Structural Modifications and Biological Evaluations of Rift Valley Fever Virus Inhibitors Identified from Chemical Library Screening

The Rift Valley fever virus (RVFV) is an emerging high-priority pathogen endemic in Africa with pandemic potential. There is no specific treatment or approved antiviral drugs for the RVFV. We previously developed a cell-based high-throughput assay to screen small molecules targeting the RVFV and identified a potential effective antiviral compound (1-N-(2-(biphenyl-4-yloxy)ethyl)propane-1,3-diamine) as a lead compound. Here, we investigated how structural modifications of the lead compound affected the biological properties and the antiviral effect against the RVFV. We found that the length of the 2-(3-aminopropylamino)ethyl chain of the compound was important for the compound to retain its antiviral activity. The antiviral activity was similar when the 2-(3-aminopropylamino)ethyl chain was replaced with a butyl piperazine chain. However, we could improve the cytotoxicity profile of the lead compound by changing the phenyl piperazine linker from the para-position (compound 9a) to the meta-position (compound 13a). Results from time-of-addition studies suggested that compound 13a might be active during virus post-entry and/or the replication phase of the virus life cycle and seemed to affect the K+ channel. The modifications improved the properties of our lead compound, and our data suggest that 13a is a promising candidate to evaluate further as a therapeutic agent for RVFV infection.


■ INTRODUCTION
Rift Valley fever (RVF) is an acute viral infection caused by the emerging mosquito-borne Rift Valley fever virus (RVFV) (genus Phlebovirus, family Phenuiviridae), which infects domestic animals and humans. The RVFV causes deadly infection among ruminants with high fever, hepatitis, acute deaths of newborns, and abortions in pregnant animals. Abortion storms are considered as a hallmark of RVFV outbreaks. 1−3 Humans are infected by mosquito bites as well as handling contaminated animal tissues and fluids while working with slaughter or taking care of infected animals in herds. In humans, RVFV infection ranges from a mild illness associated with fever and liver abnormalities to much more severe symptoms such as retinitis, encephalitis, and hemorrhagic fever. 4,5 The association of RVFV infection with miscarriage in humans has also been reported. 6,7 The RVFV causes recurrent outbreaks throughout the African countries. 8 The RVF epidemic in Yemen and Saudi Arabia in 2000 was the first of this kind outside Africa. This epidemic affected both livestock and humans, and approximately 200 humans died. 9,10 The spread of the RVFV outside Africa was mainly due to import of infected animals from epidemic countries. Notably, more than 30 different mosquito species have been identified to carry the RVFV, of which several are competent vectors and are distributed globally. 11 The RVFV has been endemic in the African subcontinent for decades, but outbreaks in Saudi Arabia, Yemen, Madagascar, and the Comoros Archipelago suggest that the geographical distribution of the RVFV is changing. 12−14 Recent studies have reported the presence of RVFV-seropositive animals in Iran and Turkey, 15,16 and other Asian countries are also at risk. Due to the presence of competent mosquito vectors in Europe, studies have emphasized that the RVFV can pose a major threat there. 17−19 Currently, there are no safe and effective treatments to prevent or cure RVFV-infected humans or livestock. So far, several natural products and synthetic chemical compounds have been reported as potent RVFV inhibitors in vitro, but none of them progressed further to become an RVFV-specific drug candidate. 20−26 Favipiravir is not RVFV-specific but has broad-spectrum activity against a number of RNA viruses, including the RVFV, and has gone through human clinical trials. 27−29 However, there is a need to develop potent efficacious antiviral compounds against RVFV infection.
