Amaryllidaceae Alkaloid Cherylline Inhibits the Replication of Dengue and Zika Viruses

Abstract

Dengue fever, caused by dengue virus (DENV), is the most prevalent arthropod-borne viral disease and is endemic in many tropical and subtropical parts of the world, with an increasing incidence in temperate regions. The closely related flavivirus Zika virus (ZIKV) can be transmitted vertically in utero and causes congenital Zika syndrome and other birth defects. In adults, ZIKV is associated with Guillain-Barré syndrome. There are no approved antiviral therapies against either virus. Effective antiviral compounds are urgently needed. Amaryllidaceae alkaloids (AAs) are a specific class of nitrogen-containing compounds produced by plants of the Amaryllidaceae family with numerous biological activities. Recently, the AA lycorine was shown to present strong antiflaviviral properties. Previously, we demonstrated that Crinum jagus contained lycorine and several alkaloids of the cherylline, crinine, and galanthamine types with unknown antiviral potential. In this study, we explored their biological activities. We show that C. jagus crude alkaloid extract inhibited DENV infection. Among the purified AAs, cherylline efficiently inhibited both DENV (50% effective concentration [EC₅₀], 8.8 μM) and ZIKV replication (EC₅₀, 20.3 μM) but had no effect on HIV-1 infection. Time-of-drug-addition and -removal experiments identified a postentry step as the one targeted by cherylline. Consistently, using subgenomic replicons and replication-defective genomes, we demonstrate that cherylline specifically hinders the viral RNA synthesis step but not viral translation. In conclusion, AAs are an underestimated source of antiflavivirus compounds, including the effective inhibitor cherylline, which could be optimized for new therapeutic approaches.

Keywords: Amaryllidaceae; Zika virus; alkaloids; antivirals; cherylline; dengue virus; flavivirus; lycorine.

FIG 1

Crinum jagus alkaloid extract’s anti-DENV activity. (A) Schematic experimental design of the antiviral assay using Crinum jagus bulbs and DENV, with DENV_GFP construct genome representation. In this system produced by Fischl and Bartenschlager (33), 103 nucleotides of DENV capsid gene were duplicated and cloned upstream of gfp, which was then fused to the capsid gene with a 2A peptide (82). Gfp is replicated as part of the viral genome and translated as a component of the polyprotein. The 2A peptide allows GFP to be released from DENV polyprotein during or following the translation process. (B) Inhibition of DENV_GFP infection with C. jagus alkaloid extract observed by inverted microscopy in Huh7 cells. Representative images are shown with cell nuclei stained with Hoechst33342 (blue) and DENV infection (green). (C) Anti-DENV activity of C. jagus crude extracts. The inhibition of DENV_GFP infection in Huh7 cells by C. jagus alkaloid extract dilutions was measured by flow cytometry. (D) Cytotoxicity of C. jagus bulbs crude alkaloid extract in Huh7 cells as measured by the XTT [2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide salt] assay.

FIG 2

Screening of the anti-DENV activity of C. jagus isolated Amaryllidaceae alkaloids (AAs). (A) Structures of the nine AAs isolated from bulbs of C. jagus. Cherylline-type alkaloids comprised cherylline, gigantelline, and gigantellinine, while crinine types comprised crinine, gigancrinine, and flexinine. Lycorine, sanguinine, and hippadine were also isolated. (B) Anti-DENV activity of AAs. Huh7 cells were treated with compounds (AAs or DMSO vehicle) for 2 h and then infected with DENV_GFP at an MOI of 0.15 for 72 h. Infectivity was visualized as GFP⁺ cells on an inverted microscope system with ×5 and ×20 (hippadine) objectives. The experiment was performed in triplicates at two dilutions. At least 3 pictures were taken from each well. One representative picture of the lowest dilution is displayed. Cher, cherylline; Giga, gigantelline; Gign, gigantellinine; Crin, crinine; Gigc, gigancrinine; Flex, flexinine; Lyco, lycorine; Sang, sanguinine; Hipp, hippadine.

