Development of Selective NS5 Inhibitors for the Broad-Spectrum Treatment of Flaviviruses

Development of Selective NS5 Inhibitors for the Broad-Spectrum Treatment of Flaviviruses is sponsored by NIH. EydisBio, Inc. received an SBIR Phase I award to develop a novel small-molecule inhibitor targeting the highly conserved NS5 protein across flaviviruses.

## [](https://pmc. ncbi. nlm.

nih. gov/articles/PMC10436986/)Abstract Flavivirus infections, such as those caused by dengue virus (DENV), West Nile virus (WNV), yellow fever virus (YFV), and Zika virus (ZIKV), pose a rising threat to global health. There are no FDA-approved drugs for flaviviruses, although a small number of flaviviruses have vaccines.

For flaviviruses or unknown viruses that may appear in the future, it is particularly desirable to identify broad-spectrum inhibitors. The NS5 protein is regarded as one of the most promising flavivirus drug targets because it is conserved across flaviviruses.

In this study, we used FL-NAH, a fluorescent analog of the methyl donor S-adenosylmethionine (SAM), to develop a fluorescence polarization (FP)-based high throughput screening (HTS) assay to specifically target methyltransferase (MTase), a vital enzyme for flaviviruses that methylates the N7 and 2’-O positions of the viral 5’-RNA cap. Pilot screening identified two candidate MTase inhibitors, NSC 111552 and 288387.

The two compounds inhibited the FL-NAH binding to the DENV3 MTase with low micromolar $\text{IC}_{50}$. Functional assays verified the inhibitory potency of these molecules for the flavivirus MTase activity. Binding studies indicated that these molecules are bound directly to the DENV3 MTase with similar low micromolar affinity.

Furthermore, we showed that these compounds greatly reduced ZIKV replication in cell-based experiments at dosages that did not cause cytotoxicity. Finally, docking studies revealed that these molecules bind to the SAM-binding region on the DENV3 MTase; and further mutagenesis studies verified residues important for the binding of these compounds.

Overall, these compounds are innovative and attractive candidates for the development of broad-spectrum inhibitors for treatment of flavivirus infections. **Keywords:**Flaviviruses, NS5, methyltransferase, high throughput screening, broad-spectrum ## [](https://pmc. ncbi.

nlm. nih. gov/articles/PMC10436986/)Results ### Binding of FL-NAH to the flavivirus MTase To find and characterize inhibitors that target the cofactor SAM-binding site of the flavivirus MTases, we developed an FP-based assay using FL-NAH, a fluorescent SAM analog (Figure 1A).

The non-hydrolyzable fluorescent SAM analog, FL-NAH, can bind the SAM-dependent MTases by mimicking SAM. 57 To develop a SAM-displacement assay for the flavivirus MTases, we expressed and purified DENV3, YFV, and WNV MTases as reported previously. 24, 53, 58 We also cloned, expressed, and purified the ZIKV MTase similarly as other MTases.

We then performed an FP assay to monitor FL-NAH binding to these representative flaviviral MTases. Our result indicated that significant FP increases were observed for FL-NAH binding to flaviviral MTases with increasing MTase concentrations, suggesting that FL-NAH binds to these flaviviral MTases dose-dependently (Figure 1B).

By fitting an experimental curve, we determined the binding affinity $\left(K_{D}\right)$ of FL-NAH to these MTases (Table 1). FL-NAH binds the WNV and DENV3 MTases with high affinity of 0. 28 μM and 0.

61 μM, respectively. The YFV and ZIKV MTases bind FL-NAH with moderate affinity of 3 μM and 3. 5 μM, respectively.

### Inhibition of FL-NAH binding to the DENV3 MTase by SAM, SAH and known inhibitor Given that FL-NAH is an analog of SAM, it is anticipated that it will bind to the SAM-binding site of the DENV3 MTase. To confirm this, we investigated if SAM and its byproduct SAH could prevent FL-NAH from binding to the DENV3 MTase. We performed the FL-NAH FP assay in the absence or presence of concentration series of SAM or SAH.

