Antiviral effects of the fused tricyclic derivatives of indoline and imidazolidinone on ZIKV infection and RdRp activities of ZIKV and DENV

Highlights

• •Fused tricyclic derivatives of indoline and imidazolidinone target ZIKV and its RdRp.
• •Some distinctive compounds are selective ZIKV inhibitors or DENV inhibitors.
• •Some broad-spectrum inhibitors of ZIKV- and DENV-infection are discovered.
• •12, 17 and 28 are more active against Asian ZIKV strain than African ZIKV strain.
• •A compound's antiviral differentiation is not always related to anti-RdRp differentiation.
• •12 majorly acts on post-infection of RNA synthesis stage of ZIKV life cycle.

Abstract

The prevalence and ravages of Zika virus (ZIKV) seriously endanger human health, especially causing significant neurological defects in both neonates as pediatric microcephaly and adults as Guillain–Barré syndrome. In this work, we studied anti-ZIKV effects of the fused tricyclic derivatives of indoline and imidazolidinone and discovered that some of them are valuable leads for drug discovery of anti-ZIKV agents. The current results show that certain compounds are broad-spectrum inhibitors of ZIKV- and dengue virus (DENV)-infection while distinctive compounds are selective ZIKV inhibitors or selective DENV inhibitors. Compounds of 12, 17 and 28 are more active against Asian ZIKV SZ-VIV01 strain than African ZIKV MR766 strain. It is valued that silylation makes six TBS compounds of 4-nitrophenyl hydrazine series and phenyl hydrazine series more active against ZIKV infection than their phenols. Time-of-addition and withdrawal studies indicate that compound 12 majorly acts on post-infection of RNA synthesis stage of ZIKV life cycle. Moreover, compounds of 12, 17 and 18 are anti-ZIKV agents with the inhibitory activities to ZIKV NS5 RdRp while 12 doesn't inhibit DENV infection even though it is a DENV RdRp inhibitor, 17 is an active agent against DENV infection but is only a weak DENV NS5 RdRp inhibitor, and 28 is inactive against DENV infection and not a DENV NS5 RdRp inhibitor. As a result, a compound's antiviral difference between ZIKV and DENV is not always related to anti-RdRp difference between ZIKV RdRp and DENV RdRp, and structural features of a compound play important roles in executing antiviral and anti-RdRp functions. Further discovery of highly potent broad-spectrum or selective agents against infection by ZIKV and DENV will be facilitated.

1. Introduction

The prevalence and ravages of Zika virus (ZIKV) seriously endanger human health. ZIKV is classified as a flavivirus (Flaviviridae family) and belonging to one of the major arboviruses spreading through arthropod vectors (mainly ticks and mosquitoes). ZIKV was first isolated from a febrile sentinel rhesus monkey in Uganda in 1947 (Dick et al., 1952) and the first confirmed human infection was reported in Uganda in 1962–63 (Wikan and Smith, 2016). ZIKV infection was originally understood to be a mild self-resolving infection and frequently misdiagnosed with other arboviruses’ infection, especially dengue virus (DENV, a flavivirus) and chikungunya virus (CHIKV, alphaviruses) (Baud et al., 2015; Campos et al., 2015; Cardoso et al., 2015; Cao-Lormeau and Musso, 2014). ZIKV becomes one of the most rapidly spreading mosquito-borne human pathogen in the tropical and subtropical regions and led to significant adverse impacts upon human health. ZIKV caused a big human epidemic in the Americas from 2015 to 2017, which made World Health Organization (WHO) declared a Public Health Emergency of International Concern in February 2016 and at last spread in 84 countries as of March 2017. It is now confirmed that ZIKV can cause significant neurological defects in both neonates as pediatric microcephaly (Bhagat et al., 2021; Calvet et al., 2016; Mlakar et al., 2016) and adults as Guillain–Barré syndrome (Brasil et al., 2016). In addition to neurologic complications, possible nonarthropod-mediated transmission in humans raises healthcare concerns (Osuna et al., 2016). Therefore, the search for efficient antivirals to ZIKV is of great medical significance. Some repurposing drugs were discovered as anti-ZIKV agents (Nascimento et al., 2021; Gardinali et al., 2020; Rosa et al., 2020; Rampini et al., 2020; Coronado et al., 2018; Munjal et al., 2017; Adcock et al., 2017; Albulescu et al., 2017; Rausch et al., 2017; Tan et al., 2017; Deng et al., 2016). Moreover, there are many Nature-originated and synthesized ZIKV inhibitors to be recently reported (Yu et al., 2022; Samrat et al., 2022; Nunes et al., 2022; Zhou et al., 2021; Cirne-Santos et al., 2021; Cataneoet al., 2021; Wardana et al., 2021; Patel and Gulick, 2021; Felicetti et al., 2020; Wang et al., 2019; García et al., 2017). Even though great efforts have been provided and invested in prevention and therapy research & development (R&D), currently there is still no approved vaccine or specific antiviral against ZIKV.

As one of flavivirues, ZIKV has a monopartite, linear, 11 kb positive sense single-stranded RNA genome, encoding a single polyprotein with multiple transmembrane domains that is cleaved, by both host and viral proteases, into 3 structural (Capsid (C), pre-membrane (prM) and envelope (E)) and 7 non-structural proteins (NS1, NS2A, NS2B, NS3, NS4, NS4B, NS5) (Hou et al., 2017; Chambers et al., 1990). As a non-structural protein, the flavivirus NS5 protein is the largest (∼ 100 kDa) and most conserved ZIKV protein among their genomic products, which is comprised of two domains, an N-terminal methyltransferase (MTase) domain and an RNA-dependent RNA polymerase (RdRp) domain at the C-terminal end (Upadhyay et al., 2017; Wang et al., 2017; Zhao et al., 2017). Similar to other members of Flaviviridae family, the RdRp domain of ZIKV NS5 (ZIKV NS5 RdRp) as a metalloenzyme plays an important role in ZIKV replication and RNA synthesis to generate positive- and negative-sense copies of the RNA genome via a de novo mechanism (Nascimento et al., 2021; Tan et al., 2019; Hou et al., 2017; Choi and Rossmann, 2009; Ackermann et al., 2001; Kao et al., 2001; Chambers et al., 1990), in that RdRp uses an RNA template but does not require a primer to elongate nascent RNA, which is different from hosts. Because viral RdRp has no homologue in humans, it is a promising drug target for the development of safe and effective anti-ZIKV drugs and hence many potential anti-RdRp agents were discovered (Nascimento et al., 2021; Yao et al., 2021; Chen et al., 2021; Gharbi-Ayachi et al., 2020; Kumar et al., 2020; Wang et al., 2018; Lim et al., 2018, 2016; Amraiz et al., 2016).

In our previous studies of antiviral drug discovery, we recently demonstrated that some of the fused tricyclic derivatives (compounds of 1 to 28, Table 1) of indoline and 4-imidazolidinone are potent DENV inhibitors with NS5 RdRp inhibitory activities (Qian et al., 2022). On the basis that there are many similar properties of ZIKV and DENV, for our pursuing discovery of selective or broad-spectrum antiviral drugs in this study, we studied the anti-ZIKV activities and inhibitory activities to ZIKV NS5 RdRp of compounds of 1 to 28 to compare with their anti-DENV activities and inhibitory activities to DENV NS5 RdRp from the previous work. The results show that certain compounds are broad-spectrum inhibitors of DENV- and ZIKV-infection while distinctive compounds are selective ZIKV inhibitors or selective DENV inhibitors. Especially, compounds of 12, 17 and 28 are moderate inhibitors to ZIKV NS5 RdRp, which are partly associated with their potent anti-ZIKV activities. In contrast, 12 is inactive against DENV infection even though 12 is an effective inhibitor to DENV NS5 RdRp, while 17 is an active agent against DENV infection but is a very weak DENV NS5 RdRp inhibitor, and 28 is inactive against DENV infection and not a DENV NS5 RdRp inhibitor.

