Analogs of the Catechol Derivative Dynasore Inhibit HIV-1 Ribonuclease H, SARS-CoV-2 nsp14 Exoribonuclease, and Virus Replication

Viral replication often depends on RNA maturation and degradation processes catalyzed by viral ribonucleases, which are therefore candidate targets for antiviral drugs. Here, we synthesized and studied the antiviral properties of a novel nitrocatechol compound (1c) and other analogs that are structurally related to the catechol derivative dynasore. Interestingly, compound 1c strongly inhibited two DEDD box viral ribonucleases, HIV-1 RNase H and SARS-CoV-2 nsp14 3′-to-5′ exoribonuclease (ExoN). While 1c inhibited SARS-CoV-2 ExoN activity, it did not interfere with the mRNA methyltransferase activity of nsp14. In silico molecular docking placed compound 1c in the catalytic pocket of the ExoN domain of nsp14. Finally, 1c inhibited SARS-CoV-2 replication but had no toxicity to human lung adenocarcinoma cells. Given its simple chemical synthesis from easily available starting materials, these results suggest that 1c might be a lead compound for the design of new antiviral compounds that target coronavirus nsp14 ExoN and other viral ribonucleases.


Plasmid Construction and Protein Purification
The genes of full-length nsp10 and nsp14 ExoN domain [1-289 amino acid (aa)] of SARS-CoV-2 (GenBank:NC_045512.2) were codon-optimized for E. coli expression and chemically synthesized (GenScript, Piscataway, NJ, USA). The synthesized nsp10 and nsp14 were fused to the N-terminal GST tag containing 3C protease cleavage site and the Nterminal 10His-SUMO containing TEV cleavage site, respectively. For nsp14 (ExoN)/nsp10 complex purification, two proteins were co-expressed in E. coli BL21(DE3) in 6 L of LB media supplemented with 50 µg/mL ampicillin and 25 µg/mL kanamycin, or 10H-SUMO-nsp14 (1-289aa) was singly expressed in E. coli BL21(DE3) in 6 L of LB media supplemented with 25 µg/mL kanamycin. Cultures were grown at 18 • C overnight after 0.5 mM IPTG induction. Cell pellets were resuspended using lysis buffer (250 mM NaCl and 20 mM HEPES pH 7.5) and disrupted using a microfluidizer (LM20, Microfluidics, Westwood, MA, USA). Lysates were cleared by centrifugation at 18,000 rpm, 4 • C for 60 min (Sorvall Lynx 4000) and cleared lysates were filtered using 0.45 µm membrane filter and incubated with glutathione resin (ThermoFisher, Rockford, IL, USA) for 2 h at 4 • C. The resin mixtures were loaded onto a gravity column and washed with lysis buffer, followed by elution with lysis buffer containing 10 mM glutathione. The fused tags were cleaved by TEV and 3C proteases during dialyzing in lysis buffer at 4 • C overnight, and the cleaved tags were then removed by a second round of glutathione resin purification, followed by a Histrap column (Cytiva, Marlborough, MA, USA) purification using ÄKTA Pure instrument. The fractions of nsp14-nsp10 complex were collected and further purified using a HiLoad 16/600 superdex 200 pg column (Cytiva, Marlborough, MA, USA) with lysis buffer.

SARS-CoV-2 nsp14 Exonuclease Activity Assays
Synthetic ssRNA substrate (RNA22, 5 -CGCAGUGAGCUCCUAAUCGCCC-3 ) was commercially obtained from IDT [17]. Final reaction mixtures contained 20 mM HEPES (pH 7.5), 5 mM DTT, 10 mM Mg 2+ , 10 µM RNA22 substrate, and 0.5 µM of nsp14-nsp10 complex in the absence or presence of inhibitors. The reactions were incubated at 30 • C for Viruses 2023, 15, 1539 6 of 18 5 min. Reactions were stopped by heating at 95 • C for 5 min. Samples were centrifuged at 18,000× g for 15 min at 4 • C. Supernatants were collected and analyzed by HPLC. The amount of RNA22 substrate was analyzed on a 1260 Infinity II Agilent Technologies HPLC instrument equipped with a Dionex DNAPac PA-100 (Thermo Scientific, Rockford, IL, USA). The processed samples (10 µL) were injected onto a Dionex DNAPac PA-100 analytical column at a flow rate of 1 mL/min and eluted with a linear gradient of 10 to 2000 mM NH 4 HCO 3 buffer (pH 7.8) for 90 min, followed by 30 min of equilibration to initial conditions. The HPLC column was maintained at 40 • C. Spectra were recorded at 256 nm. Open Lab CDS software was used to analyze and calculate the areas under the peaks in HPLC spectra. The area under the peak determined the amount of intact RNA22. IC 50 was determined by plotting inhibitor concentration against the amount of remaining intact substrate. The background was corrected by removing the smallest response and data were normalized by dividing with the largest response. The data were analyzed and plotted using GraphPad Prism (9.4.0) software.