Our findings demonstrated that SAM and SAH dose-dependently reduced FL-NAH binding to the DENV3 NS5 MTase (Figure 1C). The efficacy of SAM and SAH in inhibition of FL-NAH binding to the DENV3 MTase was quantitatively evaluated by $I C_{50 \mathit{-disp}}$, defined as the compound concentration at which FL-NAH binding to the DENV3 MTase was reduced to 50%.

The $I C_{50 \mathit{-disp}}$ values for inhibition of FL-NAH binding to the DENV3 MTase by SAM and SAH were determined as 4. 8 μM and 0. 8 μM, respectively.

It is interesting to note that SAH is more efficient in inhibition of FL-NAH binding to the DENV3 MTase than SAM. Similar finding was previously reported for other MTase. 57 Using virtual screening, we previously identified NSC12155 as a potent broad-spectrum inhibitor of flavivirus MTases.

53 Crystal structure of the NSC12155-MTase complex indicated that NSC12155 bound to the SAM-binding site of the DENV3 MTase. 53 Therefore, we used the SAM-binding site inhibitor NSC12155 to further validate the assay. Our results showed that NSC12155 dose-dependently reduced the FL-NAH binding to the DENV3 MTase with $I C_{50 \mathit{-disp}}$ of 5.

4 μM (Figure 1C). Overall, our results suggest that FL-NAH binds to the SAM-binding site of the DENV3 MTase and that its binding can be ablated by the cofactor SAM, the byproduct SAH, and small molecule inhibitor. ### FP-based FL-NAH-displacement HT screening Using SAH (25 μM) as a control inhibitor, we evaluated the suitability of the FL-NAH-displacement assay for HTS in a 96-well plate format.

Our findings demonstrated that the FL-NAH-displacement FP assay is well suitable for HTS, with a satisfactory Z’-factor of 0. 87, high signal/background (S/B) ratio of 7. 1, and low coefficient of variation (CV) of 2.

1 (Figure 1D). These parameters satisfy or are better than the guideline criteria. 59 To find compounds that target the SAM-binding site of the DENV NS5 MTase, we carried out a small-scale HTS against the NCI diversity set VI compound library, composed of 1,584 small molecules.

For HTS, each compound was applied at a single concentration of 15 μM. We used SAH at a concentration of 25 μM as an inhibitory positive control, and DMSO as a negative control for each plate. Z’-factor, S/B ratio, and CV for each plate were calculated to determine the quality of the screening assay, with the average values of 0.

79, 3. 96, and 3. 1%, respectively (Figure 2A).

The outcomes suggested a screen of excellent quality. Primary HTS screen identified 20 compounds showing inhibition more than 60% at 15 μM (Figure 2B). Eight compounds were discarded due to compound autofluorescence, fluorescence quenching, or bad chemical properties predicted from cheminformatics analysis.

Twelve compounds were re-ordered and subjected to dose-response inhibition of FL-NAH binding to the DENV3 MTase (Figure 2B). Two compounds, including NSC 111552 and 288387 (Figure 2C), showed dose-dependent inhibition of FL-NAH binding to the DENV3 NS5 MTase with $I C_{50 \mathit{-disp}}$ values of 0. 98 μM and 4.

2 μM, respectively (Figure 2D, Table 2). Because FL-NAH is an SAM-analog, inhibitors replacing FL-NAH are expected to target the SAM-binding site that is revolutionarily conserved among flavivirus MTases. 20, 60 We hypothesize that the SAM-binding site inhibitors are broad-spectrum pan-flavivirus inhibitors.

To investigate this, we tested additional flaviviral MTases. Our results showed that both compounds NSC 111552 and 288387 dose-dependently inhibited the binding of FL-NAH to the MTases of ZIKV, WNV, and YFV (Figure S1 and Table 2). In general, the $I C_{50 \mathit{-disp}}$ potency values of these two compounds against the MTases of ZIKV, WNV and YFV were higher than those against the DENV3 MTase.