Table 1.

Structures of all tested compounds (Qian et al., 2022)

2. Materials and methods

2.1. Materials

The Asian ZIKV strain SZ-WIV01 used in this study was kindly provided by Prof. Bo Zhang (Wuhan Institute of Virology, Chinese Academy of Sciences, Wuhan, China). The African ZIKV strain MR766 was kindly donated by Associate Professor Cheng Tong (Xiamen University, Xiamen, China). ZIKV NS5 RdRp (accession number: AMA12085) Plasmid (pet21a-His-Zika-NS5C(M275-S898) [Duan et al., al.,2017] was a kindly gift from Prof. Shi Yi (Institute of Microbiology, Chinese Academy of Sciences, Beijing, China). The new compounds were synthesized by the lab of Prof. Guo-Chun Zhou (School of Pharmaceutical Sciences, Nanjing Tech University, Jiangsu, China). All compounds and positive drugs were dissolved in DMSO and diluted with medium (cell-based assay) or buffer (enzymatic assay) into gradient stock solutions before use, and final concentration of DMSO is 0.1% for each well. Electro-competent Escherichia coli BL21 (DE3)-April cells were obtained from Tsingke Biotechnology. The following reagents were purchased from Sigma: NaCl, MgCl₂, ZnCl₂, glycerol, imidazole, isopropyl-β-d-1-thiogalactopyranoside (IPTG), 1,4-dithiothreitol (DTT), LB medium (powder), ampicillin (used at 100 μg/mL) and MnCl₂, ammonium acetate, and Tris base. SYTO™9 green fluorescent dye and His-Pur TM Ni-NTA resin were purchased from Thermo Fisher. Tween 20, heparin, EDTA and oligonucleotides were also purchased from Sigma. The ssRNA of polyuridylic acid (poly-U) was obtained from Shanghai Yuanye Bio-Technology. ATP was purchased from Armresco. EDTA and oligonucleotides were purchased from Sigma-Aldrich.

2.2. Methods

2.2.1. Plaque assay

Vero cells were infected with ZIKV (SZ-VIV01 or MR766) (MOI = 0.5) at 37 °C for 2 h. The infected cells were washed twice with PBS, and then a mixture of DMEM supplemented with 4% FBS and 2% low-melting agarose (Amresco, USA) was overlaid on the cells, when it was completely cooled, invert the cell culture plate at 37 °C and continue to culture for 5 days in a 5% CO₂ cell incubator. Next, the cells were fixed with 4% paraformaldehyde for 15 min at 120 hour post-infection (hpi), washed twice with PBS and stained with 0.8% crystal violet for 10 min. Finally, an enzyme-linked fluorescence spot spectrometer (CTL, USA) was used for image acquisition. The compounds exhibiting more than 50% plaque inhibition at 20 μM concentration were considered to have anti-ZIKV activity.

2.2.2. Cell cytotoxicity assay

MTT assay was used to detect the toxicity of these compounds. Cells were seeded in 96-well plates and cultured overnight. Compounds with different concentrations were added and incubated for 72 h. Then, 20 μL of 5 mg/mL MTT (Sigma-Aldrich, USA) solution was added to each well and kept at 37 °C for 4 h. Subsequently, 100 μL of 12% SDS-50% DMF (Sigma-Aldrich, USA) solution was added and incubated overnight at 37 °C. After the crystalline formamidine was completely dissolved, the optical density (OD) value was measured using a microplate reader (BioTek, USA) with a wavelength of 570 nm and a reference wavelength of 630 nm.

2.2.3. Quantitative real-time polymerase chain reaction (qRT-PCR)

Cells were infected with ZIKV (MOI = 1) and then treated with gradient concentrations of testing compound. The viral loads in the supernatant and cells were detected by qRT-PCR at different time periods after infection. Then the RNA from culture supernatant was extracted using the EasyPure® Viral DNA/ RNA kit (TransGen Biotech, China) following the manufacturer's instructions. Intracellular RNA extraction was according to trizol method. A one-step qRT-PCR kit RNA-direct™ Realtime PCR Master Mix (TOYOBO, Japan) and TaqMan probe were used to quantify the viral RNA produced. The primers (NS5 ZIKV 1086F: 5′-CCGCTGCCCAACACAAG-3′ and NS5 ZIKV 1086R: 3′-TACAGACGTTTTCTTGCAATCACC-5′) and probe (5′-FAM-AGCCTACCTTGACAAGCAGTCAGACACTCAA-TAMRA-3′) were used for amplification of the ZIKV NS5 region (Dick et al., 1952). A standard curve of the serial dilutions was used to quantify the viral RNA yield.

2.2.4. Western blot analysis

Vero cells were infected with ZIKV (MOI = 1) at 37 °C for 2 h. Next, the different amounts of compounds 12, 17 and 28 were added to continue culturing for 48 h to extract the total protein in the cells. The total protein was separated in sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Subsequently, the protein was transferred to a polyvinylidene fluoride (PVDF) membrane and incubated with ZIKV E antibody (Sino Biological, China). A specific signal was presented with a chemiluminescent substrate. The antibody signal was recorded and quantified using the Tanon-5200 Multi Imaging System.

2.2.5. Immunofluorescence assay

Vero cells were seeded at 2 × 10⁵ cells/mL in a 24-well plate with cell slides, and 500 μL/well was cultured overnight in a 37 °C, 5% CO₂ incubator. ZIKV (MOI = 1) was added to infect at 37 °C for 2 h and then the cells were washed twice with PBS to remove the viruses and then added to 500 μL of compounds diluted in concentration gradients. At the same time, virus-only as a positive control and a medium-only as negative control were set. After culturing at 37 °C for 72 h, the cell culture supernatant was discarded, and 4% paraformaldehyde was added to fix the cells for 15 min at room temperature. Cells were washed twice with PBS, PBS containing 0.1% TritonX-100 was added to permeabilize cells at 4 °C for 5 min, then the cells were washed with PBS for 5 min, and 5% BSA was added to block for 30 min at room temperature. The cells were stained overnight at 4 °C using ZIKV E antibody (Sino Biological, China). Then, the Alexa Fluor®CY5 was added for 1 h in the dark. DAPI staining was used to delineate the nucleus of cells. Microscopic examinations were performed on a Leica DMI4000B Microsystem (Wetzlar, Germany), and an image was recorded using the system.

2.2.6. Virus binding assay

Virus binding assay was performed as previously described (Yao et al., 2021). Briefly, Vero cells were seeded onto a 12-well plate to 95–100% confluence. The pre-chilled cells were incubated with ZIKV SZ-WIV01 (MOI = 1) and 50 μM of ECGC (Carneiro et al., 2016), 5 μM and 10 μM of compound 12, 50 μM of Ribavirin (RBV) or carrier ZIKV control at 4 °C for 2 h, and washed with PBS to remove unbound viruses or residual compound. The infected cells were harvested and viral RNA was determined by qRT-PCR.

