A role for glycolipid biosynthesis in severe fever with thrombocytopenia syndrome virus entry

Abstract

A novel bunyavirus was recently found to cause severe febrile illness with high mortality in agricultural regions of China, Japan, and South Korea. This virus, named severe fever with thrombocytopenia syndrome virus (SFTSV), represents a new group within the Phlebovirus genus of the Bunyaviridae. Little is known about the viral entry requirements beyond showing dependence on dynamin and endosomal acidification. A haploid forward genetic screen was performed to identify host cell requirements for SFTSV entry. The screen identified dependence on glucosylceramide synthase (ugcg), the enzyme responsible for initiating de novo glycosphingolipid biosynthesis. Genetic and pharmacological approaches confirmed that UGCG expression and enzymatic activity were required for efficient SFTSV entry. Furthermore, inhibition of UGCG affected a post-internalization stage of SFTSV entry, leading to the accumulation of virus particles in enlarged cytoplasmic structures, suggesting impaired trafficking and/or fusion of viral and host membranes. These findings specify a role for glucosylceramide in SFTSV entry and provide a novel target for antiviral therapies.

Fig 1. Haploid genetic screen identifies SFTSV entry factors.

(A) Overview of haploid genetic screen. Details provided in Materials and Methods. (B) Parental (non-mutagenized) HAP1 cells and mutagenized HAP1 cells surviving rVSV-SFTSV selection (HAP1-SFTSV^R) were subject to a single-cycle infection (8–10hr) with a panel of recombinant vesicular stomatitis viruses (rVSVs) encoding various viral glycoproteins. Infection in HAP1-SFTSV^R cells is normalized to infection in the parental HAP1 cells. SFTSV = severe fever with thrombocytopenia syndrome virus, ANDV = Andes virus, EBOV = Ebola virus. (C) Statistical Enrichment Analysis was carried out using the Chi-Square Exact Test with false discovery rate correction. Each dot represents a unique gene identified from integration site mapping in the HAP1-SFTSV^R population. The size of the dot, and number in parenthesis when provided, reflects the number of unique integration sites found within that particular gene. For clarity, only genes with 3 or more unique integration sites are plotted. (D) Schematic of glucosylceramide synthesis. GlcCer is synthesized by the membrane-bound UGCG on the cytosolic face of the cis-Golgi. The polar head of GlcCer is then flipped to the lumenal side by the transporter MDR1, where it can be further modified by lactosylceramide synthase (LCS) to form lactosylceramide, the core building block of most cellular glycosphingolipids.

Fig 2. RNAi confirms role for UGCG in SFTSV entry.

(A) U-2 OS cells were transfected with non-targeting (Neg Ctrl) or UGCG siRNAs and 72 hours post-transfection were infected with rVSV-SFTSV or VSV. Ten hours post-infection cells were harvested, immunostained for VSV M expression, and analyzed by flow cytometry. Infection levels are normalized to the negative control siRNA and expressed as relative infection. Mean ± S.E.M. for 3 independent experiments. (B) RNA was collected in parallel at the time of infection from the siRNA treated cells in (A). RT-qPCR was performed and UGCG mRNA was normalized to GAPDH mRNA and expressed relative to the negative control siRNA. Western Blot for UGCG protein expression shown in the inset. (C) U-2 OS cells were transduced with lentiviral shRNA constructs containing scrambled or UGCG-targeting shRNAs. Following puromycin selection, cells were infected with rVSV-SFTSV, VSV, or SV40 and harvested 10 hours (rVSV-SFTSV, VSV) or 24 hours (SV40) post-infection, immunostained for viral antigen, and quantified by flow cytometry. Mean ± S.E.M. for 2 independent experiments. (D) HAP1 UGCG KO cells or parental HAP1 cells were plated in 24 well plates and the next day infected with rVSV-SFTSV or VSV. Ten hours later, cells were collected, immunostained for VSV M expression, and analyzed by flow cytometry. Infection levels are normalized to parental HAP1 cells and expressed as relative infection. Mean ± S.E.M. for 3 independent experiments. (E) U-2 OS cells transfected with negative control or UGCG siRNAs were infected with rVSV-SFTSV, VSV, SV40 or Rift Valley fever virus (RVFV) 72 hours post-transfection. Cells were harvested 10 hours (rVSV-SFTSV, VSV, RVFV) or 24 hours (SV40) post-infection, immunostained for viral antigen, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. ** p<0.01, *** p<0.001, **** p<0.0001 using Student’s t-test with Bonferroni correction.