Specifically, NSC111552 was 6, 7 and 46 times less potent against the MTases of WNV, ZIKV and YFV than it was against the DENV3 MTase, respectively. NSC288387 had similar potency against the MTases of DENV3 and ZIKV and was 2–3 times less potent for the MTases of YFV and WNV. Overall, the results confirmed that NSC 111552 and 288387 are pan-flavivirus MTase inhibitors.

### The compound liability To address possible pan-assay interference compounds (PAINS) properties of the hits, compounds NSC 111552 and 288387 were screened for the PAINS properties using two _in-silico_ methods. First, chemical structures were run using the FAF4-Drugs script _via_ a web portal of the Parisian Resource in Structural Bioinformatics. 61 Both compounds were screened using the three available PAINS filter A, B & C.

In each case, compounds NSC 111552 and 288387 passed. Next, structures of NSC 111552 and 288387 were screened through the PAINS-Remover program. 62 Our results showed that both compounds NSC 111552 and 288387 passed _in-silico_ screening in PAINS-Remover.

The two compounds identified through HTS appear to be redox cyclers that generate hydrogen peroxide (H 2 O 2) in the presence of strong reducing agents such as dithiothreitol (DTT) used in nearly all of our biochemical assays. For redox cycling compounds, the observed inhibitory activity could be resulted from the promiscuous artifacts by generated H 2 O 2.

63 To rule out this possibility, we repeated the FP assay for these two compounds in the presence and absence of DTT (Figure 2D and Table 2). Our findings indicated that NSC111552 had similar potency in inhibition of FL-NAH binding to the DENV3 MTase with or without DTT (Table 2).

DTT had a slight effect on the activity of NSC288387, reducing its potency by about threefold in inhibiting FL-NAH binding to the DENV3 MTase in the presence of 1 mM DTT compared to its potency in the absence of DTT.

The results indicated that redox cycling potential of NSC288387 did not affect NSC288387’s inhibitory activity, as the addition of DTT did not lead to a lower $I C_{50 \mathit{-disp}}$ value compared to that without DTT, which would have indicated an opposite result.

Overall, the results suggest that the redox property of the compounds does not affect the inhibitory activity of these compounds on FL-NAH binding to the DENV3 MTase in the assay (Figure 2D and Table 2). ### NSC 111552 and 288387 are specific to viral MTases Inhibitors targeting the SAM-binding site of viral MTase could also affect human MTases, as SAM is the common methyl donor for nearly all MTases.

To further investigate the specificity of these compounds, we tested the NSC 111552 and 288387 compounds against a representative Human RNA MTase (hRNMT), as we did previously. 64 Our result showed that the $I C_{50 \mathit{-disp}}$ values for inhibition of hRNMT by NSC 111552 and 288387 were 63. 3 μM and 35.

5 μM, respectively (Figure 2E and Table 2). These values are significantly higher than those for viral MTases of DENV3, WNV, and ZIKV, and also slightly higher than that for the YFV MTase. The results suggest that NSC 111552 and 288387 exhibit varying selectivity towards viral and human MTases.

NSC111552 shows greater specificity towards the MTases of DENV3, ZIKV, and WNV compared to hRNMT. However, it lacks selectivity against the YFV MTase and hRNMT. On the other hand, NSC288387 displays higher specificity for the MTases of DENV3 and ZIKV compared to hRNMT, although the selectivity between the viral MTases (WNV and YFV) and hRNMT is comparatively lower.

### Inhibition of the N7 MTase activity Effectiveness of the candidate inhibitors in reducing flaviviral MTase activity was further assessed utilizing a functional N7 MTase activity assay. We first generated capped viral RNA representing the first 90 nts of authentic WNV genome as we described previously. 53 The capped WNV RNA could serve as a substrate for MTases of representative flaviviruses including DENV2, DENV3, YFV, and WNV.

23, 53, 58 We then used a commercially available EPIgenous MTase kit (PerkinElmer) based on homologous time resolved fluorescence (HTRF) to quantify released SAH from the MTase enzymatic reaction. An anti-SAH antibody coupled to Tb cryptate and SAH tagged with d2 fluorophore (SAH-d2) as a tracer is used in the HTRF experiment.