2.2.7. Time-of-drug addition assay

This assay was conducted as described previously (Yao et al., 2021), with some modifications. Vero cells were seeded in 12-well plates and cultured overnight. Next, the cells were incubated with ZIKV (MOI = 1) at 4 °C for 2 h, and washed twice with 4 °C pre-chilled PBS. Testing compound (20 μM) was added at 0, 1, 2, 3, 4, 6, 8, 10, 12 hpi, and then the remaining wells were added fresh medium and incubated at 37 °C until the end of the experiment. The cell supernatant was collected at 24 hpi and total RNA was quantified by qRT-PCR. The data were the mean (±SD) of three independent experiments with respect to ZIKV control.

2.2.8. Time of drug withdrawal

The viruses were added to the cells and adsorbed for 2 h at 4 °C, then the unbound viruses was washed away with pre-cooled PBS, and then testing compound (20 μM) in fresh medium was added to incubate at 37 °C. Withdrawal by washing of testing compound at 1, 2, 4, 6, 8, 12 hpi after dosing. After continued to culture in the maintenance medium for 24 h, the cell supernatant was collected for viral load detection. The data were the mean (±SD) of three independent experiments with respect to ZIKV control.

2.2.9. In vitro ZIKV RdRp activity assay

The ZIKV RdRp (accession number: AMA12085) was expressed and purified as reported (Saez-Alvarez et al., 2019; Xu et al., 2017). ZIKV RdRp was diluted with protein dialysis buffer to 1500 nM and stored at 4 °C for later use. In vitro RdRp polymerase activity assay was described previously (Qian et al., 2022; Saez-Alvarez et al., 2019), 1500 nM of purified RdRp protein, gradient diluted compounds of 12, 17, 28 and positive control drug Heparin (5 μM) (Saez-Alvarez et al., 2019), 2 mg/mL Poly-U, 20 mM ATP, 1 M Tris–HCl, 100 mg/mL BSA, 100 mM MnCl₂ was added to a 96-well PCR plate. After mixing, the reaction was performed at 30 °C in the dark for 60 min. The reaction was terminated with 25 mM EDTA, and then 10 μM Syto9 was added to react at room temperature for 5 min. RdRp protein activity was detected on a real-time fluorescence quantitative PCR instrument.

2.2.10. In vitro DENV RdRp activity assay

The DENV2 RdRp (accession number: QST06885) was expressed and purified, and the protein expression, purification and assay was described previously (Qian et al., 2022). DENV RdRp was diluted with protein 1 x PBS buffer to 1500 nM and stored at 4 °C for later use. Breifly, 1500 nM of purified RdRp protein, 50 μM of compounds of 12, 17 or 28 and 20 μM of positive control drug Hinokiflavone (HIN) (Coulerie et al., 2012), 2 mg/mL Poly-U, 20 mM ATP, 1 M Tris–HCl, 100 mg/mL BSA, 100 mM MnCl₂ was added to a 96-well PCR plate. After mixing, the assay was initiated by the addition of 1500 nM DENV RdRp and the fluorescence was recorded over 40 min at 30  °C. RdRp protein activity was detected on a real-time fluorescence quantitative PCR instrument.

2.3. Statistical analysis

The 50% effective concentration (EC₅₀) and 50% cytotoxic concentration (CC₅₀) of compounds were calculated according to the Reed & Muench method, The data and graphs were processed using GraphPad Prism8 software, expressed as mean ± standard deviation (Mean ± SD), comparison between groups and columns using independent t-test and one-way ANOVA, P < 0.05 was considered statistically significant.

2.4. Virtual screening assay

In this study, molecular docking was performed to detect the putative interaction between the compound and ZIKV RdRp protein. The structure of ZIKV NS5 RdRp protein was downloaded from the PBD database (PDB ID: 5U0C) (Zhao et al., 2017) and saved in pdb format. The structures of compounds of 12, 17 and 28 were drawn with ChemDraw, converted to 3D format with Chem3D, and saved as mol2 format. Autodock4.2.6 and Autodock vina1.2.0 (Trott and Olson, 2010) were used for molecular docking of the compounds and the protein, and Pymol and Discovery studio 2019 were used to visualize the docking results and analyze the docking results. Hydrophobic interaction was obtained by Ligplot 4.5.3.

3. Results

3.1. Activities of titled compounds against ZIKV replication and infection

Since thirteen of twenty-eight titled compounds are active against DENV (Qian et al., 2022), it is presumed that there are close numbers of active compounds against ZIKV in this study considering that ZIKV and DENV similarity in many aspects. As listed in Table 2, cytotoxicity assay of twenty-eight compounds by MTT was tested up to CC50 (50% toxic concentration to Vero cells) = 200 μM to Vero cells. It is valued that CC50 values of all of 28 compounds are higher than 100 μM and those of 21 compounds (including all of 6 compounds of 4-nitro aniline series) are higher than 200 μM. Initially, all of these compounds (20 μM) were primarily screened against ZIKV SZ-VIV01 (Asian strain) infection by plaque formation assay and the results (Figs. 1A and 1B) show that compared to the ZIKV control, twelve compounds of 11–17, 19, 20, and 26–28 are potent agents against ZIKV infection with quite significant differences (***P < 0.001) while six compounds of 1, 3, 5, 10, 21 and 23 express much weak anti-ZIKV activities with significant differences (**P < 0.01). Then, anti-ZIKV activities of twelve potent compounds were confirmed by RT-PCR detection for ZIKV RNA copies (Fig. 1C), which support the plaque formation results (Figs. 1A and 1B). Cell viability curves of these twelve active anti-ZIKV compounds are shown in Figs. 1D and 1E.

Table 2.

EC₅₀ (μM) ^a or inhibition rate ^b, CC₅₀ (μM) ^c and SI of titled compounds.