Fig 3. Pharmacological inhibitors of UGCG activity inhibit rVSV-SFTSV entry.

(A) A549 cells were pre-treated with the UGCG inhibitor D,L-threo-PDMP for 24 hours before infection with rVSV-SFTSV. Drug was kept in the media throughout the infection. Ten hours post-infection cells were harvested, immunostained for VSV M, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. (B) A549 cells were pre-treated with the UGCG inhibitor N-butyldeoxynojirimycin-HCl (NB-DNJ) for 24, 48, or 72 hours before infection with rVSV-SFTSV. Drug was kept in the media throughout the infection. Ten hours post-infection, cells were harvested, immunostained for VSV M, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. (C,D) A549 cells were pre-treated with NB-DNJ (C) or N-(n-Butyl) deoxygalactonojirimycin (NB-DGJ) (D) for 48 hours before infection with rVSV-SFTSV, VSV, SV40, or RVFV. Drug was kept in the media throughout infection. Ten hours (rVSV-SFTSV, VSV, RVFV) or 24 hours (SV40) post-infection, cells were harvested, immunostained for viral antigen, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. **** p<0.0001 using Student’s t-test with Bonferroni correction.

Fig 4. Analysis of downstream glycosphingolipid biosynthesis.

(A) Diagram of Ganglioside Biosynthesis. Unique enzymes are responsible for the formation of GlcCer, LacCer, GM3 and GD3 (left hand side). These substrates are then modified by common enzymes in a step-wise manner to attach additional sugar moieties. Each of the o-, a-, and b-series branches are a committed pathway, since sialyltransferases cannot convert an a-series ganglioside to a b-series (excluding the conversion of GM3 to GD3). (B) U-2 OS cells were transfected with siRNAs targeting UGCG, GM3 Synthase (GM3S), GD3 Synthase (GD3S), or a non-targeting control (Neg Ctrl) and 72 hours later were infected with rVSV-SFTSV, VSV, or SV40. Ten hours (rVSV-SFTSV, VSV) or 24 hours (SV40) post-infection cells were harvested, immunostained for viral antigen, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. (C) mRNA levels for UGCG, GM3S, and GD3S were determined for (B) using RT-qPCR, normalized to GAPDH mRNA levels, and expressed relative to the negative control siRNA. (D) U-2 OS cells were transfected with siRNAs targeting UGCG, B4GALT5, B4GALT6, both B4GALT5 and B4GALT6 (B4GALT5/6), or a non-targeting control (Neg Ctrl). 72 hours post-transfection cells were infected with rVSV-SFTSV, VSV, or SV40. Ten hours (rVSV-SFTSV, VSV) or 24 hours (SV40) post-infection cells were harvested, immunostained for viral antigen, and analyzed by flow cytometry. Mean ± S.E.M. for 3 independent experiments. (E) mRNA levels for UGCG, B4GALT5, and B4GALT6 from siRNA-transfected cells were determined by RT-qPCR, normalized to GAPDH mRNA levels, and expressed relative to the negative control siRNA. *** p<0.001, **** p<0.0001 using Student’s t-test with Bonferroni correction.

Fig 5. UGCG activity and expression important for wild-type SFTSV infection.

(A) U-2 OS cells were transfected with siRNAs to UGCG or a non-targeting control and 72 hours later were infected with SFTSV (strain HB29, MOI 0.1). Supernatants from infected cells were collected 48 hours post-infection and viral output was determined by plaque assay. Infectious output is expressed relative to the negative control siRNA. (B) U-2 OS cells were pre-treated with NB-DNJ for 48 hours before infection with SFTSV (MOI 5). Supernatants from infected cells were collected 12 hours post-infection and viral output was determined by plaque assay. Infectious output is expressed relative to the untreated control. *** p<0.001, **** p<0.0001 using Student’s t-test with Bonferroni correction.

Fig 6. Heartland virus, close relative of SFTSV, also requires UGCG for efficient entry.

(A) Phylogenetic tree based on the amino acid alignments of various phlebovirus glycoproteins (polyprotein precursor). Scale bar represents genetic distance as a percent difference (0.2 = 20%). Asterisks denote virus sequences used to generate codon-optimized M segments (G_N-G_C) for the present study. (B) U-2 OS cells were transfected with siRNAs to UGCG or a non-targeting control, then 72 hours later were infected with rVSV-SFTSV or the VSV pseudotype bearing the Heartland virus glycoprotein (VSV-HRTV). Twelve hours post-infection cells were harvested, immunostained for VSV M, and analyzed by flow cytometry. Percent infection was normalized to samples receiving the negative control siRNA and expressed as relative infection. Mean ± S.E.M. for 3 independent experiments. (C) U-2 OS cells were pre-treated with NB-DNJ for 48 hours before infection with rVSV-SFTSV, VSV-HRTV or VSV. Drug was kept in the media throughout the infection. Twelve hours post-infection cells were harvested, immunostained for VSV M, and analyzed by flow cytometry. Percent infection was normalized to the untreated control for each virus and expressed as relative infection. Mean ± S.E.M. for 3 independent experiments. *** p<0.001 using Student’s t-test with Bonferroni correction.