Previously, we have developed a high-throughput screening method for identifying potent inhibitors of RVFV infection in vitro and identified several compounds. 30 The parent compound (N 1 -(2-(biphenyl-4-yloxy)ethyl)propane-1,3-diamine) was identified as a promising hit for further evaluation ( Figure 1). The structure of the parent compound (designated as compound 1) consists of a biphenyl group that is connected, through an ether bond, to a 2-(3-aminopropylamino)ethyl chain. To investigate the important structural factors of parent compound 1, we aimed to synthesize a set of first-generation compounds with modification of the 2-(3-aminopropylamino)ethyl chain and keep the biphenyloxy part conserved. Depending on the result after the first investigation, the intention was to investigate the biphenyl part by synthesizing a set of second-generation compounds with modifications of the biphenyl part and keep the modified alkyl amine chain of interest conserved.

■ RESULTS AND DISCUSSION
Chemistry: First-Generation Compounds. To determine the importance of the 2-(3-aminopropylamino)ethyl chain for the parent compound 1, a set of first-generation compounds were synthesized. The approach for synthesizing the compounds 4d−e, 5a−c, and 5f−g is described in Scheme 1. Ether bond formation between alcohols 3a−g and 4biphenol 2 was performed under standard Mitsunobu conditions using di-isopropyl azodicarboxylate (DIAD) and triphenylphosphine (PPh 3 ) in tetrahydrofuran (THF), which gave compounds 4a−g. 31 N-Boc removal of N-Boc-protected compounds 4a−c and 4f−g was performed under acidic conditions with either trifluoroacetic acid (TFA) or HCl, which gave the corresponding deprotected compounds 5a−c and 5f−g as free amines or HCl salts.
The synthetic approach for alkyl-piperazine analogs 8c and 9a−b is described in Scheme 2. For synthesis of compounds 8a−c, N-Boc piperazine 6a or N-methyl piperazine 6b was treated with an equivalent amount of 1,4-dibromobutane or 1,5-dibromopentane in acetonitrile and excess cesium carbonate as a base. Increasing the temperature to 115°C for 20 min using microwave heating allowed the spiro salt intermediate 7a−7c to form in situ. Addition of 4-biphenol 2 in small excess and increased microwave heating to 170°C resulted in nucleophilic ring opening of the spiro salt intermediate, which gave the alkyl piperazine analogs 8a−c in a one-pot fashion in <57% yield. The initial attempt was performed using potassium carbonate as a base instead of cesium carbonate, and the reaction worked but a higher temperature was necessary for completion. In addition, an attempt of changing the solvent to dimethylformamide (DMF) was also performed but without any success. Finally, acidic N-Boc removal of compounds 8a−b with HCl gave 9a−b as di-HCl salts.
For synthesis of the N-phenyl-substituted butyl piperazine 11, the N-phenyl piperazine spiro salt 7c was isolated and reacted as a substrate with 4-biphenol. The synthetic approach for compound 11 is described in Scheme 3. The N-phenyl piperazine spiro salt 7c was synthesized from N-phenyl piperazine 6c, which was reacted with an equivalent amount of 1,4-dibromobutane in acetonitrile and potassium carbonate as a base. The reaction was performed at 115°C using microwave heating for 15 min, which gave 7c in 79% yield after isolation. Nucleophilic ring opening of the spiro salt 7c was performed using acetonitrile as a solvent in excess of cesium carbonate as a base at 170°C accomplished using microwave heating. After 25 min, the reaction was completed, and compound 11 was isolated in 89% yield.
