SREBP-dependent lipidomic reprogramming as a broad-spectrum antiviral target

Abstract

Viruses are obligate intracellular microbes that exploit the host metabolic machineries to meet their biosynthetic demands, making these host pathways potential therapeutic targets. Here, by exploring a lipid library, we show that AM580, a retinoid derivative and RAR-α agonist, is highly potent in interrupting the life cycle of diverse viruses including Middle East respiratory syndrome coronavirus and influenza A virus. Using click chemistry, the overexpressed sterol regulatory element binding protein (SREBP) is shown to interact with AM580, which accounts for its broad-spectrum antiviral activity. Mechanistic studies pinpoint multiple SREBP proteolytic processes and SREBP-regulated lipid biosynthesis pathways, including the downstream viral protein palmitoylation and double-membrane vesicles formation, that are indispensable for virus replication. Collectively, our study identifies a basic lipogenic transactivation event with broad relevance to human viral infections and represents SREBP as a potential target for the development of broad-spectrum antiviral strategies.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Integrative transcriptomic–lipidomic analysis and lipid screening. Calu-3 cells were mock infected or infected with MERS-CoV at 2 MOI and incubated in DMEM medium. At 24 hpi, cells were harvested for transcriptomic (a) and lipidomic (b) analysis, respectively. a Gene Ontology (GO) analysis of differentially expressed genes (DEGs) in the MERS-CoV infected and non-infected Calu-3 cells. DEGs were classified under six categories as indicated. Red arrow indicates lipid metabolism as the top-ranking affected metabolism pathway. b Heatmap showing the lipidomic analysis of MERS-CoV-infected vs non-infected Calu-3 cells. Each rectangle represents a lipid colored by its normalized intensity scale from blue (decreased level) to red (increased level). The hierarchical clustering analysis was based on the identified lipid metabolites with significant changes in quantity. PC phosphatidylcholines, PA phosphatidic acid, PS phosphatidylserine, PC(P−) 1-(1z-alkenyl),2-acyl-phosphatidylcholine, PE phosphatidylethanolamine, PE(P−) 1-(1z-alkenyl),2-acyl-phosphatidylethanolamine, PG phosphatidylglycerols, PI phosphatidylinositol, MGDG monogalactosyldiacylglycerol, FA fatty acids, FAHFA fatty esters, SM sphingomyelin, Cer ceramides. c Integrated transcriptomic and lipidomic analysis. Both enrichment (blue) and topological (yellow) analysis are scored, indicating the glycerophospholipid metabolism as the most affected pathway after MERS-CoV infection. The analysis was performed by MetaboAnalyst 4.0. d Scatter plot showing the lipid library screening after MERS-CoV (black dots) or influenza A(H1N1) (red dots) infection. Cell viability with 0.1 MOI and 24 h post-MERS-CoV infection, and with 0.01 MOI and 48 h post-H1N1 infection were selected as the end-point of determination after drug treatment. Shown is the normalized result by setting mock-infection as 100%, which is averaged from three independent screenings

Fig. 2

In vitro antiviral activity of AM580. a Multi-cycle MERS-CoV growth assay in the presence or absence of AM580. Huh7 cells were infected with MERS-CoV (0.01 MOI). Viral titers in cell culture supernatants were quantified by plaque assay at different time points. Differences between DMSO (open circle) and AM580 (20 µM, black square) groups were analyzed by Student’s t-test. b Western blot showed reduced MERS-CoV NP production after AM580 treatment. Huh7 cells were infected with 1 MOI MERS-CoV. c Upper panel: MERS-CoV-NP-positive cells quantitated by flow cytometry. Lower panel: immunofluorescence staining of MERS-CoV-NP antigen (green) and cell nucleus (blue). Scale bar:10 µm. d AM580 reduced MERS-CoV replication in cell culture supernatants of A549 (0.1 MOI), Calu-3 (0.1 MOI), and Vero (0.01 MOI) cells at 24 hpi. AM580 reduced MERS-CoV replication in cell lysates of monocyte-derived macrophage cells (MDM, 1 MOI) and THP-1 (0.1 MOI) at 24 hpi. Differences between DMSO and AM580 groups were analyzed by Student’s t-test. e AM580 inhibited MERS-CoV replication in human primary small airway epithelial cells (HSAEC) that were infected by 1 MOI MERS-CoV and treated with (red square) or without AM580 (black triangle). Supernatant and cell lysate were collected at the indicated time points and titrated by plaque assay and RT-qPCR assay, respectively. f, g AM580 showed anti-MERS-CoV activity in human intestinal organoid (intestinoid). f One-way ANOVA was used for comparison of the AM580 treated with the DMSO treated. g Representative images of intestinoids, after immunofluorescence staining for MERS-CoV NP (green), DAPI, and Phalloidin (purple), were 3D-imaged with a confocal microscope. Scale bar: 20 µm. h AM580 showed broad-spectrum antiviral effects against six different viruses as indicated. Plaque reduction assays were performed to evaluate antiviral activity of AM580 in MERS-CoV (Huh7 cells, magenta triangle) and SARS-CoV (Huh7 cells, magenta rectangle), ZIKV (Vero cells, black triangle), H1N1 virus (MDCK cells, black diamond), EV-A71 (RD cells, blue dot). TCID50 assays were used for AdV5 titration (HEp-2 cells, blue triangle). Shown are the PFU or TCID50 of indicated concentrations relative to controls in the absence of compound (%). The experiments were performed in triplicate and replicated twice. The results are shown as mean ± s.d. ***p < 0.001, **p < 0.01, *p < 0.05