Fig 7. Mechanistic studies on UGCG’s role in SFTSV entry.

(A) Binding and Internalization Assay. U-2 OS cells were transfected with negative control or UGCG siRNAs, replated the following day into 24 wells dishes, and the assay was performed 72 hours post-transfection. Assay details provided in Materials and Methods. vRNA levels were normalized to GAPDH mRNA levels, and are expressed relative to bound vRNA for the negative control siRNA. Mean ± S.E.M. for 3 independent experiments. (B) RNA collected from (A) was also analyzed for UGCG mRNA expression. UGCG mRNA levels were measured by RT-qPCR, normalized to GAPDH mRNA levels, and expressed relative to the negative control siRNA (bound). (C,D) The binding and internalization assay was carried out essentially as in (A) with the exception that U-2 OS cells were instead pre-treated with NB-DNJ for 48 hours prior to binding with rVSV-SFTSV (C) or SV40 (D) and qPCR for SV40 genomes did not require reverse transcription. Mean ± S.E.M. for 3 independent experiments. ** p<0.01 using Student’s t-test.

Fig 8. Immunofluorescence microscopy of incoming virus particles, 20 minutes.

(A, B) A549 cells were plated onto glass coverslips and the following day replaced with media containing NB-DNJ (200μM) or left untreated. Forty-eight hours later, cells were chilled to 4°C on ice, then rVSV-SFTSV was bound by centrifugation (1200xg, 30’, 4°C). Following centrifugation, media was replaced with pre-warmed media (37°C) and the cells placed in a 37°C incubator for 20 minutes before fixation in 2% paraformaldehyde for 10 minutes. Cells were then immunostained for viral antigen (anti-VSV M, red), cellular markers (green), and nuclei stained with DAPI (blue). Images are representative from at least 3 independent experiments. (A) A549 cells co-stained for rVSV-SFTSV (red) and early endosome marker EEA1 (green). (B) A549 cells co-stained for rVSV-SFTSV (red) and trans-Golgi marker TGN46 (green). Boxes indicate zoomed-in regions. Scale bar represents 5μm.

Fig 9. Immunofluorescence microscopy and quantification of incoming virus particles.

(A, B) A549 cells prepared as described in Fig 8 were fixed 40 minutes post-warming and co-stained for rVSV-SFTSV (red) and the early endosome marker EEA1 (green) (A) or TGN46 (B). Cells were treated with NB-DNJ (bottom panels) or left untreated (top panels). Boxes indicate zoomed-in regions. Scale bar represents 5μm. (C, D) Quantitative image analysis was performed to measure the volume of discrete VSV M-stained puncta within z-stack images in untreated and NB-DNJ treated cells at both 20 and 40 minutes post-warming. Puncta were counted for at least 6 independent z-stacks per sample for both rVSV-SFTSV (C) and VSV (D) infected cells. (**** p<0.0001 using Welch’s one-tailed t-test).

Fig 10. Immunofluorescence microscopy of rVSV-SFTSV particles following UGCG knockdown.

(A-D) U-2 OS cells were transfected with siRNAs targeting UGCG (siUGCG) or a non-targeting control (siNegCtrl) and plated onto glass coverslips. At 72 hours post-transfection cells were chilled to 4°C on ice and rVSV-SFTSV was bound by centrifugation (1200xg, 30’, 4°C). Following centrifugation, media was replaced with pre-warmed media (37°C) and the cells placed in a 37°C incubator for 20 or 40 minutes before fixation in 1% paraformaldehyde for 15 minutes. Cells were then immunostained for viral antigen (anti-VSV M, red), cellular markers (green), and nuclei stained with DAPI (blue). Images are representative from at least 3 independent experiments. (A,B) U-2 OS cells fixed after 20 minutes were co-stained for rVSV-SFTSV (red) and EEA1 (A) or TGN46 (B) (green). (C,D) U-2 OS cells fixed after 40 minutes and stained as above. Boxes indicate zoomed-in regions. Scale bar represents 5μm.