SARS-CoV-2 nsp14 N7-Methyltransferase Activity Assays
A previously published protocol was modified to measure nsp14 MTase activity [51]. . Each wash lasted for 5 min with intermittent gentle rocking. After air drying for 15 min, the filter paper was cut and transferred to individual tubes containing beta plate scintillation fluid (PerkinElmer, Waltham, MA, USA). The amount of 3 H label bound was measured in counts per minute using a Wallac scintillation counter. Graphs were plotted in Graph PAD prism.

Drug Toxicity to Human Adenocarcinoma Lung Cells
A549 cells (ATCC) were treated with varying concentration of 1c ranging from 0.75 to 50 µM for 24, 48, 72, or 96 h in a 96-well plate. Percentages of live cells was determined and compared to DMSO (solvent) control using an Alamar-Blue-based fluorescence (Ex/Em 560/590 nm) assay as per the manufacturer's protocol (ThermoFisher Scientific, Rockford, USA). The amount of fluorescence produced is proportional to the number of living cells. The experiments were carried out in triplicate.

SARS-CoV-2 Replication Inhibition in A549 ACE2 + Cells
Approximately 80% confluent monolayer of A549 ACE2 + cells in 6-well plates were washed once with DMEM and infected with SARS-CoV-2 WA1/2020 strain at 0.05 MOI. The plates were placed on a rocker for 45 min at room temperature in compound-free conditions for virus adsorption. The solutions were then removed and replaced by DMEM containing 1c at various concentrations (12.5, 25, 50 µM) and incubated at 37 • C in 5% CO 2 for 24 or 48 h. As a control, the infected cells incubated in compounds-free medium were included throughout the experiment. The supernatants were harvested followed by titration for virus yield determinations by CPE assay in Vero E6 ACE2 cells. Vero E6 ACE2 cells were Viruses 2023, 15, 1539 7 of 18 cultured in 96-well flat-bottom microplates at a density of 1.5 × 10 4 cells per well for 24 h at 37 • C in 5% CO 2 atmosphere. The cells were washed once with serum-free DMEM and infected with 10-fold serially diluted virus samples harvested from A549 ACE2 + cells. The virus-infected plates were incubated at room temperature for 45 min for virus adsorption. The virus solutions were then replaced with DMEM and further incubated at 37 • C in 5% CO 2 atmosphere for 72 h. The virus titers were determined by the uptake and subsequent extraction of neutral red dye. Briefly, cells were incubated with 0.033% neutral red dye for 3 h at 37 • C. Free dye was washed from the wells, and dye that was taken up by cells was quantified using a microplate reader with absorbance recorded at 540 nm. Absorbance values were expressed as percentages of uninfected control cells, and TCID 50 (50% tissue culture infectious dose) was determined using Prism software (GraphPad, La Jolla, CA, USA). The experiments were performed twice in triplicate.

Plaque Reduction Assay
Plaque reduction assays were performed as described previously [52]. Briefly, confluent monolayers of Vero E6 ACE2 cells were cultured in a 6-well tissue culture plate and infected with approximately 200 pfu of SARS-CoV-2 WA1/2020 strain. After 45 min on a rocker, the viruses were removed and replaced by overlay medium (DMEM containing 1% low-melting agarose and test compounds at different concentrations). After incubating the cultures for 3 d, the monolayers were fixed with 4% (v/v) formaldehyde solution and the agarose plugs were removed and stained with 1% (w/v) crystal violet solution. The concentration required to deduce the EC 50 was calculated by regression analysis of the dose-response curves generated from these data [53]. The experiments were performed three times.