In the absence of SAH, the anti-SAH Tb cryptate forms a complex with the SAH-d2 fluorophore to trigger HTRF, resulting in high HTRF. An MTase reaction will result in the production of SAH, which will compete with and displace SAH-d2 from the Tb-cryptate anti-SAH antibody, leading to reduction of the HTRF signal. We first investigated if the HRTF assay could be used to detect the SAH release from flaviviral MTase reaction.

As shown in Figure 3A, our results showed that reaction mixture in the absence of MTase generated robust HTRF signal, whereas addition of the DENV3 MTase (0. 6 μM) significantly reduced the HTRF signal.

The results suggested that the DENV3 MTase successfully utilized SAM to methylate the viral Gppp-RNA, leading to generation of m 7 Gppp-RNA and byproduct SAH, the latter of which displaced the SAH-d2 from the HTRF pair, leading to reduced HTRF signal (Figure 3A). Using this HTRF function assay, we measured the inhibitory efficacy of hit compounds in inhibition of the methyl transfer activity of viral MTases of DENV3, ZIKV, WNV, and YFV.

As shown in Figure 3B, compounds NSC 111552 and 288387 dose-dependently inhibited the viral MTase activity with $I C_{50 \mathit{-HTRF}}$ values in low micromolar range (Table 3). In addition, by using a 32 P-labeled substrate, we measured the N7 MTase activities of the WNV MTase in the presence of NSC 111552 and 288387.

This experiment allowed us to directly visualize and quantify N7-methylation of the RNA guanine cap using traditional thin layer chromatography (TLC) method, as we described previously. 24, 53 We first _in vitro_ generated RNAs representing the first 90 nucleotides of the WNV genome, both unmethylated G*pppA-RNA and N7-methylated m7G*pppA-RNA where the asterisk indicated that the following phosphate was 32 P labeled.

58, 64 As shown in Figure 3C, the WNV NS5 MTase effectively N7-methylated the G*pppA-RNA to produce the cap-0 structure (m 7 G*pppA). The conversion of G*pppA-RNA to m 7 G*pppA-RNA by the WNV MTase was dose-dependently inhibited by concentration series of NSC 111552 and 288387. We determined the $\mathit{IC}_{\mathit{50-TLC}}$ values of 1.

1 μM and 1. 6 μM in inhibition of the WNV MTase for NSC 111552 and 288387, respectively (Table 3). Overall, the above experiments clearly show that compounds NSC 111552 and 288387 strongly inhibit the N7 MTase activity of representative flaviviruses.

### Cytotoxicity and Antiviral Analyses. To further characterize these candidate inhibitors, we investigated the cytotoxicity of NSC 111552 and 288387 to human A549 lung carcinoma cells, using a WST-8 cell viability kit, as we described previously.

65, 66 Our results showed that NSC288387 was moderately toxic to the A549 cells with cytotoxicity $\text{CC}_{50}$ of 57 μM, whereas NSC111552 was not toxic to the A549 cells until very high concentration, with $\text{CC}_{50}$ estimated as 187 μM (Figure 4A, Table 3). Next, we carried out a cell-based replicon study to investigate if NSC 111552 and 288387 could lower viral replication in cell culture.

Our results showed that NSC 111552 and 288387 dose-dependently inhibited DENV2 replication using a replicon (Replic) cell line of DENV serotype 2 (DENV2) (Figure 4B, Table 3). The $\text{EC}_{50}$ values for NSC 111552 and 288387 for DENV2 were found to be 5. 0 μM and 11.

4 μM, respectively (Figure 4B, Table 3). We further carried out an antiviral immunofluorescence assay (IFA) to investigate the antiviral efficacy of these two inhibitors against DENV2 and ZIKV. We showed that NSC 111552 and 288387 were potent inhibitors against both DENV2 and ZIKV (Figure 4C, Table 3).

We next carried out an antiviral plaque reduction assay (PRA) to investigate the antiviral efficacy of these two inhibitors against ZIKV. By using a Venus-expressing ZIKV we generated previously 67, we showed that NSC 111552 and 288387 were potent inhibitors against ZIKV, with $\text{EC}_{50}$ values of 1. 4 μM and 0.