Biology: First-Generation Compounds. To evaluate the first-generation compounds, their antiviral activities were analyzed using the rRVFVΔNSs::Katushka virus assay essentially as described previously. 30 Briefly, A549 cells were infected with rRVFVΔNSs::Katushka (multiplicity of infection (MOI) = 0.1) together with serial dilutions of test compounds and incubated for 16 h. Then, the number of virus-infected cells was quantified by monitoring the expression of Katushka fluorescent proteins in a Trophos plate runner HD (Trophos, Roche Group). To determine the cytotoxic concentration (CC) 50 values, A549 cells were treated with three-fold serially diluted compounds for 24 h, and the cytotoxicity was measured using a resazurin cell viability assay (Sigma-Aldrich). The results from the antiviral and cytotoxicity assays are summarized in Table 1. Changing the 2-(3-aminopropylamino) ethyl chain in compound 1 to a shorter ethyl (compound 5b) or propyl amine (compound 5c) chain resulted in reduction of antiviral activity but improved the cytotoxicity profile. On the other hand, replacing the 2-(3-aminopropylamino) ethyl chain with a hexyl amine chain (compound 5a) showed a similar antiviral effect to compound 1 but increased the cytotoxicity. Introducing a heterocycle instead of the amino functionality for the ethylamine-substituted compound 5b to an N-piperidine (compound 4d) or Nmorpholine (compound 4e) decreased the antiviral potency but showed an improved cytotoxicity profile. However, the Npiperazine-substituted compound 5f showed similar antiviral potency to the ethylamine-substituted compound 5b. Increasing the length of the alkyl piperazine chain to butyl (compound 9a) and pentyl (compound 9b) showed increased antiviral activity with an increased length of the alkyl ligand, but the propyl (compound 5g) showed reduced antiviral activity. Butyl and pentyl piperazine compounds 9a and 9b showed similar potency with EC 50 = 12.8 ± 0.2 μM and EC 50 = 11.7 ± 1.8 μM to the parent compound 1 (12 μM), and butyl piperazine compound 9a showed a similar cytotoxicity profile (CC 50 = 74.8 μM ± 1.2) to compound 1 (CC 50 = 86 μM ± 9); meanwhile, pentyl piperazine compound 9b showed increased cytotoxicity. Introducing an N-methyl in the piperazine moiety (8c) decreased the antiviral potency, and for the N-phenyl (11), the antiviral activity was completely eliminated.
The antiviral activity identified from the fluorescence assay was further confirmed with an orthogonal qRT-PCR assay. A549 cells were infected with rRVFVΔNSs::Katushka (MOI = 1) and treated and incubated with the compound for 16 h. Thereafter, viral RNA was purified, and the amount of synthesized viral RNA was quantified by qRT-PCR ( Figure  2). As shown in Figure 2, the qRT-PCR results resembled the results obtained from the fluorescence assay. Analogues (4d− e) were unable to inhibit the RNA expression at the concentrations used in the qRT-PCR assay. Also, the analogues 5b−c, 5f−g, and 8c were also unable to inhibit the viral RNA expressions. In the fluorescence assay, we observed that analogues 5a and 9a−b had similar EC 50 values to compound 1. These three analogues also had similar patterns to compound 1 regarding inhibition of viral RNA expression. When taking into account the CC 50 values, butyl piperazine compound 9a was considered the best among the firstgeneration analogues with a similar cytotoxicity profile to compound 1. The N-phenyl piperazine analogue 11 did not show any inhibition during the fluorescence assay, and a similar phenomenon was also observed by the qRT-PCR assay.
Chemistry: Second-Generation Compounds. Based on the first-generation compounds, we concluded that the butyl piperazine ligand in compound 9a was the best choice for further modification of the biphenyl part. Compound 9a with a butyl piperazine ligand showed an equal antiviral activity and toxicity profile to the parent compound 1 with a 2-(3aminopropylamino)ethyl ligand. For synthesis of the butyl piperazine analogs, the spiro salt 7a was synthesized and isolated and used as a substrate for further synthesis of the butyl piperazine analogs 12a−m by nucleophilic ring opening of the spiro salt. Acidic removal of the Boc group using HCl gave 13a−m as di-HCl salts in 25−72% yields calculated over two steps. The synthetic approach for compounds 13a−m is described in Scheme 4. For synthesis of compound 16, the above approach was not successful. Instead, compound 16 was synthesized in two steps described in Scheme 5. First, diphenylmethanol 14 was O-alkylated with 1,4-dibromobutane and sodium hydride as a base to give compound 15 in 22% yield. Second, N-alkylation of piperazine with 1 equiv of compound 15 and excess of cesium carbonate gave compound 16 in 57% yield.