Molecular Docking
The 2-dimensional structure of compound 1c was optimized for 3D using discovery studio and the structure was minimized using CHARMM force field in Accelrys discovery studio pipeline as described earlier [54,55]. The energy-minimized structure was first used to perform a blind docking on the AutoDock Vina program followed by site-specific docking using Glide (Schrodinger), where grids were restricted in the 10 best scoring binding poses. The lowest energy binding pose were selected based on free energy of binding (G binding ). The structures were analyzed and visualized in pymol and UCSF Chimera 1.8.
For reasons of comparison, the new compounds 2a-c were prepared as close analogs of 1c by reaction of 3,4-dihydroxy-5-nitrobenzaldehyde (obtained from 3-nitrovanillin refluxed in 48% HBr) and the corresponding acid hydrazide (Scheme 2). For reasons of comparison, the new compounds 2a-c were prepared as close analogs of 1c by reaction of 3,4-dihydroxy-5-nitrobenzaldehyde (obtained from 3-nitrovanillin refluxed in 48% HBr) and the corresponding acid hydrazide (Scheme 2).

Inhibition of HIV-1 RNase H
Initially, dynasore and compounds 1a-l and 2a-c were evaluated for their inhibition of HIV-1 RNase H ( Table 1). The HIV-1 RT-associated RNase H inhibition assay is wellestablished and suitable for the identification of hit compounds from compound series and libraries [49,50]. As expected, dynasore exhibited a considerable inhibition of HIV-1 RNase H, which was higher than the inhibition by the positive control RDS1643 [50]; however, addition of a nitro group to the dynasore structure in compound 1c led to the strongest HIV-1 RNase H inhibitor of this series of test compounds, with an IC50 value of 0.57 µM. In terms of structure-activity relationships, it became clear that a catechol scaffold is necessary for a high HIV-1 RNase H inhibition, as evidenced by dynasore, 1c, 1e, and 1f. The 2,3-dihydroxy-4-methoxyphenyl derivative 1e and its 2,3,4- NMR, IR, and MS methods (see Supplementary Materials). Compounds 1d and 1f-k are known compounds, and analytical data were consistent with published data of these compounds [43][44][45][46][47].

Inhibition of HIV-1 RNase H
Initially, dynasore and compounds 1a-l and 2a-c were evaluated for their inhibition of HIV-1 RNase H ( Table 1). The HIV-1 RT-associated RNase H inhibition assay is wellestablished and suitable for the identification of hit compounds from compound series and libraries [49,50]. As expected, dynasore exhibited a considerable inhibition of HIV-1 RNase H, which was higher than the inhibition by the positive control RDS1643 [50]; however, addition of a nitro group to the dynasore structure in compound 1c led to the strongest HIV-1 RNase H inhibitor of this series of test compounds, with an IC50 value of 0.57 µM. In terms of structure-activity relationships, it became clear that a catechol scaffold is necessary for a high HIV-1 RNase H inhibition, as evidenced by dynasore, 1c, 1e, and 1f. The 2,3-dihydroxy-4-methoxyphenyl derivative 1e and its 2,3,4-Scheme 2. Reagents and conditions: (i) Acid hydrazide, EtOH, reflux, 2 h, 64-86%.