2 μM respectively (Figure 4D, Table 3). ### Direct binding of NSC 111552 and 288387 to the DENV3 NS5 MTase By using microscale thermophoresis (MST), we investigated if NSC 111552 and 288387 directly bound to the DENV3 NS5 MTase protein. Our results showed that NSC 111552 and 288387 dose-dependently bound to the DENV3 NS5 MTase with a binding affinity $\left(K_{D}\right)$ of 4.

6 μM and 3. 2 μM, respectively (Table 3). These data confirmed direct binding of NSC 111552 and 288387 to the DENV3 NS5 MTase (Figure 5).

To gain insight into the binding modality and possible mechanism of enzyme inhibition, _in-silico_ docking of both identified leads was performed using Schrodinger Maestro software and crystal structure of DENV3 NS5 MTase bound to the substrate S-Adenosyl methionine (PDB ID: 3P97). 43 The inhibitors NSC 288387 and 111552 were observed to dock consistently at the SAM binding pocket.

The primary interactions observed were consistent with residues common to SAM binding with both inhibitors interacting with Val132, Lys105, Thr104 and His110. Docking was validated by overlay of native structure and ligand with docked SAM to observe fit accuracy. Sinefungin, a SAM analog, and SAH, the enzymatic byproduct, were also docked as reference inhibitors (Figure S2).

Inhibitor NSC111552 interacted with more SAM binding residues at higher frequency and at greater consistency compared to NSC288387. Notably, docking of NSC111552 predicted extensive pi-cation interactions with Lys105 and Pi-Pi stacking interactions with His110 (Figure 6A). Multiple poses of NSC111552 displayed frequent hydrogen bonding with Lys105 through its carbonyl oxygen.

NSC111552’s hydroxyl groups interacted through hydrogen bonding with His110, Asp131 and Lys105 with a deprotonated hydroxyl group likely forming a salt bridge with Lys180.

The hydrogen bonding of the hydroxyl groups often occurred via protein backbone interactions at residues Asp131 and Val132 as observed in top MM/GBSA scoring poses with similar interactions occurring at residues Thr104 and Lys130 in less frequent and lower scoring poses. Significantly, NSC111552 interacted with an Asp131 in addition to the 4 residues both inhibitors have in common with SAM.

Lastly, NSC111552 interactions with these critical residues are of greater projected frequency and consistency. The docking of NSC288387 resulted in considerably less frequent consistent interactions. The increased aromaticity of NSC288387 did not get greater Pi-Pi stacking or Pi-cation interactions when compared to NSC111552.

Filtering by MM/GBSA scoring, favorable poses showed the carbonyl oxygen hydrogen bonding between Glu149, Lys105, Thr104 and Gly148 alternately in posing. NSC288387 also interacted with Lys105 via pi-cation and hydrogen bond formation at protein backbone.

Notably, NSC288387 had greater frequency of Pi-Pi stacking interactions with residue Phe133 and a unique pi-cation interaction with Arg160 and Lys180 distinct from the interactions observed with inhibitor NSC111552. While NSC288387 did interact with SAM binding residues Val132, His110, Lys105 and Thr104, these were infrequent relative to 111552.

NSC288387 had a superior average MM/GBSA scoring relative to inhibitor NSC111552 by factor of −20 Kcal/mol which is a function of the substantially lower energetic favorability of the deprotonated NSC111552 docking in the binding site. It should be noted that the MM/GBSA score is a calculation based on end-point free energy estimates and therefore tends to be more negative than experimentally determined binding free energy.

Observing different poses of the two leads, we observed that NSC288387 had greater conformational consistency between poses relative to inhibitor NSC111552. ### Biochemical confirmation of residues interacting with SAM, NSC 111552 and 288387 To confirm the essential roles of the critical residues identified in the molecular modeling, we performed mutagenesis study on all key residues of the DENV3 MTase by functional MTase assays.

To this end, we generated ten mutants of the DENV3 MTase contacting either SAM or the inhibitors. Except the H110A mutant that was not soluble, all other nine mutants, including T104A, K105A, D131A, V132A, F133A, G148A, E149A, R160A, and K180A, were expressed and purified (Figure 7A). We first used the HTRF-based N7 MTase activity assay as described above to quantify the mutant MTase activities compared to that of wild-type (WT).