Biology: Second-Generation Compounds. The butyl piperazine analogues with modification of the biphenyl part showed a wide range of antiviral and cytotoxic activities ( Table  2). Changing the biphenyl part to a smaller phenyl group (13f) completely removed the antiviral activity. In addition, introducing a more bulky group such as diphenyl methyl (16) instead of the biphenyl reduced the antiviral activity significantly. For the investigation of the structural isomers of the biphenyl part, the antiviral activity was similar for both the para-position of 9a (EC 50 = 12.8 μM ± 0.2) and the metaposition of 13a (EC 50 = 13.8 ± 5.3 μM) but reduced when the phenyl was located in the ortho-position of 13b (EC 50 = 63.5 μM ± 11.1). On the contrary, an improved cytotoxicity profile was observed when the phenyl was located in the meta-position (CC 50 = 144.8 ± 5.5 μM). Then, the phenyl was changed to benzyl (13c−e), and among them, the highest antiviral activity was observed for the benzyl located in the para-position of 13c (EC 50 = 15.4 ± 4.7 μM) compared to the ortho-position of 13d (EC 50 = 22.9 ± 5.2) and the meta-position of 13e (EC 50 = 31.2 ± 15.8). Benzyl-substituted compounds (13c−e) showed a similar cytotoxicity profile to the meta-biphenyl analog 13a. Introducing a more bulky substituent such as cyclohexyl (13l) instead of phenyl (9a) in the para-position increased the antiviral activity (EC 50 = 8.8 ± 0.5 μM) but also increased the cytotoxicity. Extreme toxicity was observed when the outer phenyl ring was replaced with adamantine (13m). To investigate if addition of an electron-donating, electronwithdrawing, or steric group could affect the antiviral activity, an acetyl (13g), dimethylamine (13h), or isopropyl (13j) substituent was introduced in the meta-position of the outer ring of the biphenyl part. For the meta-acetyl-substituted compound 13g, the antiviral activity decreased (EC 50 = 41.9 ± Table 1. The Antiviral Activity, Cytotoxicity, and Selectivity Index of First-Generation Compounds a a EC 50 , half-maximum effective concentration; CC 50 , cytotoxicity concentration at 50%. Both values are representatives of two independent experiments with two replicates each time. ND, not detectable. 27.3 μM), while the meta-N,N-dimethylamine-substituted compound 13h (EC 50 = 15.5 ± 0.7 μM) displayed improved antiviral activity but showed high toxicity. Good antiviral activity was detected when the isopropyl substituent was introduced to either the para-(13i) or ortho-position (13k). In contrast, addition of an isopropyl group to the biphenyl ring may have enhanced the antiviral activity for 13i−k but diminished the cytotoxicity profile. One interesting observation was that analogues with low EC 50 values (13i−m) exhibited high toxicity. This suggested that the observed antiviral activities of these compounds were actually cytotoxic effects.
Similar to the first-generation compounds, the antiviral activity of the second-generation compounds identified from the fluorescence assay was further confirmed by qRT-PCR. Compared to 13a, compounds (13b−e and 13 g−h) showed low or no inhibitory effect on RVFV RNA expression. We had similar observation for compounds 13b−e and 13 g−h when we detected the EC 50 values by the fluorescence assay. In contrast, compounds (13i−m) having lower EC 50 values than 13a in the fluorescence assay also efficiently inhibited the RVFV RNA expression in qRT-PCR. However, as mentioned earlier, these compounds were highly cytotoxic; therefore, the observed antiviral potency was most probably due to their  toxic effect on the cells and not because they inhibited the virus itself ( Figure 3). Taking into account both antiviral activity and cytotoxicity of all the compounds, 13a was considered the best to be further explored in a mode-of-action study, although 13a has similar antiviral activity to compound 1. However, 13a had a better cytotoxicity profile than compound 1.