Cmpd	CC50 (μM)	Anti-ZIKV				Anti-DENV 2 h
		Plaque formation		RT-PCR		RT-PCR
		EC50 (rate)	SI	EC50	SI	EC50	SI j
1	> 200	(37%)	/ e	ND g	/	- i	/
2	> 200	NA d	/	ND	/	–	/
3	> 200	(51%)	/	ND	/	–	/
4	> 200	NA	/	ND	/	–	/
5	> 200	(43%)	/	ND	/	–	/
6	> 200	NA	/	ND	/	–	/
7	> 200	NA	/	ND	/	–	/
8	> 200	NA	/	ND	/	5.94 ± 0.71	> 33.6
9	> 200	NA	/	ND	/	6.22 ± 0.22	> 32.1
10	> 200	(40%)	/	ND	/	–	/
11	> 200	10.56 ± 0.26	> 18.9	ND	/	7.81 ± 0.56	> 25.6
12	> 200	4.20 ± 2.35	> 47.6	2.91 ± 0.53	> 68.7	–	/
		13.61 ± 1.78 f	> 14.6 f	/	/
13	> 200	4.06 ± 3.31	> 49.2	2.87 ± 1.36	> 69.6	9.35 ± 3.03	> 21.4
14	158.89 ± 1.54	6.48 ± 2.35	24.5	8.44 ± 0.49	18.8	10.85 ± 0.11	14.6
15	186.39 ± 5.74	5.21 ± 1.36	35.7	5.12 ± 0.89	36.4	4.79 ± 0.82	> 38.9
16	> 200	2.97 ± 2.51	> 67.3	2.53 ± 0.17	> 79.0	5.64 ± 0.72	> 35.4
17	> 200	1.78 ± 1.97	> 112.3	1.69 ± 0.49	> 118.3	6.67 ± 1.78	> 30.0
		12.42 ± 0.02 f	> 11.1 f	/	/
18	> 200	NA	/	ND	/	–	/
19	186.09 ± 9.83	7.07 ± 3.26	26.3	ND	/	10.97 ± 0.63	> 16.9
20	122.57 ± 7.28	9.80 ± 4.13	12.5	ND	/	8.08 ± 1.24	> 13.9
21	> 200	(56%)	/	ND	/	–	/
22	> 200	NA	/	ND	/	–	/
23	129.10 ± 5.69	(61%)	/	ND	/	–	/
24	> 200	NA	/	ND	/	–	/
25	> 200	NA	/	ND	/	7.29 ± 0.63	> 27.4
26	117.92 ± 7.82	4.79 ± 1.42	24.6	3.78 ± 1.91	31.1	5.59 ± 0.52	> 21.1
27	> 200	5.25 ± 1.62	> 38.0	3.01 ± 0.12	> 66.4	5.97 ± 0.49	> 33.5
28	145.56 ± 5.53	4.02 ± 3.16	36.2	2.75 ± 2.49	52.9	–	/
		7.36 ± 2.07 f	19.7 f	/	/
RBV	> 200	47.88 ± 4.68	> 4.1	37.15 ± 4.73	> 5.38	46.78 ± 5.12	> 4.27
		49.88 ± 0.06 f	> 4.0 f	/	/

^aExperiments of plaque formation assay or RT-PCR assay conducted independently in triplicate. Asian ZIKV SZ-VIV01 strain was used in all experiments except African ZIKV MR766 strain was also used for plaque formation assay for compounds of 12, 17, 28 and ribavirin.

^bInhibition rate in parenthesis by 20 μM of tested compound.

^cMTT experiments conducted independently in triplicate.

^d“NA” represents “20 μM of testing compound expresses less than 30% inhibition of ZIKV replication”.

^e“/” means “not available”.

^ffrican ZIKV MR766 strain was used.

^g“ND” means “not detected”.

^hData from ref (Qian et al., 2022).

ⁱ“-“ represents “50 μM of testing compound expresses less than 50% inhibition of DENV2 replication”.

^jSI was calculated based on new CC₅₀ in this work.

Fig. 1.

Primary screening of twenty-eight compounds (20 μM) against ZIKV SZ-VIV01 (Asian strain) infection and cytotoxicity assay of twelve active compounds. (A) Antiviral effects of twenty-eight compounds on the formation of ZIKV plaques; (B) Plaque assay to detect twelve active compounds (20 μM) against ZIKV infection; (C) RT-PCR method to detect anti-ZIKV activity of twelve active compounds (20 μM); (D) and (E) Cytotoxicity assay of twelve anti-ZIKV active compounds to Vero cells; the data is the mean (±SD) of three experiments with respect to ZIKV control; carrier (0.1% DMSO) without ZIKV inoculation as Mock control, and carrier (0.1% DMSO) with ZIKV inoculation as ZIKV control, RBV (20 μM) as a positive control. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Based on above primary screening results, the anti-ZIKV activities of twelve potent compounds (****P < 0.0001) were submitted to determine EC₅₀ (50% effective concentration against ZIKV infection) values with gradient concentrations of each compound by ZIKV plaque formation assay (Fig. 2A and 2B) and RT-PCR quantification of ZIKV RNA copies (Fig. 2C and 2D). Anti-ZIKV activities of titled compounds are summarized in Table 2 and the overall results (Table 2) show that anti-ZIKV activities of these compounds show quite resemblance to those of anti-DENV even though there are some differences. Among twelve active compounds against ZIKV infection, two compounds of 12 and 28 are only active against ZIKV infection while ten compounds of 11, 13–17, 19, 20, 26, 27 are active against both of ZIKV and DENV. Eight compounds of 11, 13, 14, 16, 17, 19, 26 and 27 are more potent against ZIKV infection than against DENV infection but only two compounds of 15 and 20 from them are more potent inhibitors against DENV than against ZIKV. Meanwhile, compounds of 8, 9 and 25 exhibit selective inhibition against DENV infection. Furthermore, we conducted plaque formation assays of three selected compounds of 12, 17, 28 and ribavirin against African ZIKV MR766 strain. The results (Table 2) showed that while the positive drug of ribavirin exhibits very close inhibitory potency against both ZIKV strains, compounds of 12, 17 and 28 are active against MR766 strain infection but their anti-MR766 strain activities are less potent than their anti-SZ-VIV01 strain activities.

Fig. 2.

The anti-ZIKV activity of twelve active compounds based on gradient concentrations. (A) and (B) Antiviral effects of twelve active compounds on ZIKV RNA replication, the data are the mean (±SD) of three experiments with respect to ZIKV control; (C) and (D) Twelve active compounds’ inhibition of ZIKV plaque formation, the data are the mean (±SD) of three experiments with respect to ZIKV control; the data is the mean (±SD) of three experiments with respect to ZIKV control; carrier (0.1% DMSO) with ZIKV inoculation as ZIKV control, RBV (20 μM) as a positive control. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

3.2. Structure-cytotoxicity relationship of titled compounds

As active compounds’ SI (selectivity index, CC₅₀/EC₅₀) values are higher than 13 (Table 2) in this study, compounds’ anti-ZIKV activities are not obviously associated with their cytotoxicities. To effectively direct further modification, structure-cytotoxicity relationship of titled compounds was analysed, which can provide useful information. Apart from twenty-one compounds having higher than 200 μM of CC₅₀ values, seven compounds show their CC₅₀ values are less than 200 μM, which are six active anti-ZIKV compounds of 14 and 15 of 4-nitrophenyl hydrazine series and 19, 20, 26 and 28 of phenyl hydrazine series, and one inactive compound 23 of phenyl hydrazine series, informing that five out of seven compounds are belonging to phenyl hydrazine series. Compounds of 19, 20 and 26 of phenyl hydrazine series are more toxic than their corresponding counterparts of 8, 9 and 15 of 4-nitrophenyl hydrazine series, and 2, 3 and 6 of 4-nitro aniline series, respectively. Comparison of two diphenols of 12 and 23, and two 3′-OTBS-4′-OMe analogs of 17 and 28, discloses less cytotoxicities of 12 and 17 of 4-nitrophenyl hydrazine series than those of 23 and 28 of phenyl hydrazine series; meanwhile, increasing cytotoxicities from 6, 15 and 26 of three 4′-OTBS compounds were observed. It can be deduced from these data that the scaffold of phenyl hydrazine series is more likely to produce cytotoxicity than that of 4-nitrophenyl hydrazine series and 4-nitro aniline series.

Whereas, 14 of 4-nitrophenyl hydrazine series is more cytotoxic than 25 of phenyl hydrazine series, which is different from the cytotoxicitiy variation of corresponding silylated compounds of 17 and 28, indicating that the cytotoxicity derived from nitro substitution at 4′’-position depends on substituents and their combinations. Increment of cytotoxicity by conversion of phenol 9 to silylated 15 or phenol 25 to silylated 28 may be due to more lipophilicity change; whereas, silylated 17 is less cytotoxic than its corresponding phenol 14, CC₅₀ values of phenol 3 of 4-nitro aniline series and its corresponding silylated 6 are higher than 200 μM, cytotoxicities of phenol 20 and its silylated product 26 are at the same level, revealing that the effect of silylation on the cytotoxicity varies with the basic structural features.

