Identification of SARS-CoV-2 Mpro inhibitors containing P1' 4-fluorobenzothiazole moiety highly active against SARS-CoV-2

The COVID-19 pandemic persists as a serious public health threat. Although vaccines were rapidly developed, the emergence of SARS-CoV-2 variants with spike mutations raises concerns about vaccine efficacy and the loss of activity of several monoclonal antibodies. Only a few antivirals (remdesivir, nirmatrelvir, and molnupiravir) are available, and their potency can be insufficient; for example, nirmatrelvir has been associated with virologic rebound and potential resistance. Given SARS-CoV-2’s high replication rate and ability to develop drug resistance, there is an urgent need for more potent, safer, and better-tolerated therapeutics. Building on prior Mpro inhibitors (e.g., GRL1720, GRL0920, and 5h), the authors sought to design improved inhibitors and identified two novel 4-fluorobenzothiazole-containing compounds, TKB245 and TKB248, that specifically target Mpro and exhibit superior antiviral potency in vitro and efficacy in vivo.

The study builds on prior work targeting SARS-CoV-2 Mpro and related coronavirus proteases. Earlier inhibitors such as GRL-0920 and compounds like 5h showed activity but had limitations in potency and pharmacokinetics. Existing authorized antivirals (remdesivir, molnupiravir, nirmatrelvir) face issues including moderate potency, potential for resistance, and clinical rebound with nirmatrelvir-ritonavir. Mpro’s sequence and structure are conserved among betacoronaviruses, and experimental Mpro inhibitors often retain activity across variants, suggesting Mpro is a robust target. Fluorination is known to enhance lipophilicity, membrane penetration, metabolic stability, and oral bioavailability, motivating the design of fluorinated analogs of 5h to improve antiviral and pharmacologic properties.

- Compound design and synthesis: Starting from benzothiazole-containing Mpro inhibitor 5h, a series of fluorinated analogs were synthesized by introducing fluorine atoms on the phenyl ring and optimizing P3 and P2 moieties. TKB245 and TKB248 were developed, differing by a central carbonyl (TKB245) versus thiocarbonyl (TKB248). - Enzyme inhibition assays: SARS-CoV-2 Mpro activity was measured using a FRET-based assay (BPS Bioscience). Selectivity was tested against human cysteine proteases cathepsin L and calpain using commercial kits. IC50 values were calculated from dose-response curves. - Cell-based antiviral assays: Multiple cell lines (VeroE6; VeroE6 with P-gp inhibitor CP-100356; HeLa-ACE2-TMPRSS2; A549-ACE2-TMPRSS2) were infected with SARS-CoV-2 strains/variants at defined MOIs. Antiviral activity was quantified by RT-qPCR of viral RNA from culture supernatants. Cytotoxicity (CC50) was assessed by WST-8 assay. - Variant coverage: Activity was tested against alpha (B.1.1.7), beta (B.1.351), gamma (P.1), delta (B.1.617.2), kappa (B.1.617.1), and omicron (BA.1, BA.2, BA.5) variants. - Intracellular concentration measurements: VeroE6 and HeLa-ACE2-TMPRSS2 cells were incubated with 10, 50, 100 µM of TKB245, TKB248, or nirmatrelvir for 6 h; cells were washed, extracted, and analyzed by LC/MS to quantify intracellular compound levels. - Pharmacokinetics (PK): In PXB (human liver chimeric) mice, 10 mg/kg i.v. and p.o. dosing of TKB245, TKB248, and nirmatrelvir were administered; plasma concentrations were quantified by LC-MS/MS to calculate t1/2 and oral bioavailability (F). Earlier PK for 5h, TKB125, TKB198 used ICR mice. - In vivo efficacy: Human ACE2 knock-in mice were intranasally infected with SARS-CoV-2 Delta or Omicron (BA.1, BA.2). Two hours post infection, mice received intraperitoneal TKB245 (100 mg/kg BID) or TKB248 (120 mg/kg BID) versus vehicle. Lungs were collected at 2 and 3 dpi to determine infectious viral titers by plaque assay on VeroE6/TMPRSS2. Body weight, survival, micro-CT imaging, and histopathology were also assessed. - Native mass spectrometry: Authentic and glycine-modified Mpro were incubated with inhibitors and analyzed by ESI-QTOF native MS to assess monomer-dimer equilibria and inhibitor binding stoichiometry. - X-ray crystallography: Co-crystal structures of Mpro with TKB245 (PDB 8DOX) and TKB248 (PDB 8DPR) were solved. Crystals were grown by hanging drop vapor diffusion; data collected at SPring-8 BL41XU; structures refined with standard crystallographic software. Binding modes, covalent adducts to Cys145, and interactions within S1/S2/S3/S1' subsites were analyzed. - Statistics: Wilcoxon rank-sum tests compared lung viral titers between treated and vehicle groups; repeated ANOVA included day and strain as fixed effects and individual as random effect.