FIG 3

Cherylline displays anti-DENV activity. (A) Impact of cherylline treatment on DENV_R2A (purple) and ZIKV_R2A (pink) replication was tested in Huh7.5 cells (MOI, 0.001), and viral-dependent luciferase luminescence was measured at 48 hpi. Cell viability (ATP) was assessed at 48 h. Results are displayed as fold changes in viability and replication, with 1 meaning no change compared to matched concentrations of DMSO-treated cells. (B) Treatment of Huh7 with cherylline dampened infectivity of DENV_GFP replication (MOI, 0.15) in a dose-dependent manner, as measured by flow cytometry 72 hpi. (Left) representative dot plots; (right) percentage of infected Huh7 cells using 0.6 to 100 μM cherylline and 10 to 500 μM ribavirin as a positive control. The nonlinear regression curve fit is shown with the 95% of confidence interval in finer lines. (C) Representative pictures of Huh7 treated with different concentrations of cherylline and infected with DENV_GFP at an MOI of 0.15. Ribavirin was used as a positive control, and DMSO as a negative control. Infected cells were visualized on an inverted microscope system at 72 hpi.

FIG 4

Cherylline displays antiviral activity against DENV and ZIKV. (A) Viral titers were measured by plaque assay on Vero E6 cells using the cytopathic WT DENV2 16881s strain. Huh7.5 cells were infected and treated 2 hpi with compounds. Then, 48 h later, supernatants were harvested and plated on Vero E6 cells (left). Fold changes in viability (ATP) were calculated at 48 hpi in Huh7.5 cells (right). (B) Viral titers were measured as in panel A on Vero E6 cells using cytopathic WT Uganda MR766 ZIKV strains (left) and French Polynesia H/PF/2013 (right). (C) THP-1 cells were treated with two concentrations of each compound in triplicates for 2 h and infected with VSV-G pseudotyped HIV-1GFP at an MOI of 1. The antiretroviral nevirapine was used as a positive control, and DMSO (vehicle), as negative a control. Results were analyzed by flow cytometry 72 hpi.

FIG 5

Cherylline blocks DENV replication during RNA synthesis. (A) Infectivity of DENV in Huh7 cells continuously treated with cherylline compared to cells treated only for 2 hpi. DMSO was used as a negative control, and ribavirin as a positive antiviral control. Representative pictures taken at ×5 with an inverted microscope system are shown. (B) Schematic explanation of time-of-drug-addition and -removal rationale. +, addition of compound; –, removal. (C) Time of drug addition. (D) Time of drug removal. For panels C and D, DENV_GFP was used at an MOI of 0.1; DMSO was used as a negative control, and the fold infectivity relative to the control was calculated; 1-fold means that the level of infection is the same as control. NITD008 (NITD) was used at 10 μΜ as a positive antiviral control targeting RNA synthesis. Cherylline was used at 50 μΜ. The impact of addition and removal of NITD and cherylline at 0, 2, 4, 7, 12, and 24 hpi on the percentage of DENV_GFP-infected Huh7 cells was monitored by flow cytometry at 72 hpi. Experiments was performed in triplicates twice. Means with the standard error of the mean (SEM) are shown. (E) Schematic representation of subgenomic replicon sg-DVs_R2A_WT and subgenome sg-DVs_R2A_GND used in panels F and G. Sgs were transfected into Huh7.5 cells, which were then treated with DMSO, NITD008 (5 μΜ), or cherylline (50 μΜ) at the indicated time. (F) Luciferase levels were quantified at 48 and 72 h postelectroporation (hpe) with sg_DVs_R2A_WT and normalized over levels detected in DMSO-treated cells. (G) Luciferase levels were quantified at 4 hpe with RdRp-deficient sg_DVs_R2A_GND and normalized over levels in cells treated with DMSO.

FIG 6

Cytotoxicity and antiviral effects of cherylline in PBMCs. (A) The cytotoxicity of antiviral concentrations of cherylline in PBMCs was measured by ATP production and normalized over DMSO at 48 h posttreatment. Means with SEM are shown. Dotted lines represent the concentration used in the antiviral assay. Ch, cherylline. (B) Antiviral activity of cherylline in PBMCs. PBMCs were infected with DENV_GFP preincubated with anti-Envelope antibody 4G2 at an MOI of 2, treated with cherylline (30 μM) or DMSO (0.05%), and analyzed 72 hpi by flow cytometry. The experiment was performed in triplicates (uninfected) and six-plicates (infected). Representative plots are shown.