bento I., Panneerselvam S., Karpics I., Schneider T. R., Garcia-Alai M. M., Niebling S., Gunther C., Schmidt C., Schubert R., Han H., Boger J., Monteiro D. C. F., Zhang L., Sun X., Pletzer-Zelgert J., Wollenhaupt J., Feiler C. G., Weiss M. S., Schulz E. C., Mehrabi P., Karnicar K., Usenik A., Loboda J., Tidow H., Chari A., Hilgenfeld R., Uetrecht C., Cox R., Zaliani A., Beck T., Rarey M., Gunther S., Turk D., Hinrichs W., Chapman H. N., Pearson A. R., Betzel C., Meents A.. X-ray screening identifies active site and allosteric inhibitors of SARS-CoV-2 main protease. Science. 2021;372:642–646. doi: 10.1126/science.abf7945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. El-Baba T. J., Lutomski C. A., Kantsadi A. L., Malla T. R., John T., Mikhailov V., Bolla J. R., Schofield C. J., Zitzmann N., Vakonakis I., Robinson C. V.. Allosteric Inhibition of the SARS-CoV-2 Main Protease: Insights from Mass Spectrometry Based Assays*. Angew. Chem., Int. Ed. 2020;59:23544–23548. doi: 10.1002/anie.202010316. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Douangamath A., Fearon D., Gehrtz P., Krojer T., Lukacik P., Owen C. D., Resnick E., Strain-Damerell C., Aimon A., Ábrányi-Balogh P., Brandão-Neto J., Carbery A., Davison G., Dias A., Downes T. D., Dunnett L., Fairhead M., Firth J. D., Jones S. P., Keeley A., Keserü G. M., Klein H. F., Martin M. P., Noble M. E. M., O’Brien P., Powell A., Reddi R. N., Skyner R., Snee M., Waring M. J., Wild C., London N., von Delft F., Walsh M. A.. Crystallographic and electrophilic fragment screening of the SARS-CoV-2 main protease. Nat. Commun. 2020;11:5047. doi: 10.1038/s41467-020-18709-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Fosgerau K., Hoffmann T.. Peptide therapeutics: current status and future directions. Drug Discovery Today. 2015;20:122–128. doi: 10.1016/j.drudis.2014.10.003. [DOI] [PubMed] [Google Scholar]
  52. Cheng S.-C., Chang G.-G., Chou C.-Y.. Mutation of Glu-166 Blocks the Substrate-Induced Dimerization of SARS Coronavirus Main Protease. Biophys. J. 2010;98:1327–1336. doi: 10.1016/j.bpj.2009.12.4272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Chen S., Zhang J., Hu T., Chen K., Jiang H., Shen X.. Residues on the dimer interface of SARS coronavirus 3C-like protease: dimer stability characterization and enzyme catalytic activity analysis. J. Biochem. 2007;143:525–536. doi: 10.1093/jb/mvm246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Hsu W. C., Chang H. C., Chou C. Y., Tsai P. J., Lin P. I., Chang G. G.. Critical assessment of important regions in the subunit association and catalytic action of the severe acute respiratory syndrome coronavirus main protease. J. Biol. Chem. 2005;280:22741–22748. doi: 10.1074/jbc.M502556200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Chou C. Y., Chang H. C., Hsu W. C., Lin T. Z., Lin C. H., Chang G. G.. Quaternary structure of the severe acute respiratory syndrome (SARS) coronavirus main protease. Biochemistry. 2004;43:14958–14970. doi: 10.1021/bi0490237. [DOI] [PubMed] [Google Scholar]
  56. Silvestrini L., Belhaj N., Comez L., Gerelli Y., Lauria A., Libera V., Mariani P., Marzullo P., Ortore M. G., Palumbo Piccionello A., Petrillo C., Savini L., Paciaroni A., Spinozzi F.. The dimer-monomer equilibrium of SARS-CoV-2 main protease is affected by small molecule inhibitors. Sci. Rep. 2021;11:9283. doi: 10.1038/s41598-021-88630-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Shitrit A., Zaidman D., Kalid O., Bloch I., Doron D., Yarnizky T., Buch I., Segev I., Ben-Zeev E., Segev E., Kobiler O.. Conserved interactions required for inhibition of the main protease of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Sci. Rep. 2020;10:20808. doi: 10.1038/s41598-020-77794-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Baker E. G., Williams C., Hudson K. L., Bartlett G. J., Heal J. W., Porter Goff K. L., Sessions R. B., Crump M. P., Woolfson D. N.. Engineering protein stability with atomic precision in a monomeric miniprotein. Nat. Chem. Biol. 2017;13:764–770. doi: 10.1038/nchembio.2380. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Ozga K., Berlicki Ł.. Design and Engineering of Miniproteins. ACS Bio Med. Chem. Au. 2022;2:316–327. doi: 10.1021/acsbiomedchemau.2c00008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Storz G., Wolf Y. I., Ramamurthi K. S.. Small Proteins Can No Longer Be Ignored. Annu. Rev. Biochem. 2014;83:753–777. doi: 10.1146/annurev-biochem-070611-102400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Dai L., Yang Y., Kim H. R., Zhou Y.. Improving computational protein design by using structure-derived sequence profile. Proteins: Struct., Funct., Bioinf. 2010;78:2338–2348. doi: 10.1002/prot.22746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Huang P.-S., Boyken S. E., Baker D.. The coming of age of de novo protein design. Nature. 2016;537:320–327. doi: 10.1038/nature19946. [DOI] [PubMed] [Google Scholar]