Inhibition of HIV-1 RNase H
Initially, dynasore and compounds 1a-l and 2a-c were evaluated for their inhibition of HIV-1 RNase H ( Table 1). The HIV-1 RT-associated RNase H inhibition assay is wellestablished and suitable for the identification of hit compounds from compound series and libraries [49,50]. As expected, dynasore exhibited a considerable inhibition of HIV-1 RNase H, which was higher than the inhibition by the positive control RDS1643 [50]; however, addition of a nitro group to the dynasore structure in compound 1c led to the strongest HIV-1 RNase H inhibitor of this series of test compounds, with an IC 50 value of 0.57 µM. In terms of structure-activity relationships, it became clear that a catechol scaffold is necessary for a high HIV-1 RNase H inhibition, as evidenced by dynasore, 1c, 1e, and 1f. The 2,3-dihydroxy-4-methoxyphenyl derivative 1e and its 2,3,4-trihydroxyphenyl analog 1f were slightly more active than the 3,4-dihydroxyphenyl parent compound dynasore, but less active than 1c. Among the non-catechol derivatives, only 3-nitrovanillyl 1b and 4-N-diethylaminophenyl 1k revealed activities below 10 µM (IC 50 = 7-8 µM). The latter compound was distinctly more active than its 4-N-dimethylamino analog 1i, indicating an activity increasing effect by longer alkyl groups, while the corresponding acetamido and N-methylpiperazinyl derivatives 1j and 1l were inactive. Compounds 2a-c, which are nitrocatechol analogs of 1c differing in their acyl moiety, were much less active than 1c.
This underlines the importance of the 3-hydroxy-2-naphthoic acyl fragment in 1c for HIV-1 RNase H inhibition.

Inhibition of ExoN Activity of SARS-CoV-2 nsp14/nsp10
Based on its high HIV-1 RNase H inhibitory activity, compound 1c was selected for SARS-CoV-2 ExoN exoribonuclease inhibition studies. Nsp10 interacts with the ExoN domain of nsp14 to stabilize it and stimulates its nuclease activity [23,56]. Degradation of the ssRNA substrate by nsp14/nsp10 complex was monitored using an HPLC-based assay in the absence or presence of the inhibitor. SARS-CoV-2 nsp14/nsp10 ExoN activity was strongly inhibited by 1c with an IC 50 value of 2.2 µM (Figure 1). Thus, the nsp14 inhibitory activity of 1c lies in the range of the recently disclosed nsp14/nsp10 inhibitors patulin and ebselen [18,19]. The 1c analogs 2a, 2b, and 2c, were also tested to judge the importance of the 3-hydroxy-2-naphthoic acid scaffold for nsp14 inhibitory activity. But compounds 2a-c had reduced inhibitory activities compared with 1c ( Figure 1). There was complete inhibition of the ExoN activity by 1c at 100 µM, whereas 2a, 2b, and 2c inhibited ExoN activity by about 65%, 53%, and 37%, respectively ( Figure 1).

Inhibitor 1c Does Not Interfere with nsp14 MTase Activity
Nsp14 is a bifunctional enzyme with ExoN (N-terminal) and N7-MTase (C-terminal) activities. The MTase activity is required for viral RNA capping, mRNA translation, and stabilization [51]. We tested the possibility of dual-function targeting by inhibitor 1c. Full-

Inhibitor 1c Does Not Interfere with nsp14 MTase Activity
Nsp14 is a bifunctional enzyme with ExoN (N-terminal) and N7-MTase (C-terminal) activities. The MTase activity is required for viral RNA capping, mRNA translation, and stabilization [51]. We tested the possibility of dual-function targeting by inhibitor 1c. Full-length nsp14 MTase activity was assessed in a biochemical assay using synthetic GpppA mRNA cap substrate in the absence or presence of the inhibitor (Figure 2). Nsp14 methylates GpppA to m7GpppA by transferring the [3H]-CH 3 moiety from [3H]-S-adenosyl methionine (SAM). Tritium incorporation in the methylated product was quantified using a DEAE-cellulose filter binding assay, followed by liquid scintillation counting. In positive control reactions, the recombinant purified nsp14 protein efficiently methylated the GpppA substrate in a concentration-dependent manner (Figure 2a). As negative controls, m 7 GpppA was not methylated, as determined by a background signal similar to incubations lacking nsp14. We observed no significant differences in the MTase activity as measured by tritium incorporation in the final product in the absence and presence of inhibitor 1c (Figure 2b). Results indicate that 1c did not interfere with the MTase activity of the nsp14 protein; therefore, it inhibits only its ExoN function. Moreover, inhibition of the nsp14 ExoN activity (due to 1c binding) does not affect MTase domain function. activity by the analogs is shown at 100 µM concentration, as indicated on the graph. RNA22 substrate incubated under the same conditions (in the absence of both nsp14/10 and inhibitors) served as un-degraded control (100% intact substrate). Experiments were performed in triplicate. Error bars show standard deviations.