Mode-of-Action Studies. The time-of-addition assay was performed to determine at which stage in the RVFV infection cycle compound 13a had an effect. Compound 13a (50 μM) was added at different time points of virus (rRVFVΔNSs::Katushka) infection: preinfection (−1 h before infection), during virus addition (0 h), at early post-entry (2 h post infection (hpi) and 4 hpi), and at late stages of virus infection (6 and 8 hpi). The compound was also added to the cells 1 h before infection, incubated for 1 h, and then removed just before virus addition (−1 to 0 h) as indicated in Figure 4. The experiment was terminated at 13 hpi, and infectivity was assessed by the fluorescence assay, as described previously. As demonstrated in Figure 4, data suggested that the highest antiviral potency was observed when compound 13a was added in the beginning of virus infection. These results suggested that compound 13a was active at early stages of the viral replication cycle, most likely during or just after virus entry. The compound 13a progressively lost its potency when added at late stages of the infection cycle.
It is now well-evident that virion fusion and entry to the host cells are largely regulated by ion channels. Studies have shown that the functionality of ion channels plays a crucial role during entry or post-entry stages of several viruses. 34 For example, the hepatitis C virus (HCV) requires ion channels for its successful infection cycle. 35,36 Therefore, ion channels could be new targets to counteract virus infections, with a potentially broadspectrum antiviral activity useful for future pandemics. Previous antiviral screens of chemical compounds identified several clinically approved ion channel inhibitors as membrane fusion blockers of the HCV. 37−39 Many of them were closely related to our compound 13a. Recently, it has been reported that the K + channel regulates the post-entry stages of the Bunyamwera virus (BUNV), a closely related virus to the RVFV, and the authors showed that blocking the K + channel with chemical compounds inhibits BUNV infection. 40,41 Based on data from our time-of-addition experiments and the abovementioned facts, we hypothesized that compound 13a might influence ion channels. Therefore, we performed resting membrane potential experiments to evaluate if compound 13a has effects on the ion channel (i.e., the K + channel).
The K + channel plays a vital role to maintain the charge difference across the cell membrane (the resting membrane potential). Due to the changes of the ion channel's status (either open or closed), the membrane potential can either become more positive (depolarization) or more negative (hyperpolarization). This scenario can be monitored using a membrane potential-sensitive dye, bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4 (3)). 40,42 An increased DiBAC4(3) fluorescence intensity indicates cell depolarization, while a decreased fluorescence intensity means cellular hyperpolarization. We measured the membrane potential of A549 cells when treated with compound 13a or quinidine as a positive control to validate the assay. Quinidine is a known K + channel blocker that leads to depolarization. Before performing the membrane potential experiment, we examined the antiviral activity of quinidine and confirmed that it inhibits RVFV infection (EC 50 value = 146 μM) (Figure 5a). We then performed the membrane potential assay, and A549 cells treated with compound 13a (25 and 50 μM) exhibited an increased DiBAC4(3) fluorescence intensity (depolarization) similar to quinidine (200 μM), compared to cells that remained untreated, which showed a very low or decreased fluorescence intensity (hyperpolarization) (Figure 5b). This indicated that compound 13a either blocked directly or had an indirect effect on the K + channel.

■ CONCLUSIONS
The highly pathogenic RVFV has great health and socioeconomic impacts on endemic countries and could spread to new regions, with potentially devastating consequences. It is crucial to develop better therapeutics to prevent public and animal health threats. In this study, we investigated the SAR of the novel antiviral compound 1, previously identified from chemical library screening specifically for the RVFV. 30 In the first-generation compounds, we examined the importance of the (3-aminopropylamino)ethyl chain of compound 1. Here, we showed that the length of the alkyl amine chain seemed to be important to retain the antiviral activity, but it was also affecting the cytotoxic profile. A general observation was that for compounds with shorter alkyl amine chains (5b−c and 5f− g), the antiviral efficacy and toxicity against A549 cells decreased, and for compounds with longer alkyl amine chains (5a and 9a−b), the antiviral efficacy and toxicity increased. In addition, the presence of a hydrogen bond-donating amine functionality seemed to be necessary for the antiviral activity. When N−H for the ethyl piperazine analog 5f was replaced with −CH2 (4d) or oxygen (4e), the antiviral activity was dramatically reduced. Similarly, when the hydrogen bonddonating amine functionality of the butyl piperazine analog 9a was substituted with N-methyl (8c) or N-phenyl (11), the antiviral activity was reduced or completely abolished. The butyl piperazine compound 9a and the parent compound 1 showed similar antiviral activity and toxicity to A549 cells, but taking into account the CC 50 values, butyl piperazine compound 9a was considered the best candidate for further modifications.