3.3. Analysis of SAR against ZIKV and DENV infections

As summarized in Table 2, twelve compounds of 11–17, 19, 20, and 26–28 are discovered to be potent anti-ZIKV agents, which is helpful for Structure-activity relationship (SAR) analysis. In addition, six compounds of 1, 3, 5, 10, 21 and 23 reach higher than 30% to marginally higher than 50% inhibition rate of ZIKV infection at 20 μM, which can provide supplementary SAR information.

Linear indoline carboxylic amide 1 of 4-nitro aniline series exhibits 37% inhibition rate but other two linear indoline carboxylic hydrazine derivatives of 7 of 4-nitrophenyl hydrazine series and 18 of phenyl hydrazine series express much lower than 30% inhibition rates (Fig. 1A), suggesting that linear scaffold of amide is potentially superior to that of hydrazine counterparts.

Then, comparison of cyclic analogs of 2 with 8 and 19 bearing 4′-NO₂, or 3 with 9 and 20 containing 4′-OH, or 4 with 10 and 21 linking 3-(2′-furyl), or 5 with 11 and 22 having 3′-OMe and 4′-OH, or 6 with 15 and 26 connecting 4′-OTBS is valued as follows. Compound 2 of 4-nitro aniline series is inactive to ZIKV and DENV infection, 8 of 4-nitrophenyl hydrazine series is inactive to ZIKV infection but active against DENV infection, and 19 of phenyl hydrazine series is active agent against ZIKV and DENV infection. Compound 3 of 4-nitro aniline series reaches about 51% inhibition rate at 20 μM to ZIKV infection and no inhibition to DENV infection, 9 of 4-nitrophenyl hydrazine series is inactive to ZKV infection and active against DENV infection, while 20 of phenyl hydrazine series is active against ZIKV and DENV infection. Two 2-furyl analogs of 10 of 4-nitrophenyl hydrazine series and 21 of phenyl hydrazine series show 40% and 56% inhibition rates, respectively, but the inhibition rate of 2-furyl analog of 4 of 4-nitro aniline series is less than 10% (Fig. 1A). Compound 5 of 4-nitro aniline series exhibits 43% inhibition to ZIKV infection and no inhibition to DENV infection, 22 of phenyl hydrazine series is not active against ZIKV and DENV infection but 11 of 4-nitrophenyl hydrazine series is active against ZIKV and DENV infection. Silylated 6 of 4-nitro aniline series does not exhibit inhibitory activity to ZIKV infection but silylation makes 15 of 4-nitrophenyl hydrazine series and 26 of phenyl hydrazine series be active against ZIKV infection and both of 15 and 26 also are active against DENV infection. These observations disclose that cyclic scaffolds of 4-nitrophenyl hydrazine series and phenyl hydrazine series are more privileged than cyclic 4-nitro aniline series.

Next, two cyclic phenol derivatives of 3 (4′-OH) and 5 (3′-OMe-4′-OH) of 4-nitro aniline series express 51% and 43% inhibition rates at 20 μM, respectively, and non-phenol or phenol-blocked cyclic derivatives of 2 (4′-nitro), 4 (3-(2-furyl)) and 6 (4′-OTBS) of the same series are no inhibition to ZIKV infection. For 4-nitrophenyl hydrazine series, 4′-nitro analog of 8 and 4′-OH analog of 9 are not anti-ZIKV agents but they are potent and selective inhibitors to DENV infection; meanwhile, 4′-nitro analog of 19 and 4′-OH analog of 20 of phenyl hydrazine series are broad-spectrum inhibitors to ZIKV and DENV but their anti-DENV activities are slightly less active than their corresponding counterparts of 8 and 9. OH and OMe substituted at 3′- or 4′-position are appropriate groups for 4-nitrophenyl hydrazine series in that compounds of 11–14 are active anti-ZIKV agents. Among them, compound 14 of 3′-OH and 4′-OMe is less active anti-ZIKV agent and slightly more toxic than other compounds of 11–13; on the other hand, diphenol 12 exhibits no inhibition to DENV infection and is a potent and selective inhibitor to ZIKV infection but 11, 13 and 14 are brad-spectrum inhibitors to both of ZIKV and DENV. Different from its corresponding counterparts of 11–14, compounds of 22–25 of phenyl hydrazine series by substitutions of OH and OMe at 3′- or 4′-position are not anti-ZIKV agents and only 3′-OH-4′-OMe compound 25 shows selective anti-DENV activity as 8 and 9 of 4-nitrophenyl hydrazine series. Moreover, diphenol 23 of phenyl hydrazine series shows only 61% inhibition rate at 20 μM but diphenol 12 of 4-nitrophenyl hydrazine series is active against ZIKV infection (EC₅₀ = 4.2 μM). Interestingly, six silylated analogs of 15–17 of 4-nitrophenyl hydrazine series and 26–28 of phenyl hydrazine series are potent agents against ZIKV infection and compound 28 was discovered to be inactive against DENV infection (Qian et al., 2022) as compound 12. Furthermore, these silylated compounds are more active against ZIKV infection than their corresponding phenols of 9, 11, 14 and 20, 22, 25 (phenols of 9, 22 and 25 are inactive anti-ZIKV compounds!), respectively, and compound 17 is the most potent anti-ZIKV agent in this study, indicating that anti-ZIKV activities of these silylated analogs are not derived from possible hydrolysis of OTBS. Generally, silylation enhances anti-ZIKV and anti-DENV activities of 4-nitrophenyl hydrazine series and phenyl hydrazine series except 28 without anti-DENV activity; however, silylation actually decreases antiviral activities of 4-nitro aniline series that 4′-OTBS derivative of 6 is inactive against ZIKV and DENV infection even its corresponding phenol 3 possesses 51% inhibition rate at 20 μM against ZIKV infection. These data reveal that among cyclic compounds, 4-nitrophenyl hydrazine series seems superior to phenyl hydrazine series against ZIKV infection.

As shown in Table 2, three selected compounds of 12, 17 and 28 are active against Asian and African ZIKV strain infection but they have the antiviral selectivity between ZIKV strains, that anti-Asian ZIKV SZ-VIV01 strain is 3.2, 7.0 or 1.8 times more active than anti-African ZIKV MR766 strain by 12, 17 or 28, respectively.

Considering that antiviral activity, cytotoxicity, selectivity and novelty of structure and activity, 12 and 28 as selective ZIKV inhibitors and 17 as a broad-spectrum inhibitor to ZIKV and DENV and the most active ZIKV inhibitor were selected for further study in this work.