As shown in Figure 7B, the DENV3 MTase mutations, T104A, L105A, D131A, G148A, and E149A, significantly decreased the N7-MTase activity of the DENV3 MTase, suggesting that these mutations are critical for the DENV3 MTase activity. In contrast, mutations, V132A and F133A, did not have any effects on the DENV3 MTase N7 activity, whereas the R160A and K180A mutations slightly reduced the MTase activity.

The majority of these residues form contact with the methyl donor SAM. For example, the sidechains of Thr104 and Asp131 form hydrogen bonds with SAM; Lys105, V132 and F133 form a cage to hold the adenosine group of SAM. Our results are in agreement with previous findings on the WNV MTase.

68 We next performed the FP assay with WT and mutant DENV3 MTases as described above. Our result suggests that the DENV3 MTase mutants V132A, F133A, E149A, and R160A are nearly as active as WT in binding of FL-NAH, whereas the mutations at positions 104, 131, and 148 led to complete loss of their binding to FL-NAH (Figure 7C). Mutants K105A and K180A retained about 20% binding ability to FL-NAH compared to WT.

The results of FL-NAH binding which access SAM-binding are consistent with those from the HTRF functional MTase assay as described above. Two discrepancies between the two assays are on residues E149 and K180. E149A is active in the FP assay but completely inactive in the HTRF assay, indicating that Glu149 may participate in the MTase function other than binding to SAM.

K180A is about 20% active in binding to FL-NAH. However, K180A is nearly 90% active in the HTRF functional assay. The sidechain of Lys180 forms close contact with the leaving methyl group of SAM.

It is reasonable that K180A mutation decreased the binding of FL-NAH. It is currently unclear why the K180A mutation retains its MTase activity. It is possible that K180A facilitates release of SAH, the by-product of the MTase reaction, so that the enzyme turnover rate is faster.

We next tested the effects of these mutations on inhibition of the DENV3 MTase activity by the identified inhibitors, NSC 111552 and 288387. Because only V132A, F133A, R160A, and K180A were biologically active, we only conducted the experiment with these mutants.

Our result showed that mutations Val132 and Phe133 significantly decreased the inhibitory efficacy of NSC111552 on the MTase activity, whereas R160A and K180A did not have significant effects. For NSC288387, mutations V132A, F133A, and R160A did not affect the compound’s inhibitory activity, whereas K180A slightly enhanced the inhibitory activity of the compound. These results are consistent with those from the docking studies.

We further tested inhibition of the FL-NAH binding activity to different mutants by NSC 111552 and 288387. Our results showed that the IC50-disp values for the inhibition of FL-NAH binding by NSC111552 were 12. 5-, 6.

0-, 6. 2-, and 3. 5-fold higher against the active mutants V132A, F133A, E149A, and R160A, respectively, than against the WT.

In contrast, the $I C_{50 \mathit{-disp}}$ values of compound NSC288387 for the V132A, F133A, E149A, and R160A mutants were 1. 0-, 1. 7-, 2.

0-, and 1. 6-fold higher, respectively, than for the WT MTase. (Figure 7E).

Finally, a thermal shift assay (TSA) was performed to check the melting point $\left(\right. T m \left. \right)$ changes in mutant DENV3 MTases in presence of NSC111552, compared to those of DMSO control (Table 4).

NSC288387 could not be evaluated using this assay due to compound fluorescence interference. Our results showed that binding NSC111552 to nearly all the mutants, except F133A, G148, and K180A, led to negative changes of $T_{m}$ values, suggesting that NSC111552 destabilized the mutant MTases instead of stabilizing them.

Binding of NSC111552 to the F133A, G148A, and K180A mutants led to significantly reduced $T_{m}$ changes compared to that to the WT (Table 4). All these results suggest that these mutations weaken the binding of NSC111552 to the MTases. ## [](https://pmc.

ncbi. nlm. nih.

gov/articles/PMC10436986/)Discussion Most of the flaviviruses as well as SARS-CoV-2 use N7 and 2’-O MTases to protect their RNA inside the host cells. 34, 35, 72–75 These MTases are essential for virus survival. Previous reports have suggested that flavivirus MTase can serve as a potential therapeutic target to develop therapeutics to combat flavivirus infection.