In the second-generation compounds, we examined the importance of the biphenyl part of the butyl piperazine compound 9a. An investigation of the structural isomers by exchanging the para-biphenyl (9a) with ortho-biphenyl (13a) or meta-biphenyl (13b) or substitution of phenyl to benzyl (13c−e) resulted in no improvement in the antiviral activity. However, when the biphenyl was located in the ortho-position (13a), the CC 50 value was improved to almost double compared to the biphenyl in the para-position (9a). When the biphenyl group was replaced with a phenyl (13f) or 1,2diphenyl methyl (16), the antiviral activity was completely removed or significantly reduced. Addition of an isopropyl group to the biphenyl ring induced antiviral activity but showed more toxicity (13i−k). Therefore, considering both the antiviral activity and the cytotoxicity profile, compound 13a was the best compound with a selectivity index (SI) of 10. In addition, mode-of-action studies suggested that 13a could inhibit the post-entry or the early replication phase of the RVFV life cycle, and it affected the K + channel. In the future, it would be interesting to investigate the pharmacokinetics and antiviral efficiency of compound 13a in an animal model. To conclude, our studies have identified a novel compound with the potential to be further developed as an antiviral drug against the emerging and potentially deadly RVFV infection for which there are no available therapeutics.
General Synthetic Procedure. Procedure A: Synthesis of Compounds 4a−g (R 2 = Boc) Using the Mitsunobu Reaction, Exemplified for Compound 4a. Biphenyl-4-ol (86.2 mg, 0.506 mmol), tert-butyl 6-hydroxyhexylcarbamate 3a (100 mg, 0.46 mmol), and PPh 3 (145 mg, 0.55mmol) were dissolved in 1 mL of THF. To the solution, DIAD (0.11 mL, 0.55 mmol) was added, and the reaction was stirred at rt for 9 days. The resulting mixture was diluted with EtOAc and washed two times with brine. The organic phase was dried with Na 2 SO 4 , filtrated, and concentrated. Purification with flash chromatography [P:E 6:1] over silica gave 172 mg of the Bocprotected compound 4a, which was taken directly to the next reaction.
Procedure B: Synthesis of Compounds 5a−c and Deprotection of the Boc Group Using TFA, Exemplified for Compound 5a. Boc-protected compound 4a dissolved in 1.5 mL of DCM and 1.5 mL of TFA was added. After ca. 2.5 h of reaction, it was diluted with water and DCM. NaOH (aq) (2 M) was added until basic pH. The aqueous phase was extracted three times with DCM. Organic phases were combined and washed with brine, dried with anhydrous Na 2 SO 4 (s), filtrated, and concentrated. Free amine was redissolved in Et 2 O, and 4 M HCl (dioxane) was added until acidic pH. The resulting mixture was concentrated and triturated with Et 2 O three times, which gave 5a as a HCl salt (52 mg, 0.17 mmol) in 37% yield.
Procedure C: Deprotection of the Boc Group Using HCl, Exemplified for Compound 5g. Boc-protected 4g (160 mg, 0.404 mmol) was dissolved in DCM (1 mL). HCl (4 M) in dioxane (1.6 mL) was added, and the reaction was stirred at rt for 3 h. The resulting mixture was concentrated and triturated three times with Et 2 O, which gave 5g (116 mg, 0.315 mmol) as a solid in 60% yield calculated over two steps.