3.4. Active compounds can effectively inhibit the expression of ZIKV E protein

In order to further disclose whether or not the titled active compounds inhibit ZIKV infection and replication, inhibition of ZIKV E protein expression by compounds of 12, 17 and 28 was evaluated by Western Blot and immunofluorescence analysis using the antibody of ZIKV E protein to represent the changes of ZIKV E protein, which reflect the progression of ZIKV replication and infection (Abraham and Wood, 2022; Li et al., 2021; Rockstroh et al., 2017). Western Blot results (Fig. 3A-3D) show that there is strong concentration-dependent inverse relationship between three titled compounds of 12, 17, 28 and ribavirin with ZIKV E protein expression, which ZIKV E protein expression was significantly inhibited by 5 μM of 12 (Fig. 3A), 2.5 μM of 17 (Fig. 3B), 10 μM of 28 (Fig. 3C) or 50 μM of ribavirin (Fig. 3D). EC₅₀ value of 12 from Western-Blot results is very close to those from RT-PCR and plaque formation; meanwhile, EC₅₀ values of 17 and 28 from Western-Blot results are slightly higher than those from their corresponding RT-PCR and plaque formation. Immunofluorescence detection (Fig. 3E) of anti-ZIKV effects of 12, 17 and 28 at 20 μM reveals that 12, 17 and 28 strongly inhibit the expression of ZIKV E protein while almost no reduction of viable cells compared to the Mock control; meanwhile, the same concentration (20 μM) of RBV cannot fully abrogate ZIKV E protein expression and slightly less living cells were observed than the Mock control. These data support that compounds of 12, 17 and 28 can effectively inhibit ZIKV replication and infection.

Fig. 3.

Compounds of 12, 17 and 28 inhibit the expression of ZIKV E protein. Western Blot detection of inhibition of ZIKV E protein expression by 12 (A), 17 (B), 28 (C) and Ribavirin (D); (E) Immunofluorescence detection of inhibition of ZIKV E protein expression by 12, 17 and 28. The data were the mean (±SD) of three experiments (except for (D) one experiment) with respect to ZIKV control; RBV as a positive drug control, carrier (0.1% DMSO) without ZIKV inoculation as Mock control, and carrier (0.1% DMSO) with ZIKV inoculation as ZIKV control.

3.5. Compound 12 majorly acts on RNA synthesis stage of ZIKV life cycle

In order to understand further details of antiviral mode of active compounds against ZIKV infection, binding assay for entry, time-of-addition and time-of-withdrawal were carried out to identify the time window in ZIKV viral cycle in which life cycle our compounds act on (Fig. 4). As we were interested in ZIKV inhibition in this study, considering that 17 is a broad-spectrum inhibitor to both ZIKV and DENV while 12 and 28 are selective agents against ZIKV and 12 is less toxic than 28, compound 12 was selected as a representative compound to study the antiviral modes of action. Firstly, compound 12 was tested its ability to interfere ZIKV's binding/absorbing to the surface of Vero cells that ZIKV and epigallocatechin gallate (EGCG) (Carneiro et al., 2016), compound 12 or RBV were incubated with Vero cells for 2 h at 4 °C, during which period viruses only bind onto host cells but cannot enter into the cells. The binding assay results (Fig. 4A) indicated that 5 and 10 μM of 12 and 20 μM of RBV cannot prevent ZIKV from adsorbing on the surface of Vero cells while 50 μM of EGCG effectively interferes the binding of ZIKV to the surface of Vero cells.

Fig. 4.

Compound 12′s action on ZIKV life cycle. (A) and (B) Binding assay of 12′s deterrence of ZIKV attachment onto the surface of host Vero cells and the time diagram of binding assay experiments, EGCG (50 μM) used as positive binding compound and RBV (20 μM) used as negative compound; (C) and (D) Time-sharing addition assay of 12 (20 μM), and the assay diagram of time-sharing addition experiments; (E) and (F) Time-sharing withdrawal assay of 12 (20 μM), and the assay diagram of time-sharing addition experiments. The data were the mean (±SD) of three independent experiments with respect to ZIKV control; carrier (0.1% DMSO) with ZIKV inoculation as ZIKV control. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Then, time of addition was conducted (Fig. 4C and 4D). After ZIKV inoculation of Vero cells at 4 °C for 2 h, post-treatment was conducted at pre-determined time points of addition of 12 to examine at which time point, the inhibition effect of 12 exerts to make the viral replication inhibited. As shown in Fig. 4C and 4D, compared with ZIKV control, post-treatment of 12 at 0, 1, 2, 3, 4, 6 and 8 hpi makes ZIKV RNA copies significantly decreased. As 12′s continuous presence from addition to detection, ranges of drug action at 0–2 hpi cover at the entry stage, the virus translation stage and the RNA synthesis stage, those at 2–6 hpi go through the virus translation stage and the RNA synthesis stage while those at 6–12 hpi belong to the RNA synthesis stage (Hou et al., 2017; Chambers et al., 1990; Qing et al., 2009; Kato et al., 2016). Therefore, the observation that no synergistic effect or combined effect by post-treatment of 12 at 0, 1 and 2 hpi on inhibition of ZIKV replication signifies that 12 is not majorly targeting the entry stage. Meanwhile, post-treatments of 12 at 3, 4, 6 and 8 hpi make inhibition of ZIKV replication at the close level, meaning that inhibition of the virus translation stage is also unlikely the major event. Whereas, post-treatment of 12 at 10 hpi at the late RNA synthesis stage only exhibits mild inhibition and post-treatment of 12 at 12 hpi at matured ZIKV progeny stage does not show significantly inhibitory activity. These data ascertain that the RNA synthesis stage by post-treatment is confirmed inhibition stage. Intriguingly, it is difficult to explain why inhibition by 12 is gradient increased along with addition time from 0 to 4 hpi. It is also unknown that addition of 12 at 6 hpi during transition period of the translation stage into the RNA synthesis stage engenders less activity (**P < 0.01) against ZIKV than at 4 hpi of the translation stage and 8 hpi of the RNA synthesis stage.

Later on, time of withdrawal assay (Fig. 4E and 4F) was conducted that drug withdrawal of 12 at 1 hpi does not show obvious inhibition of the entry stage of ZIKV replication and at 2 hpi exhibits weak inhibition (*P < 0.05) of the entry stage, these results are parallel to those of Fig. 4C. Withdrawal at 4 hpi shows significant inhibition of ZIKV replication which reflects the overall inhibition effects of the entry stage and the viral translation stage, indicating that the inhibition by 12 of viral translation stage plays ignorable role. Withdrawal of 12 at 6 and 8 hpi, in which time windows cover the translation stage and the RNA synthesis stage, exhibits more potent inhibition of ZIKV replication and shows significant difference with withdrawal at 4 hpi, indicating that 12′s inhibition of the RNA synthesis stage is the most important event. Anti-ZIKV activity by withdrawal of 12 at 12 hpi results in synergistic effect of inhibition at RNA synthesis stage and matured ZIKV progeny stage even though inhibition by 12 at matured ZIKV progeny stage is marginally active (Fig. 4C, 10 hpi).

3.6. Anti-RdRp activities of 12, 17 and 28 are part of their inhibitory activities to ZIKV infection

As compound 12 was demonstrated to inhibit ZIKV RNA synthesis stage in this study as above, it is valued and interesting to explore whether or not 12, 17 and 28 inhibit ZIKV NS5 RdRp activity. When purified ZIKV RdRp synthesizes dsRNA using Poly-U as ssRNA template and ATP as nucleotide source, inhibition assay of RdRp activity by these compounds was measured by a real-time assay based on the bound fluorescent quantity of fluorescent dye SYTO9 (Saez-Alvarez et al., 2019), which binds dsRNA but not ssRNA template molecules. The assay readout means less synthesis of dsRNA leading to more fluorescence, indicating that testing compound holds more inhibitory activity to RdRp. As shown in Fig. 5, the inhibitory effects of 12, 17 and 28 on ZIKV NS5 RdRp activity are significant and three compounds directly inhibit RdRp activity in a dose-dependent manner with IC₅₀ values of 8.52, 9.89 and 10.67 μM, respectively, supporting that these compounds’ anti-ZIKV activities are associated with their inhibitory activities against RdRp function. As their relatively low anti-RdRp activities, these compounds of 12, 17 and 28 very likely have other inhibitory mode/s to achieve the inhibition of ZIKV infection. Elucidation of other drug target/s of these compounds will be studied in the future.

