Rapid incorporation of Favipiravir by the fast and permissive viral RNA polymerase complex results in SARS-CoV-2 lethal mutagenesis

Coronaviruses are large positive-strand RNA viruses requiring rapid and accurate replication of ~30 kb genomes. The SARS-CoV-2 pandemic has driven efforts to develop antivirals targeting the viral RNA-dependent RNA polymerase (RdRp). Nucleoside analogues (NAs) are metabolized into active triphosphates and incorporated by error-prone viral RdRps, leading to chain termination or mutagenesis. Coronavirus replication is complicated by the proofreading exoribonuclease (ExoN, nsp14), which can excise certain NAs, potentially diminishing efficacy. Favipiravir (T-705), a broad-spectrum NA used against other RNA viruses, was hypothesized to inhibit SARS-CoV-2, but its efficacy and mechanism in coronaviruses were unclear. This study investigates whether favipiravir inhibits SARS-CoV-2 via incorporation by the RdRp and consequent lethal mutagenesis, and characterizes the kinetic properties and fidelity of the SARS-CoV/SARS-CoV-2 polymerase complex that might enable such action even in the presence of ExoN.

Prior work established that coronavirus nsp14 ExoN enhances replication fidelity and can remove some incorporated NAs, challenging NA-based therapies. Nonetheless, NAs including remdesivir and favipiravir have shown antiviral potential across RNA viruses. For favipiravir, studies in influenza, Coxsackie B3, and Ebola support a predominant mechanism of lethal mutagenesis with enhanced transition mutations; other reports suggested chain termination as an alternative mechanism, leaving its antiviral mode of action context-dependent and controversial. Comparative viral RdRp kinetics indicate most viral polymerases elongate at 4–20 s⁻¹ at 30–37 °C, implying that unusually rapid polymerases could trade fidelity for speed. This context frames the present examination of coronavirus RdRp speed, fidelity, and susceptibility to favipiravir-driven mutagenesis despite ExoN activity.

- Polymerase complex assembly and activity assays: Recombinant SARS-CoV/SARS-CoV-2 RdRp complexes were prepared with nsp12 and cofactors nsp7 and nsp8 (including an nsp7–nsp8 fusion to enhance activity). Activity was characterized using primer–template (PT) and self-priming hairpin (HP) RNA substrates. Processivity and elongation were assessed via steady-state time courses and gel-based product analysis. - Nucleoside analogue incorporation: Favipiravir (T-705) ribonucleotide triphosphate (RTP) and its non-fluorinated analogue T-1105-RTP were tested for incorporation by omitting ATP and/or GTP to force analogue usage. Reactions were run with defined analogue concentrations and monitored over time on PT and HP substrates to map incorporation events and elongation competence. - Kinetic analysis: Rapid quench-flow experiments (EDTA quench) quantified stepwise elongation rates across multiple positions, with fits to models yielding average incorporation rates and nucleotide concentration dependencies. Maximal elongation rates were derived from error-weighted fits across NTP titrations on HP and PT substrates. - Mismatch and analogue competition: In the absence of ATP, reactions containing GTP, UTP, and CTP quantified native GTP–U mismatch incorporation kinetics. Parallel reactions with added T-1105-RTP measured analogue incorporation relative to native mismatch rates to determine comparative efficiencies. - Cell culture infection and deep sequencing: Cells were infected with SARS-CoV-2 in the presence or absence of favipiravir (e.g., 500 µM). Viral RNA from supernatants was subjected to deep sequencing to quantify mutation frequencies and spectra across the genome. Antiviral activity was assessed by genome copy number, cytopathic effect (CPE), and infectious particle yield (e.g., TCID50). - Protein expression/purification and substrate preparation: nsp12, nsp7, and nsp8 were expressed in E. coli, purified via affinity chromatography, and assembled into active complexes. RNA substrates were synthesized, annealed, and validated. T-1105-RTP was synthesized and purified; T-705-RTP was obtained commercially. Compound stability was verified by spectroscopy and HPLC in assay buffers. - Data analysis: Band intensities were quantified from denaturing PAGE. Kinetic parameters (rates, burst amplitudes) were extracted from linear/exponential fits. Statistical analyses for mutation frequency employed Pearson’s χ² test with Yates’ continuity correction.