Procedure E: Spiro Salt Formation, Exemplified for Compound 7a. Boc-protected piperazine 6a (500 mg, 2.685 mmol) was dissolved in MeCN (3 mL), and K 2 CO 3 (1.113 g, 8.054 mmol) and 1,4-dibromobutane (337 μL, 2.819 mmol) were added. The reaction was capped (microwave vessel) and heated using a sand batch at 120°C for 40 min. The resulting mixture was diluted with CHCl 3 (product soluble in CHCl 3 ) and filtrated to remove K 2 CO 3 solids. The organic phase was extracted with Milli-Q water twice. The combined water phases were co-concentrated using absolute EtOH. The resulting colorless precipitate was dissolved in CHCl 3 and dried with anhydrous Na 2 SO 4 (s), filtrated, and evaporated. After drying under vacuum, it gave 7a (757 mg, 2.356 mmol) as a colorless solid in 87% yield.
Procedure F: Synthesis of Substituted Biphenols Using the Suzuki Coupling Reaction, Exemplified for Compound 10a. 3-Bromophenol (1.26 g, 7.29 mmol), benzeneboronic acid (1,78 g, 14.57 mmol), K 2 CO 3 (2.52 g, 18.22 mmol), and Pd(PPh 3 ) 2 Cl 2 (26 mg, 0.037 mmol) were mixed with a solution of 20 mL of dioxane and 5 mL of water followed by reflux overnight. The resulting black mixture was diluted with Et 2 O and washed with H 2 O. The water phase was extracted with additional Et 2 O. The combined organic phases were dried with Na 2 SO 4 , filtrated, and concentrated. Flash chromatography over silica [8.5:1 H:E] gave 10a (0.91 g, 5.35 mmol) as a colorless solid in 73% yield.
Procedure G: Nucleophilic Ring Opening of the Spiro Salt, Exemplified for Compound 12a. The spiro salt 7a (50 mg, 0.156 mmol) and 3-biphenol 10a (29.13 mg, 0.171 mmol) were mixed in 0.8 mL of MeCN. Cs 2 CO 3 (152 mg, 0.467 mmol) was added, and the reaction was microwave heated at 170°C for 25 min. The resulting mixture was diluted with EtOAc and washed two times with brine. The organic phase was dried with anhydrous Na 2 SO 4 , filtrated, and concentrated.  mmol) was added, and the reaction was stirred at 55°C. After 16 h, additional PPh 3 (137 mg, mmol) and DIAD (0.1 mL, mmol) were added, and the reaction was further stirred at 70°C for 24 h. The resulting mixture was diluted with EtOAc and washed two times with brine. The organic phase was dried with Na 2 SO 4 , filtrated, and concentrated. Purification using flash chromatography [E:P 1:7] gave 222 mg of the Boc-protected compound 1 in 45% yield. For removal of the Boc groups, 33 mg of the product was dissolved in 0.4 mL of DCM and 0.5 mL of 4 M HCl (dioxane was added). After 8 h, the resulting mixture was concentrated, and the retained solid was triturated three times with Et 2 O, which gave compound 1 (mg, mmol) as a di-HCl salt in 75% yield. 1  3-(Biphenyl-4-yloxy)propan-1-amine (5c). 5c was synthesized by procedures A and B in 42% yield as a free amine. 1  1-(2-(Biphenyl-4-yloxy)ethyl)piperazine (5f). 5f was synthesized by procedures A and B in 41% yield as a free amine. 1  FBS, with multiplicity of infection (MOI) = 0.1. The growth medium was removed, 100 μL of the virus and the compound mixture was added to the cells, and the plate was incubated for 16 h at 37°C in 5% CO 2 . Later, the medium was removed, and cells were fixed for 1 h with 3% paraformaldehyde (PFA); then, the cellular nuclei were stained with 0.1% DAPI for 15 min. The wells were washed with PBS, and the number of infected cells/well was counted by a Trophos plate runner HD (Trophos, Roche Group) following the expression of the Katushka protein by the virus. Simultaneously, the total number of cells/well was also counted following the DAPI staining. GraphPad Prism software version 9.2.0 (GraphPad Software, La Jolla, CA, USA) was used to calculate the EC 50 value with nonlinear regression analysis with a variable slope. All laboratory work with the rRVFVΔNSs::Katushka virus was performed under biosafety level 2 conditions as approved by the Swedish Work Environment Authority.