Fig. 5.

Inhibition of ZIKV RdRp activity by compounds of 12 (A), 17 (B) and 28 (C) in concentration-dependent manner. The data are the mean (±SD) of three experiments, with respective to DMSO control, Heparin (5 μM) (Saez-Alvarez et al., 2019) as a positive drug control. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

3.7. Docking analysis of inhibition of ZIKV RdRp activity by compounds of 12, 17 and 28

In order to understand whether and how compounds of 12, 17 and 28 bind to and interact with the structural features of ZIKV NS5 RdRp, we analysed the putative binding and interaction between 12, 17 or 28 with RdRp domain of ZIKV NS5 by docking analysis. The docking model was generated from the RdRp domain of ZIKV NS5 (PDB code: 5U0C) (Zhao et al., 2017) by using Autodock 4.2.6 and Autodock Vina 1.2.0 and the docking results are shown in Fig. 6.

Fig. 6.

Putative binding of 12, 17 and 29 to ZIKV NS5 RdRp (PDB ID: 5U0C) (Zhao et al., 2017) by docking analysis. (A) 12 interacts with ZIKV RdRp palm domain and finger domain; (B) 17 interacts with ZIKV RdRp palm domain; (C) 28 interacts with ZIKV RdRp palm.

As shown in Fig. 6A, compound 12 and RdRp establish two conventional hydrogen bonds by exocyclic N—H with Ala469, 3′-OH of phenol with Asp535; one Pi-alkyl interaction by phenyl ring (attached by 4″-nitro and 1″-hydrazine) with Arg473; one Pi-sulfur interaction by phenyl ring of indoline with Cys711; and one Pi-anion interaction by phenyl ring (attached to C3) with Asp666. Interactions of RdRp with compound 17 (Fig. 6B), belonging to the same class as 12, form one conventional hydrogen bond by exocyclic N—H and Ser472; one carbon hydrogen bond by one hydrogen of 4′–OCH₃ and Asp665, another carbon hydrogen bond by oxygen of carbonyl and Lys359; one Pi-donor hydrogen bond by phenyl ring of indoline and Ala474; two Pi-alkyl interactions by Arg473 with phenyl ring (attached to 4″-nitro and hydrazine) and 3-phenyl ring; three alkyl-alkyl interactions by Ile475 and Arg473 with methyl of OTBS, and by pyrroline from indolie and Lys359. Compound 28 also has strong interactions with RdRp, in that there is one conventional hydrogen bond by exocyclic N—H and Gln605, one carbon hydrogen bond by oxygen from carbonyl and Gly604, one Pi-sigma interaction between phenyl ring from indoline with Thr608, one alkyl-alky interaction by methyl from 3′-OTBS with Ile799. Taken together, orientation of whole compound is more or less affected by 4″-nitro and bulky TBS. As compound 17 is 4-nitrophenylhydrazine series and bearing 3′-OTBS, its orientation has resemblance in some respects to 12 of 4-nitrophenylhydrazine series, and a little similarity to 28, which contains 3′-OTBS. These putative interactions could make 12, 17 and 28 tightly binding to RdRp.

3.8. Activity differentiation of compounds of 12, 17 and 28 against ZIKV, DENV and their RdRp enzymes

Combined data of this work and previous results show that compound 17 is a broad-spectrum agent against ZIKV and DENV infection while compounds of 12 and 28 are potent agents against ZIKV but they are not DENV inhibitors. In addition, compounds of 12, 17 and 28 possess moderate inhibitory activities against ZIKV NS5 RdRp. Considering that ZIKV and DENV have many similar aspects and share close activity of many viral inhibitors and target inhibitors, we wondered whether 12, 17 and 28 have inhibitory activity to DENV NS5 RdRp.

Inhibition of DENV NS5 RdRp was assayed as previously reported and the results were shown in Fig. 7A. Compound 28 has almost no inhibition to DENV NS5 RdRp activity, in accordance with the fact that 28 is a selective compound against ZIKV infection; however, compound 17 exhibits very weak inhibitory activity against DENV NS5 RdRp, implying that 17′s anti-DENV activity is not mainly associated with its inhibition of RdRp function. Surprisingly, 12 was discovered to be an active DENV NS5 RdRp inhibitor (Fig. 7A) even though 12 is not an anti-DENV agent. Then, anti-DENV assay of plaque formation by 12 was conducted again and the repeated result (Fig. 7B) confirmed that 12 is really not an anti-DENV agent, showing that 12′s inhibitory activity towards DENV RdRp does not make 12 be an active agent against DENV infection. Reasons for 12′s antiviral activity differentiation against DENV and ZIKV may originated from 12′s selective permeability or/and some viral specific factors making 12 metabolic alteration between DENV and ZIKV, which makes 12 no chance to execute its inhibitory activity to DENV2 NS5 RdRp in cell-based anti-DENV assay. Importantly, compounds of 17 and 28 displaying different inhibitory activities to ZIKV RdRp and DENV RdRp illustrates there are critical functional differences between ZIKV RdRp and DENV RdRp; meanwhile, compound 12′s inhibitory activities to ZIKV RdRp and DENV2 RdRp reveals that some of their basic functions are similar. Furthermore, anti-ZIKV activities of compounds of 12, 17 and 28 are related with their inhibitory activities to ZIKV RdRp but their inhibitory activities to DENV RdRp are not the main sources for their anti-DENV activities.

Fig. 7.

Inhibition of DENV NS5 RdRp by compounds of 12, 17 and 28 and compound 12′s anti-DENV-2 infection. (A) Compounds of 12, 17 and 28 inhibit DENV NS5 RdRp activity based on the relationship of time-fluorescence (the bound fluorescent quantity of SYTO 9), HIN (Coulerie et al., 2012) is a positive drug control, NC is the negative control with substrate only, and RdRp (DMSO) is the DMSO control; the data are the mean (± SD) of three experiments, P values are calculated with respect to DMSO control. (B) Repeated plaque formation assay of compound 12 was conducted to inhibit DENV2 infection; carrier (0.1% DMSO) without DENV inoculation as Mock control, and carrier (0.1% DMSO) with DENV inoculation as DENV control.

4. Discussion

ZIKV and DENV, arboviruses transmitted by the Aedes aegypti and Aedes albopictus mosquitos, belong to the Flavivirus genus. Since currently no vaccine is available against more recently ZIKV and there is one elusive DENV vaccine, while the diseases by DENV and ZIKV infection does not exist effective drug treatment, these flaviviruses made big endemic medical burden problems in recent years. One of major problems comes from damaging host neural system by ZIKV infection. As ZIKV infection is closely related to neonatal microcephaly and adult Guillain-Barre syndrome (Calvet et al., 2016; Bhagat et al., 2021; Mlakar et al., 2016), its alarming is always in some special regions. Until now, there is no detailed evidence how ZIKV conceals and invades the neural system, this makes it difficult to block ZIKV invasion of the neural system. Therefore, selective or/and broad-spectrum medications are urgently needed. Moreover, effective anti-ZIKV agents will help understand the mechanism of ZIKV life cycle, infection, and invasion of neural system.