- Potent antiviral activity: TKB245 showed EC50 ≈ 0.03 µM (mean ± SD 0.03 ± 0.02 µM) and EC95 ≈ 0.53 µM in VeroE6 cells; Mpro enzymatic IC50 = 0.007 ± 0.002 µM. TKB248 exhibited EC50 0.22 ± 0.08 µM; Mpro IC50 ≈ 0.074 ± 0.034 µM. - Broad variant coverage: TKB245 was highly active against alpha, beta, gamma, delta, kappa, and omicron (BA.1/BA.2/BA.5) with EC50 range 0.014–0.056 µM (variant- and isolate-dependent). TKB248 EC50 range 0.070–0.430 µM. Both outperformed nirmatrelvir on a molar basis in VeroE6 assays (Table 2). - Superior performance across cell types: In VeroE6 with P-gp inhibitor CP100356, EC50s were 0.001 µM (TKB245), 0.03 µM (TKB248), and 0.025 µM (nirmatrelvir). In HeLa-ACE2-TMPRSS2, TKB245 EC50 0.021 µM; in A549-ACE2-TMPRSS2, 0.0027 µM. TKB245 was more potent than ensitrelvir and molnupiravir in side-by-side assays. - Intracellular accumulation: Compared to nirmatrelvir, TKB245 accumulated 39-fold (50 µM) and 19.5-fold (100 µM) higher in VeroE6; 2.3-fold higher in HeLa-ACE2-TMPRSS2 at both 50 and 100 µM. TKB248 showed dramatically higher intracellular levels versus nirmatrelvir: 4800-fold (50 µM) and 2289-fold (100 µM) in VeroE6; 318-fold (50 µM) and 165-fold (100 µM) in HeLa-ACE2-TMPRSS2. - Pharmacokinetics: In PXB mice, TKB245 t1/2 ≈ 3.82 h (p.o.), oral F ≈ 48%; TKB248 t1/2 ≈ 4.34 h, oral F ≈ 72%. Nirmatrelvir t1/2 ≈ 1.03 h, oral F ≈ 56% (same model). TKB245/248 improved PK over earlier analogs (e.g., 5h t1/2 ≈ 0.278 h). - In vivo efficacy: In hACE2 knock-in mice infected with Delta or Omicron variants, TKB245 (100 mg/kg BID, i.p.) and TKB248 (120 mg/kg BID, i.p.) significantly reduced lung viral titers versus vehicle. Repeated ANOVA: TKB245 reduced titers by −1.34 log10 PFU/g (95% CI: −1.85, −0.82); TKB248 by −1.01 (95% CI: −1.41, −0.60). No significant changes in body weight, survival, micro-CT, or histopathology at early time points, and no signs of acute toxicity. - Mechanism and structural insights: Native MS showed TKB245 and nirmatrelvir promote Mpro dimerization, predominantly forming dimers bound by two inhibitor molecules; TKB248 required higher concentration to show similar effects, consistent with its higher IC50. X-ray structures revealed both TKB245 and TKB248 covalently bond to Cys145 forming hemithioketal adducts, occupy S1' with a 4-fluorobenzothiazole moiety (fluorine pointing to solvent), and make key H-bonds in S1/S3 pockets; the central dimethyl-bicyclo[3.1.0]-proline fills S2. Nirmatrelvir lacks a P1' moiety, leaving S1' unoccupied.

The study addresses the urgent need for more potent SARS-CoV-2 antivirals by developing Mpro inhibitors with enhanced potency, breadth, and pharmacokinetics. TKB245, featuring a P1' 4-fluorobenzothiazole and a ketone warhead, demonstrates superior in vitro potency across multiple target cells and variants, surpassing nirmatrelvir, molnupiravir, and ensitrelvir. The compounds’ favorable PK and significant reductions in lung viral titers in hACE2 knock-in mice support their therapeutic potential. Mechanistically, promoting and stabilizing Mpro dimerization upon inhibitor binding, along with covalent engagement of Cys145 and effective occupation of the S1' subsite by the 4-fluorobenzothiazole, likely contribute to the high antiviral activity. The structural data underscore the importance of the P1' moiety—absent in nirmatrelvir—in fully engaging the S1' pocket. Fluorination appears to enhance lipophilicity and intracellular accumulation, aligning with the observed high intracellular concentrations and improved PK. Collectively, these findings validate the design strategy and highlight TKB245/248 as promising leads for COVID-19 therapeutics.

TKB245 and TKB248 are novel, orally available SARS-CoV-2 Mpro inhibitors with a P1' 4-fluorobenzothiazole moiety that show potent, broad antiviral activity against multiple variants, superior in vitro efficacy across diverse cell types, favorable pharmacokinetics, and significant in vivo reduction of viral titers in hACE2 knock-in mice. Structural analyses reveal covalent binding to Cys145 and effective occupancy of the S1' pocket, rationalizing their potency and suggesting an advantage over inhibitors lacking a P1' group. These compounds represent strong candidates for further development. Future work should include comprehensive safety/toxicology studies, oral dosing efficacy studies in multiple animal models, optimization of the warhead and P1' chemistry for potency and selectivity, evaluation of resistance potential, and clinical translation.

- Animal model constraints: hACE2 knock-in mice did not exhibit pronounced lung pathology at early time points, limiting assessment of disease-modifying effects beyond viral titer reductions. - Dosing route: In vivo efficacy was demonstrated with intraperitoneal dosing; oral efficacy in infected animals was not reported. - Mechanistic gaps: The mechanism behind TKB248’s markedly higher intracellular accumulation remains unclear. A direct comparative evaluation of warheads (ketone vs nitrile) was not completed. - Resistance assessment: While cross-variant activity was shown, systematic selection and characterization of resistance mutations were not performed. - Sample sizes and time points: Small group sizes (n=5) and early euthanasia (2–3 dpi) may limit generalizability of in vivo findings.