20, 43, 44, 53, 58, 64, 76, 77 It has been demonstrated that preventing N7 methylation can stop viral replication in general. 23, 24 Currently, a number of inhibitors that target the SAM-binding site of the NS5 MTase have been discovered.

20, 43, 44, 53, 58, 64, 76, 77 Sinefungin, an analogue of SAM, has been demonstrated to block the NS5 MTase of DENV and WNV with an $\text{IC}_{50}$ value in lower micromolar range, however, their antiviral efficacy in cell system is poor.

30, 76 Previously our lab has also reported a SAM-competitive inhibitor NSC 12155, showed inhibitory activities toward the NS5 MTases of WNV and YFV _in vitro_ with the $\text{IC}_{50}$ value in low micromolar range. 53 NSC 12155 also showed antiviral activity in cell-based experiments against WNV, DENV-2, and JEV, with $\text{EC}_{50}$ values ranging from 1. 00 to 7.

00 μM 53. Several other groups have shown development of small molecules against flavivirus MTase. 38, 41, 42 In this article, we developed an FP-based HTS platform to target the SAM-binding site of MTases from different flaviviruses.

The viral MTase is a crucial enzyme involved in producing RNA cap structures necessary for RNA translation and hiding viral RNA from the host immune response. Using this HTS screening method we were able to identify two inhibitors against flavivirus MTase. Using an FP-based assay, we first screened the chemical libraries from the NCI Diversity set VI library.

A readily available HTRF MTase functional assay was used to further characterize the top hit compounds (Cisbio, MA). Moreover, we used cell-based assays to assess their antiviral effectiveness and possible cytotoxicity. Through this procedure, two effective broad spectrum flavivirus MTase inhibitors were identified.

The binding of FL-NAH, a fluorescent SAM analog, was inhibited by NSC 111552 and 288387 with low micromolar $I C_{50 \mathit{-disp}}$. The anti-MTase activity of these compounds was further assessed using two different methods: HTRF and TLC with 32 P-labeled substrate.

Similar outcomes from these two methods demonstrated the effectiveness of these two compounds in inhibiting flavivirus MTase activity with low micromolar $\text{IC}_{50}$. Additionally, we demonstrated the direct binding of NCS 111552 and 288387 to the DENV-3 MTase using MST, with binding affinities in the low micromolar range. Using cell-based assays, we also assessed their antiviral strength and cytotoxicity.

Our findings demonstrated that these compounds have therapeutic indices of around 10 and single digit micromolar $\text{EC}_{50}$ values against DENV-2 and ZIKV viruses’ replication. Finally, we used mutagenesis to confirm the mechanism of action of these compounds. We also evaluated the selectivity of NSC111552 and NSC288387 against hRNMT.

The $I C_{50 \mathit{-disp}}$ values for NSC111552 and NSC288387 against hRNMT were observed to be 63. 3 μM and 35. 5 μM, respectively (Figure 2E and Table 2).

Although these values are significantly higher (nearly or >10-fold) compared to those for the DENV3, ZIKV, and WNV MTases, they are only 1. 5-fold and 3. 5-fold higher than the values for the YFV MTase.

Further modification of the compounds will be necessary to optimize selectivity. These results suggest multiple paths forward for compound optimization. Given the size, simplicity and chemical functionality of NSC111552, this compound has great potential for optimization notably through further substitutions to probe binding pocket or via isosteric replacement of hydroxyl groups to address possible PAINS properties of molecule.

As an example, replacement of hydroxyl groups with chlorine retains consistent docking interactions and posing of NSC111552 while introducing halogen binding with Gly148, Asp131 and Val 132. For NSC288387, using Maestro’s ligand designer, NSC288287 was analyzed for possible areas of improvement; three areas were of immediate interest.