Inhibitor 1c Does Not Interfere with nsp14 MTase Activity
Nsp14 is a bifunctional enzyme with ExoN (N-terminal) and N7-MTase (C-terminal) activities. The MTase activity is required for viral RNA capping, mRNA translation, and stabilization [51]. We tested the possibility of dual-function targeting by inhibitor 1c. Fulllength nsp14 MTase activity was assessed in a biochemical assay using synthetic GpppA mRNA cap substrate in the absence or presence of the inhibitor (Figure 2). Nsp14 methylates GpppA to m7GpppA by transferring the [3H]-CH3 moiety from [3H]-Sadenosyl methionine (SAM). Tritium incorporation in the methylated product was quantified using a DEAE-cellulose filter binding assay, followed by liquid scintillation counting. In positive control reactions, the recombinant purified nsp14 protein efficiently methylated the GpppA substrate in a concentration-dependent manner (Figure 2a). As negative controls, m 7 GpppA was not methylated, as determined by a background signal similar to incubations lacking nsp14. We observed no significant differences in the MTase activity as measured by tritium incorporation in the final product in the absence and presence of inhibitor 1c (Figure 2b). Results indicate that 1c did not interfere with the MTase activity of the nsp14 protein; therefore, it inhibits only its ExoN function. Moreover, inhibition of the nsp14 ExoN activity (due to 1c binding) does not affect MTase domain function.

Inhibitor 1c Docks in the Catalytic Site of nsp14-nsp10 Complex
The heterodimerization of nsp14 with nsp10 is a prerequisite for ExoN activity [56]. Therefore, to probe the molecular details of inhibitor 1c interaction with nsp14, we utilized the high-resolution Cryo-EM structure of SARS-CoV-2 nsp10-nsp14 (WT)-RNA complex (PDB ID: 7N0B) for molecular docking simulation using Glide™ [57]. Initial docking simulations were performed using a large enough grid to include the entire complex. The best-scoring binding pose among the 10 top poses with RMSD of 2.5 Å preferentially occupies the catalytic site on the RNA-protein binding interface (Figure 3a-c). The in silico binding energy was calculated to be −7.2 kcal/mol. Inhibitor 1c appears to bind to the EDD catalytic site residues, which are critical for substrate interaction [57]. The meta-hydroxyl group forms a hydrogen bond with aspartic acid 90, and glutamic acid 191 is in close proximity of the nitro group forming a salt bridge.
best-scoring binding pose among the 10 top poses with RMSD of 2.5 Å preferentia occupies the catalytic site on the RNA-protein binding interface (Figure 3a-c). The in sil binding energy was calculated to be −7.2 kcal/mol. Inhibitor 1c appears to bind to the ED catalytic site residues, which are critical for substrate interaction [57]. The meta-hydro group forms a hydrogen bond with aspartic acid 90, and glutamic acid 191 is in clo proximity of the nitro group forming a salt bridge.

Compound 1c Does Not Affect Cell Viability
To investigate whether 1c might have drug-like properties suitable for controlli SARS-CoV-2 infections, its toxicity to human lung cancer cells was investigated. On nontoxic compounds are suitable for antiviral drug development. Human A549 NSC Residues are depicted as sticks, while 1c is depicted using ball-and-stick modeling with elemental color representations. Mg 2+ ions are shown in green and the dotted line shows the possible interaction with Mg 2+ . Amino acid residues, as sticks, and inhibitor 1c, in ball and sticks, are depicted using elemental colors.

Compound 1c Does Not Affect Cell Viability
To investigate whether 1c might have drug-like properties suitable for controlling SARS-CoV-2 infections, its toxicity to human lung cancer cells was investigated. Only nontoxic compounds are suitable for antiviral drug development. Human A549 NSCLC (non-small-cell lung cancer) cells were applied for toxicity testing with 1c ( Figure 4). Compound 1c showed no toxicity to A549 cells even at concentrations of 50 µM for four days. (non-small-cell lung cancer) cells were applied for toxicity testing with 1c ( Figure 4) Compound 1c showed no toxicity to A549 cells even at concentrations of 50 µM for four days.