- Favipiravir induces lethal mutagenesis in SARS-CoV-2: Deep sequencing of viral populations grown with 500 µM favipiravir showed a 3-fold increase in total mutation frequency versus no-drug controls (p < 0.001, Pearson’s χ² with Yates’ correction). - Transition mutations increased: There was a 12-fold enrichment of G→A and C→U transitions, consistent with favipiravir acting predominantly as a guanosine-like purine analogue. This increased mutational burden correlated with reduced infectious particle yield. - Coronavirus RdRp speed and processivity: The nsp12/7/8 complex is highly active, displaying two modes of processivity—distributive behavior on PT substrates (with ~60% intermediates across timepoints) and dominant processive elongation on HP substrates. - Fastest viral RdRp measured: Quench-flow kinetics yielded maximal elongation rates of 150 ± 30 s⁻¹ (HP) and 91 ± 4 s⁻¹ (PT) under assay conditions, substantially exceeding typical viral RdRp rates (5–20 s⁻¹; HCV/dengue 4–18 s⁻¹). Extrapolation suggests 600–700 s⁻¹ at physiological temperatures. - Low fidelity and mismatch tolerance: The polymerase exhibited frequent misincorporations and read-through under conditions lacking cognate NTPs, indicating permissive active-site dynamics consistent with speed–fidelity tradeoffs. - Efficient analogue incorporation: Omission of ATP and/or GTP led to rapid, multiple incorporations of T-1105-RTP and T-705-RTP on both PT and HP substrates, with incorporation as purine analogues (opposite U and C), not substituting for pyrimidines. - T-1105 incorporation vs native mismatch: At 1 µM T-1105-RTP (with 50 µM GTP, UTP, CTP) analogue incorporation occurred on a similar timescale as native GTP–U mismatch and was quantified to be ~5-fold faster than the natural mismatch at 50 µM GTP over 0–60 s (from burst amplitude and linear rate analyses). - ExoN escape evidenced in cells: Despite coronavirus proofreading, favipiravir treatment in cell culture increased mutational diversity, implying that incorporated favipiravir can evade complete ExoN correction to drive population collapse.

The study addresses whether favipiravir can effectively target the coronavirus RdRp to inhibit SARS-CoV-2. Results show that the unusually fast and permissive SARS-CoV/SARS-CoV-2 RdRp readily incorporates favipiravir nucleotides, which then propagate C→U and G→A transition mutations throughout the genome. Infected cell populations displayed a significant rise in overall mutation frequency and specific transition biases alongside decreased infectious yield, supporting lethal mutagenesis as the predominant antiviral mechanism in the coronavirus context. The kinetic characterization reveals that coronavirus RdRp operates at rates exceeding those of other RNA viruses, with a concomitant reduction in fidelity and tolerance for mismatches and analogue incorporation. These properties help explain how favipiravir can be incorporated and retained even in the presence of the ExoN proofreading machinery, ultimately overwhelming viral population fitness. Collectively, the data position the RdRp as a critical vulnerability and support nucleoside analogue strategies for COVID-19 therapy.

This work demonstrates that favipiravir is efficiently incorporated by the fast, low-fidelity SARS-CoV/SARS-CoV-2 RdRp and predominantly inhibits SARS-CoV-2 through lethal mutagenesis, increasing C→U and G→A transitions and reducing infectivity. The polymerase exhibits the highest viral RdRp elongation rates reported, with distinct processivity modes and permissive mismatch incorporation, features that collectively enable nucleoside analogue action despite coronavirus proofreading. These findings validate the coronavirus RdRp as an Achilles heel and support development and optimization of nucleoside analogues for COVID-19. Future work should: (i) quantify in vivo pharmacodynamics and resistance barriers; (ii) define structural bases for analogue selectivity versus ExoN excision; (iii) optimize analogue chemotypes for improved incorporation and retention; and (iv) test rational combinations (e.g., with ExoN modulators) to enhance mutagenesis-driven antiviral efficacy.

- Most evidence derives from in vitro biochemical assays and cell culture infections; clinical efficacy and safety in humans were not addressed. - While mutagenesis is predominant, the precise contribution of alternative mechanisms (e.g., context-dependent chain termination) was not exhaustively dissected for coronaviruses. - Kinetic measurements were performed under defined assay conditions and extrapolated to physiological temperatures; in vivo polymerase dynamics may differ. - Deep sequencing captures population-level mutation spectra but may underrepresent rare events or context-dependent proofreading effects.