First, while Lys 105 participates in hydrogen bonding with the ether oxygen, the terminal methyl group creates a steric clash in the pocket. Two areas of the molecule are solvent exposed in binding pocket (shown in gray halos in Figure 6) which may be exploited by the medicinal chemist to possibly gain greater binding affinity and specificity.

As an example, the largely solvent exposed terminal benzene ring is positioned in close proximity to charged residues, notably Asp 146, which could introduce ionic interactions with derivatives of this lead compound. In conclusion, these compounds offer a basis for further pharmacological modification to create more effective and specific broad-spectrum inhibitors against flaviviruses MTase. ## [](https://pmc.

ncbi. nlm. nih.

gov/articles/PMC10436986/)Materials and Methods ### Expression and Purification of flavivirus MTases The plasmid His-DENV3 NS5-pET28a encoding full-length DENV3 NS5 MTase was used to transform bacteria _E. coli_ to express and purify the NS5 protein as described previously. 44, 45 In brief, plasmid was transformed into _E.

coli_ BL21(DE3) cells. The next day, a single transformed colony was used to inoculate 50 ml of LB media with 50 μg/ml of kanamycin and cultured in a shaker with 250 rpm at 37°C for overnight. The overnight grown pre-culture was added to a culture medium (3 L) and then continued to grow at 37°C at 250 rpm until $\text{OD}_{600}$ reached 0.

9. Isopropyl -D-thiogalactoside (IPTG) was added to the cells to induce protein expression at a final concentration of 0. 5 mM.

The cells were then allowed to grow at 16 °C for overnight. The cells were harvested by centrifugation at 5,000 g for 10 min. The bacterial pellet was resuspended in 150 ml of lysis buffer containing 20 mM Tris, pH 8.

0, 0. 5 M NaCl, 10% Glycerol, 10 mM β-mercaptoethanol. The resuspended cells were sonicated with 70% power for 30 second “on” and 2 min “off” for 6 cycles The lysate was centrifuged at 15,000 g for 30 min.

The supernatant containing the soluble protein was loaded onto an Ni-NTA column equilibrated with the lysis buffer without β-mercaptoethanol. The resin was washed with the lysis buffer containing 20 mM imidazole. The bound DENV3 NS5 MTase was eluted with lysis buffer containing 300 mM imidazole.

The eluted fractions were analyzed by SDS-PAGE. The fractions containing the protein of interest were combined, concentrated to 5 ml and loaded onto a Superdex 75 column equilibrated with a protein storage buffer containing 25 mM HEPES, pH 7. 5, 0.

5 M NaCl, 10% Glycerol, 1 mM DTT. The fractions containing purified DENV3 NS5 MTase were pooled, concentrated, aliquoted, flash-frozen in liquid nitrogen and stored at −80 °C until further use. All mutant plasmids were generated by Gene Universal and used to express and purify mutant proteins as described above.

Similar procedures to those outlined above for the DENV3 NS5 MTase protein were used to produce the NS5 MTases of YFV, WNV, ZIKV and hRNMT, as we described previously. 24, 58, 64, 78 An FP-based assay using FL-NAH, a fluorescent analog of the methyl donor SAM, was carried out for inhibitor screening against DENV3 NS5. The reaction solution used for the screening assay was composed of 20 mM HEPES, pH 8.

5, 50 mM NaCl, 10% glycerol, 1 mM DTT, 0. 1% triton X-100, 50 nM FL-NAH, and 0. 5 μM DENV3 NS5.

The NCI diversity set VI library was used in the assay, which was conducted in a 25 μL reaction volume in 96-well black polypropylene plates (1,584 compounds, 20 plates). In the beginning, DENV 3 NS5 was incubated with DMSO or the inhibitor for 30 min at room temperature. After 30 minutes, FP was determined by adding FL-NAH to the reaction and utilizing 485 nm and 528 nm for excitation and emission, respectively.

Millipolarization (mP) units were used to define FP. To determine $I C_{\text{50}-\text{disp}}$ defined as 50% inhibition of FP resulted from FL-NAH binding to the DNEV3 NS5 MTase by a compound, the assay was done with 50 nM