Recent years many active compounds against DENV and/or ZIKV were reported, including repurposing medications (Nascimento et al., 2021; Rosa et al., 2020; Rampini et al., 2020; Munjal et al., 2017; Adcock et al., 2017; Gardinali et al., 2020; Deng et al., 2016), natural products and synthesized compounds (Coronado et al., 2018; Albulescu et al., 2017; Tan et al., 2017; Rausch et al., 2017; Zhou et al., 2021; Cirne-Santos et al., 2021; Cataneoet al., 2021; Wardana et al., 2021; Patel and Gulick, 2021; Felicetti et al., 2020; Wang et al., 2019; Yu et al., 2022; Samrat et al., 2022; Nunes et al., 2022; García et al., 2017); however, there is still no effective drug for the treatment of ZIKV and DENV infection in the clinic, illustrating there is a long way to go for effective treatment of ZIKV and related viruses as the extreme difficulty of R&D for anti-DENV and anti-ZIKV. On the other hand, even though none inhibitors are currently in clinical development, these reported inhibitors actually reveal useful information and provide enlightening envisage for further modification by diversification and derivatization and pave a way for future successful drug R&D of anti-ZIKV and anti-DENV. As RdRp and NS2B-NS3^pro are the most promising targets, RdRp inhibitors (Nascimento et al., 2021; Tan et al., 2019; Choi and Rossmann, 2009; Ackermann et al., 2001; Kao et al., 2001) and NS3 protease inhibitors (Coronado et al., 2018; Albulescu et al., 2017; Patel and Gulick, 2021) are more focused research fields. On the basis of our former exploration of the fused bicyclic derivative of pyrrolidine and 4-imidazolidinone (Weng et al., 2017) and their functionalized derivatives (Jaratsittisin et al., 2020), and the fused tricyclic derivatives of 1,2,4,5-tetrahydroimidazo[1,5-a]quinolin-3(3aH)-one (Xu et al., 2020) as anti-flavivirus agents, we diversified structures to design indoline derivatives of 1 to 28 (Table 1) and discovered some of them to be anti-DENV agents with DENV NS5 RdRp inhibitory activity (Qian et al., 2022). In this study, we further worked on their anti-ZIKV and anti-ZIKV RdRp activities of these compounds of 1 to 28. The results (Table 2) show that among 28 compounds, twelve compounds are potent anti-ZIKV agents (***P < 0.001 to the control) and six compounds express limited anti-ZIKV infection activity (**P < 0.01 to the control). Among them, ten compounds of 11, 13–17, 19, 20, 26, 27 are broad-spectrum agents and active against viral infection by ZIKV and DENV; while eight broad-spectrum compounds of 11, 13, 14, 16, 17, 19, 26 and 27 show more potent activities against ZIKV infection than against DENV infection. Meanwhile, two compounds of 12 and 28 are selective anti-ZIKV inhibitors and three compounds of 8, 9 and 25 exhibit selective inhibition against DENV infection. As compounds of 12, 17 and 28 exhibit inhibition of anti-ZIKV NS5 RdRp, their anti-ZIKV activities are at least partly associated with their inhibitory activities to ZIKV RdRp; on the contrary, compound 17 is a weak DENV NS5 RdRp inhibitor which is almost not responsible to its anti-DNV activity while 28 is not a DENV NS5 RdRp inhibitor and not an anti-DENV agent. Activity differentiation of 28 against ZIKV and DENV infection is parallel to that of its respective ZIKV RdRp and DENV RdRp; however, antiviral differentiation of 12 does not stem from 12′s action of ZIKV RdRp and DENV RdRp, which compound 12 is inactive against DENV infection even though 12 is an active DENV NS5 RdRp inhibitor. It is unknown whether 12′s selective permeability or/and some viral specific factors makes antiviral activity alteration between DENV and ZIKV.

To sum up above, some key structural features hold common/sharing characteristics while others can make antiviral differentiation between ZIKV and DENV, and/or anti-RdRp differentiation between ZIKV RdRp and DENV RdRp. On the other hand, as their anti-RdRp activities cannot fully accomplish their antiviral potencies, these compounds of 12, 17 and 28 quite possibly are multi-targeting antiviral agents. Moreover, compounds of 12, 17 and 28 are more active against Asian ZIKV SZ-VIV01 strain than African ZIKV MR766 strain, indicating that these compounds also have antiviral selectivity between ZIKV strains. These data suggest that different epidemic ZIKV strains need more specific medications; therefore, it is possible to serve precision treatment by this series of compounds against various flaviviruses and their strains and variants. Further elucidation of dominant factor/s for the mode of action is important.

TBS is not be a good druggable pharmacophore but it is a good pilot of hydrophobic construction unit and sometimes silylation can enhance antiviral activity from reported observation (Vernekar et al., 2015) and our previous results (Xu et al., 2020). In this work, five silylated analogs of 15–17 of 4-nitrophenyl hydrazine series and 26–27 of phenyl hydrazine series are more active against ZIKV and DENV infection than their corresponding phenols. Silylated 28 is more active against ZIKV infection than its inactive free phenol 25; contrarily, 28 was discovered to be inactive against DENV infection but its respective phenol 25 is active against DENV infection. These observations reveal that the inhibitory activity of TBS-protected compound is not because of its hydrolysis into its corresponding free phenol even though it is possible to have a partial hydrolysis in long time. These results are useful information for how to introduce hydrophobic pharmacophores in future modification and optimization.

Time-of-addition and time-of-withdrawal experiments of 12 identified RNA synthesis stage of post-treatment to be the most key time window to interfere ZIKV viral cycle by 12 (Fig. 4). These observations support the results that 12, 17 and 28 are RdRp inhibitors. Besides, drug actions on entry stage, virus translation stage and matured ZIKV progeny stage are very weak but cannot be completely excluded. It is puzzled that inhibition by 12 is gradient increased along with drug addition time from 0 to 4 hpi while addition of 12 at 6 hpi engenders less activity (**P < 0.01) against ZIKV than at 4 hpi and 8 hpi. These data will help for our further search of the cellular target/s.

In summary, the fused tricyclic derivatives of indoline and imidazolidinone were discovered to be valuable leads for drug discovery of anti-ZIKV and anti-DENV. Structural features of a compound play important roles in executing its antiviral differentiation between ZIKV and DENV and anti-RdRp differentiation between ZIKV RdRp and DENV RdRp. On the basis of the current promising core structural element of the active compounds, the further discovery of highly potent broad-spectrum or selective agents against ZIKV and DENV infection and elucidation of major drug targets of these compounds can be expected.

Author contributions statement

Guang-Feng Zhou: expression of ZIKV NS5 RdRp. Guang-Feng Zhou, Jian-Xia Xue, Xue-Rong Gu, Chang-Bo Zheng and Chunyan Li: virology research work. Liu-Meng Yang: supervision of virology research work. Feng Li and Weiyi Qian: design and synthesis of compounds. Guo-Chun Zhou: supervision of design and synthesis of compounds. Si-Dong Xiong, Guo-Chun Zhou and Yong-Tang Zheng: Conceptualization and manuscript preparation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

• Data will be made available on request.