Compound 1c Inhibits SARS-CoV-2 Replication
Next, the antiviral activity of compound 1c was evaluated using A549 ACE2 + cells infected with SARS-CoV-2 ( (non-small-cell lung cancer) cells were applied for toxicity testing with 1c ( Figure 4). Compound 1c showed no toxicity to A549 cells even at concentrations of 50 µM for four days.
To quantify the antiviral activities of 1c and 2a-c as EC 50 values, plaque reduction assays were carried out ( Figure 6, Table 2). The most active antiviral compound was 1c with an EC 50 value of 10.2 µM (Table 2). Analogs 2a and 2b showed weaker activity, while nicotinoyl derivative 2c was inactive at concentrations of up to 50 µM (Table 2). EC 50 of analogs 2a and 2b were 4.3-and 3.4-fold times higher compared to 1c, respectively. The observed tendency of antiviral activity among compounds 1c and 2a-c is consistent with their nsp14/nsp10 ExoN inhibitory activity. These findings support inhibition of ExoN as the molecular mechanism for the antiviral activity of 1c. About fivefold higher doses of 1c were required for the inhibition of SARS-CoV-2 replication than for inhibition of ExoN nsp14/nsp10 activity in vitro, perhaps due to reduced cellular uptake of the drug. To quantify the antiviral activities of 1c and 2a-c as EC50 values, plaque reduction assays were carried out ( Figure 6, Table 2). The most active antiviral compound was 1c with an EC50 value of 10.2 µM (Table 2). Analogs 2a and 2b showed weaker activity, while nicotinoyl derivative 2c was inactive at concentrations of up to 50 µM ( Table 2). EC50 of analogs 2a and 2b were 4.3-and 3.4-fold times higher compared to 1c, respectively. The observed tendency of antiviral activity among compounds 1c and 2a-c is consistent with their nsp14/nsp10 ExoN inhibitory activity. These findings support inhibition of ExoN as the molecular mechanism for the antiviral activity of 1c. About fivefold higher doses of 1c were required for the inhibition of SARS-CoV-2 replication than for inhibition of ExoN nsp14/nsp10 activity in vitro, perhaps due to reduced cellular uptake of the drug.

Discussion
In this study, the nontoxic dynasore analog 1c was identified as a new dual HIV-1 RNase H and SARS-CoV-2 nsp14/nsp10 exoribonuclease inhibitor. In addition, 1c blocked SARS-CoV-2 replication in infected cells; thus, it is a promising lead antiviral drug candidate suitable for further development. The DEDD motif ribonucleases HIV-1 RNase H and SARS-CoV-2 ExoN with Mg 2+ -coordinated active sites are both targets of 1c. Further efforts towards the design of improved dual HIV-1 RNase H and SARS-CoV-2 nsp14/nsp10 inhibitors are currently underway. Notably, the non-nitro derivatives 1e and 1f with substituted catechol ring systems also displayed reasonable HIV-1 RNase H inhibitory activities, and can be considered for drug optimization studies to tackle SARS-CoV-2 infections more efficiently. Results obtained from MTase inhibition assays support the conclusion that ExoN and MTase domains are functionally independent, consistent with prior findings on SARS-CoV-2 and SARS-CoV nsp14 [22,51,58].
Thus, 1c adds to the current arsenal of clinically applied acyl-hydrazone-based drugs and anti-infectives [68,69]. Future synthetic efforts will deal with the replacement of the hydrazone core fragment by other molecular systems, which cannot undergo hydrolysis but conserve the activity of 1c. In addition, prodrug strategies to mask the aromatic hydroxyl groups (e.g., by esterification) of 1c might be considered if necessary.
The established ribonuclease inhibition suggests a vital role of the nitrocatechol moiety for the activity of 1c; however, the 3-hydroxy-2-naphthoic component of 1c also contributed significantly to viral nuclease inhibition. A conceivable interaction of the nitro-substituted catechol with the nuclease active site was corroborated by nsp14-nsp10 docking experiments. In addition, this scaffold is a crucial component of salient human catechol-Omethyltransferase (hCOMT) inhibitors such as entacapone, which are clinically applied for the treatment of Parkinson's disease [70]. Available pharmacokinetic data of these hCOMT inhibitors are expected to be relevant and useful for the further development of 1c, or a derivative of 1c, as an antiviral drug.