  63. Winnifrith A., Outeiral C., Hie B. L.. Generative artificial intelligence for de novo protein design. Curr. Opin. Struct. Biol. 2024;86:102794. doi: 10.1016/j.sbi.2024.102794. [DOI] [PubMed] [Google Scholar]
  64. Watson J. L., Juergens D., Bennett N. R., Trippe B. L., Yim J., Eisenach H. E., Ahern W., Borst A. J., Ragotte R. J., Milles L. F., Wicky B. I. M., Hanikel N., Pellock S. J., Courbet A., Sheffler W., Wang J., Venkatesh P., Sappington I., Torres S. V., Lauko A., De Bortoli V., Mathieu E., Ovchinnikov S., Barzilay R., Jaakkola T. S., DiMaio F., Baek M., Baker D.. De novo design of protein structure and function with RFdiffusion. Nature. 2023;620:1089–1100. doi: 10.1038/s41586-023-06415-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Stranges P. B., Kuhlman B.. A comparison of successful and failed protein interface designs highlights the challenges of designing buried hydrogen bonds. Protein Sci. 2013;22:74–82. doi: 10.1002/pro.2187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Silva D.-A., Yu S., Ulge U. Y., Spangler J. B., Jude K. M., Labão-Almeida C., Ali L. R., Quijano-Rubio A., Ruterbusch M., Leung I., Biary T., Crowley S. J., Marcos E., Walkey C. D., Weitzner B. D., Pardo-Avila F., Castellanos J., Carter L., Stewart L., Riddell S. R., Pepper M., Bernardes G. J. L., Dougan M., Garcia K. C., Baker D.. De novo design of potent and selective mimics of IL-2 and IL-15. Nature. 2019;565:186–191. doi: 10.1038/s41586-018-0830-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Anishchenko I., Pellock S. J., Chidyausiku T. M., Ramelot T. A., Ovchinnikov S., Hao J., Bafna K., Norn C., Kang A., Bera A. K., DiMaio F., Carter L., Chow C. M., Montelione G. T., Baker D.. De novo protein design by deep network hallucination. Nature. 2021;600:547–552. doi: 10.1038/s41586-021-04184-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Cano-Muñoz M., Lucas J., Lin L. Y., Cesaro S., Moog C., Conejero-Lara F.. Conformational Stabilization of Gp41-Mimetic Miniproteins Opens Up New Ways of Inhibiting HIV-1 Fusion. Int. J. Mol. Sci. 2022;23:2794. doi: 10.3390/ijms23052794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Polo-Megías D., Cano-Muñoz M., Berruezo A. G., Laumond G., Moog C., Conejero-Lara F.. Exploring Highly Conserved Regions of SARS-CoV-2 Spike S2 Subunit as Targets for Fusion Inhibition Using Chimeric Proteins. Int. J. Mol. Sci. 2022;23:15511. doi: 10.3390/ijms232415511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Polo-Megías D., Cano-Muñoz M., Gantner P., Laumond G., Decoville T., Grezzani J., La Rocchia I., Salinas-Garcia M. C., Cámara-Artigas A., Gavira J. A., Conejero-Lara F., Moog C.. Potent HIV-1 miniprotein inhibitors targeting highly conserved gp41 epitopes. Int. J. Biol. Macromol. 2025;310:143157. doi: 10.1016/j.ijbiomac.2025.143157. [DOI] [PubMed] [Google Scholar]
  71. Ahmed M. S., Farag A. B., Boys I. N., Wang P., Menendez-Montes I., Nguyen N. U. N., Eitson J. L., Ohlson M. B., Fan W., McDougal M. B., Mar K., Thet S., Ortiz F., Kim S. Y., Solmonson A., Williams N. S., Lemoff A., DeBerardinis R. J., Schoggins J. W., Sadek H. A.. FDA approved drugs with antiviral activity against SARS-CoV-2: From structure-based repurposing to host-specific mechanisms. Biomed. Pharmacother. 2023;162:114614. doi: 10.1016/j.biopha.2023.114614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Paciaroni A., Libera V., Ripanti F., Orecchini A., Petrillo C., Francisci D., Schiaroli E., Sabbatini S., Gidari A., Bianconi E., Macchiarulo A., Hussain R., Silvestrini L., Moretti P., Belhaj N., Vercelli M., Roque Y., Mariani P., Comez L., Spinozzi F.. Stabilization of the Dimeric State of SARS-CoV-2 Main Protease by GC376 and Nirmatrelvir. Int. J. Mol. Sci. 2023;24:6062. doi: 10.3390/ijms24076062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Polo-Megías D., Cano-Muñoz M., Berruezo A. G., Laumond G., Moog C., Conejero-Lara F.. Investigating vulnerability of the conserved SARS-CoV-2 spike’s heptad repeat 2 as target for fusion inhibitors using chimeric miniproteins. Int. J. Biol. Macromol. 2024;262:130132. doi: 10.1016/j.ijbiomac.2024.130132. [DOI] [PubMed] [Google Scholar]
  74. Xia B., Kang X.. Activation and maturation of SARS-CoV main protease. Protein Cell. 2011;2:282–290. doi: 10.1007/s13238-011-1034-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ci5c01708_si_001.pdf (723.4KB, pdf)

Data Availability Statement

Data supporting the findings of this study, including molecular data in PDB format and simulation scripts, are available at https://github.com/rlinslab/miniprotein-mpro-data. The Rosetta software is available at https://rosettacommons.org, GROMACS is available at https://www.gromacs.org, the PBEE software is available at https://github.com/chavesejf/PBEE, and Plumed is available at https://www.plumed.org.

Articles from Journal of Chemical Information and Modeling are provided here courtesy of American Chemical Society

RESOURCES