Cellular Toxicity Assay. The resazurin cell viability assay (Sigma-Aldrich) was used to analyze the cellular toxicity of synthesized analogues, described previously. This assay measures the metabolic activity of living cells and is based on the oxidoreduction of the nontoxic indicator blue dye resazurin. Viable cells with active metabolism can reduce resazurin into resorufin, which is pink and fluorescent. Briefly, A549 cells (approximately 10,000/well) were seeded in a black-wall transparent-bottom 96-well plate and incubated at 37°C in 5% CO 2 overnight. Cells were then treated with compound concentrations starting from 300 μM with 2-fold serial dilutions down to 2.34 μM and incubated at 37°C in 5% CO 2 for 24 h. To analyze the cell survival/toxicity, 10 μL (40 μM final concentration) of resazurin was added per well and incubated for 3−4 h at 37°C in a 5% CO 2 incubator, and the resorufin fluorescence intensity was measured by a Trophos plate runner HD (Trophos, Roche Group). The CC 50 value was then calculated with GraphPad Prism software version 9.2.0 (GraphPad Software, La Jolla, CA, USA) following the nonlinear regression analysis with a variable slope.
Viral RNA Extraction and qRT-PCR. Cell seeding, virus infection with rRVFVΔNSs::Katushka (MOI = 1.0), and compound addition were carried out in the same way as previously described. 33 Briefly, A549 cells (approximately 10,000/well) were seeded in transparent 24-well plates and incubated at 37°C in 5% CO 2 overnight. Cells were then infected with rRVFVΔNSs::Katushka (MOI = 1.0) together with the compound (50 and 25 μM) and incubated at 37°C in 5% CO 2 for 16 h. Then, the virus inoculum was discarded, cells were washed with PBS and lysed with proteinase K, and the total cellular RNA was extracted. Extraction of viral RNA and quantification of the viral load were performed as previously described. 33 Time-of-Addition Assay. The fluorescent cell focus assay was used for the time-of-addition assay, as previously described. 33 In short, A549 cells were infected at an MOI of 0.1, and 50 μM compound 13a was added prior to infection (−1 h), at the time of infection (0 h), and at 2, 4, 6, and 8 h after infection. An additional experiment was to treat cells with 50 μM of the compound 1 h before infection (−1 h) and then remove the cell medium, containing the compound, at the time of infection (−1 to 0 h). The infection was assayed by the fluorescent cell focus assay at 13 h post infection. The cellular nucleoli were stained with 0.1% DAPI and counted as described in the previous section.
Resting Membrane Potential Assay. Before performing the resting membrane potential assay, we performed the dose− response activity of quinidine and compound 13a, similar to the effective concentration assay. The only exception was that both compounds were serially diluted in two-fold steps, from 400 to 3.12 μM for quinidine and from 100 to 1.56 μM for compound 13a. The resting membrane potential assay was performed as previously described. 40 Briefly, A549 cells were treated with 20 μM DiBAC 4 (3) (Sigma) for 20 min at 37°C in the dark. After labeling, cells were washed and treated with either quinidine (200 μM) as a positive control or compound 13a (25 and 50 μM) for 16 h. Labeled cells were also left untreated as the DiBAC 4 (3) control. Wide-field images were taken with an Olympus CKX53 fluorescence microscope.
Nuclear magnetic resonance (NMR) spectra and highperformance liquid chromatography (HPLC) analysis of the compounds; purity of the compounds used in the experiments analyzed by NMR and HPLC (PDF)