Fig. 3

In vivo antiviral activity of AM580. a DDP4 transgenic mice were treated by intraperitoneal inoculation of AM580 (red square) or 0.1% DMSO (placebo control, black triangle) for 3 days starting 6 h post-challenge with 50PFU of MERS-CoV. Survivals and clinical disease were monitored for 14 days or until death. Differences in survival rates between groups were compared using Log-rank (Mantel–Cox) tests. b Daily body weights of surviving mice. Student’s t-test was used to compare different groups on each day post-infection. c Lung and brain tissues were collected for detection of viral yields using both plaque assay and RT-qPCR assays (Supplementary Figure 7). Difference were compared with DMSO-treated groups using Student’s t-test. d Representative histologic sections of lung tissues from the indicated groups with hematoxylin and eosin (H&E) staining. Greater alveolar damage and interstitial inflammatory infiltration were present in the DMSO group. e–h Balb/c mice were treated with intranasal AM580 (red square), zanamivir (positive control, blue dots), or 0.1% DMSO (negative control, black triangle) for 3 days starting 6 h post-challenge with 100 PFU of influenza A (H7N9) virus. Shown are survival rate (e), mean body weight (f), lung viral load (g), and representative lung sections stained by H&E (h). The same statistical analyses were performed as described in a–d. Results are presented as mean values ± s.d. ***p < 0.001,**p < 0.01, *p < 0.05. Scale bar: 20 µm

Fig. 4

SREBPs were essential for MERS-CoV replication. a AM580 decreased cellular lipid droplets (LDs) and cholesterol levels. Huh7 cells were infected with MERS-CoV at 0.01 MOI for 24 h, in the presence of 0.1% DMSO or 20 µM AM580 or mock-infected. Cells were fixed and stained with DAPI (blue) and BODIPY 493/503 lipid probe (green) for LDs detection, or with filipin (blue) for intracellular cholesterol visualization. Scale bar: 20 µm. b The expression of multiple genes in fatty acid (upper) and cholesterol (lower) synthesis pathways were analyzed by RT-qPCR, respectively. * indicates p < 0.05 comparing DMSO and AM580 by one-way ANOVA. c siRNA knockdown of either SREBP1 (two distinct siRNA 1_1 and 1_2) or 2 (siRNA 2_1 and 2_2) or both SREBP1 and 2 (siRNA 1_2 and 2_2) reduced MERS-CoV replication. The knockdown efficiency and specificity was evaluated with western blot. siRNA-treated cells (100 nM siRNA for single knockdown and 50 nM each for double knockdown for 48 h) were infected with MERS-CoV (0.001 MOI and 24 hpi). Viral titer in the cell culture supernatant was evaluated with plaque assay. One-way ANOVA was used for comparison with the control siRNA pre-treated group. d Hyper-expression of nuclear form SREBP1 and SREBP2 by transfection to Huh7 cells for 24 h, respectively, followed by MERS-CoV infection and (1 MOI, 12 hpi) AM580 treatment as indicated. Overexpression of n-SREBP1 and n-SREBP2 were analyzed by western blot using anti-flag tag antibodies. Differences in viral titer were compared with the vector-transfected control and analyzed using Student’s t-test. e AM580 inhibited transactivation of lipogenic genes such as FAS (blue dots) and HMGCS (red square) but not LXRE (black triangle). Huh7 cells transfected with the indicated reporter plasmid were treated with serial-dilutions of AM580 for 24 h. For virus infection assays, AM580 was added after virus absorption; for luciferase assay, AM580 was added 6 h after plasmid transfection. Student’s t-test was used to compare the AM580-treated with DMSO-treated groups. The experiments were performed in triplicate and replicated twice. The results are shown as mean ± s.d. **p < 0.01, *p < 0.05

Fig. 5

AM580 interacted with n-SREBP to block lipogenic transactivation. a AM580 blocked n-SREBP1 and SRE binding. DNA-binding activity of nuclear-extracted SREBP1 (n-SREBP1) to double-strand DNA (a mimic of SRE) immobilized onto the wells of microtiter plates. Betulin, negative control; competitor dsDNA, positive control. One-way ANOVA was used for comparison with the DMSO-treated group. b AM580 bound with n-SREBP1 instead of SRE. Black line indicates AM580 was added and incubated in SRE-dsDNA-immobilized wells, washed before addition of n-SREBP1; red line with square indicates AM580 was pre-incubated with n-SREBP1 before adding to SRE-dsDNA-immobilized wells. Students’ t-test was done between groups with same concentrations of AM580 but different treatments. The experiments were performed in triplicate and replicated twice. The results are shown as mean ± s.d. **p < 0.01, *p < 0.05. c AM580 was predicted to occupy the SRE-recognition sites of both SREBP1 and 2. Shown is the 3D molecular docking analysis. Potential interaction surfaces on SREBPs (red) are shown, while AM580 (green) is displayed in stick and mesh representation. Partial sequence alignment of the DNA-binding domains of human and mouse SREBP1 and 2 is shown. Tyr335, the key residue for AM580 binding is highlighted with a box. d Structure of an AM580-derived probe (AM580dp) showing the locations of designated groups with specific functionalities. Azido-AM580 was synthesized through the reaction between azido-PEG5-amine and AM580. AM580dp was further synthesized through the addition of tri-functional crosslinker with azido-AM580. e Cellular distribution of azido-AM580. AM580 was used as a negative control due to the lack of phosphine-specific azido group. Scale bar:10 µm. f Tyr335 was critical for AM580 and n-SREBP1 interaction. AM580dp was immobilized on streptavidin beads and incubated with the cell lysate that were transfected with either WT or Y335R mutant constructs. After pull-down, western blot was employed to detect n-SREBP1 using both anti-n-SREBP1 and anti-Flag antibodies. Overexpression of RAR-α protein was used as a positive control for the pull-down capacity of AM580dp, while azido-AM580 was used as a negative control to exclude non-specific binding

Fig. 6

Suppressed SREBP-dependent pathways reduced virus replication. a Inhibition the fatty acid synthesis reduced virus replication, while exogenous palmitate or oleic acid rescued MERS-CoV replication. Huh7 cells infected with MERS-CoV (0.01 MOI) were treated with DMSO, or C75 (FAS inhibitor), or AM580 in the absence (white bars) or presence of supplemental exogenous palmitate (black bars) or oleic acid (red bars). Viral titers in culture supernatants after 24 hpi are shown. Differences between groups were analyzed by one-way ANOVA test. b–d Virus rescue assays were performed for H1N1 virus (0.001 MOI), EV-A71 (0.001 MOI), and human AdV5 (100TCID50) in different cell lines as described in a. Viral titer for different viruses were analyzed by Students’ t-test. *p < 0.05, **p < 0.01. The experiments were performed in triplicate and replicated twice for confirmation. The results are shown as mean ± s.d. e AM580 inhibited DMVs formation. Vero cells were infected with 3 MOI of MERS-CoV and treated with DMSO or AM580 for another 12 h before processing for the staining before transmission electron microscopy. Virus-infected cells without treatment showed perinuclear clusters of DMVs (red box) and the lack of DMVs production upon AM580 treatment. Representative electron microscopy images were selected from a pool of over 30 images captured in two separate experiments. Scale bar of upper panel: 1 µm, lower panel: 200 nm. f AM580 did not affect viral protein expression of nsp3 and nsp4. Huh7 cells were co-transfected with MERS-CoV nsp3 and nsp4 with flag tag, and treated with AM580 at 6 h post-transfection for another 24 h. Protein expression level was evaluated by western blot using β-actin as an internal control. g AM580 reduced viral protein palmitoylation. A549 cells were transfected with HA plasmid of H1N1 virus. Drug treatment with DMSO (0.1%), 5 µM 2-BP (positive control inhibitor), or 20 µM AM580 was carried out post-transfection, while cell lysates were harvested 24 h later. Total HA (input) and palmitoylated HA of different groups were analyzed using western blot. In all assays above, AM580 (20 µM) was added after virus absorption and maintained in the cell culture medium

Fig. 7

Schematic diagram showing the proteolytic processing of precursor SREBPs to generate n-SREBPs, downstream effector pathways (i.e. n-SREBP1-mediated fatty acid synthesis and n-SREBP2-mediated cholesterol synthesis), the positive control inhibitors and their pharmacological targets, and reporter gene constructs that were used in this study