Genome-scale CRISPR screen identifies TMEM41B as a multi-function host factor required for coronavirus replication

Abstract

Emerging coronaviruses (CoVs) pose a severe threat to human and animal health worldwide. To identify host factors required for CoV infection, we used α-CoV transmissible gastroenteritis virus (TGEV) as a model for genome-scale CRISPR knockout (KO) screening. Transmembrane protein 41B (TMEM41B) was found to be a bona fide host factor involved in infection by CoV and three additional virus families. We found that TMEM41B is critical for the internalization and early-stage replication of TGEV. Notably, our results also showed that cells lacking TMEM41B are unable to form the double-membrane vesicles necessary for TGEV replication, indicating that TMEM41B contributes to the formation of CoV replication organelles. Lastly, our data from a mouse infection model showed that the KO of this factor can strongly inhibit viral infection and delay the progression of a CoV disease. Our study revealed that targeting TMEM41B is a highly promising approach for the development of broad-spectrum anti-viral therapeutics.

Fig 1. Genome-wide CRISPR screening to identify genes associated with alpha-CoVs TGEV-induced cell death.

(A) Identification of TGEV host factors using the porcine genome-scale CRISPR/Cas9 knockout (PigGeCKO) library. Transformed PK-15-Cas9 cells were either mock-treated or challenged with TGEV (MOI = 0.001). Surviving cells from each round of virus challenge were isolated, with the subsequent PCR amplification and sequencing of sgRNA. (B-D) Scatter plots showing sgRNA-targeted sequence frequencies and the extent of enrichment in transformed PK-15-Cas9 cells (mock-treated versus TGEV infected) in three rounds of TGEV screening; (B) first, (C) second and (D) third, respectively. Counts_TGEV1st/Counts_TGEV2nd and Counts_TGEV3rd are the average reads from the paired-end sequencing for each round. Log2(fold change) is the median log 2 ratio between normalized sgRNA count of TGEV challenged and mock-treated populations. (E and F) Venn diagrams indicate the scope of overlapping enrichment in specific sgRNA targeting sequences for the three TGEV screening rounds amongst the top (E) ~0.1% and (F) ~0.5% of averaged sgRNAs reads. FACS, Fluorescence Activated Cell Sorting; TGEV, Transmissible gastroenteritis virus; sgRNA, small guide RNA.

Fig 2. TMEM41B is a host factor required for TGEV replication.

(A) Verification of candidate genes enriched in the genome-wide CRISPR screen using a second, focused-CRISPR library screen. Scatter plots compare sgRNA targeted sequence frequencies and the extent of enrichment in the transformed PK-15-Cas9 cells (mock-treated versus TGEV infected) for the third (TGEV3rd) and fifth (TGEV5th) rounds of TGEV screening. Counts_TGEV3rd and Counts_TGEV5th represent the average values of the read counts from paired-end sequencing. Log2(fold change) is the median log 2 ratio between normalized sgRNA count of TGEV challenged and mock-treated cell populations. (B) Quantification of virus infectivity (TCID₅₀) in culture supernatant collected 24 hpi from TGEV-infected (MOI = 0.1) WT and KO (ANPEP, TMEM41B, LPP, TAX1BP1, BARHL12) cell lines. (C) Immunofluorescence assays for detection of the TGEV N protein in WT cells and five selected genes (ANPEP, TMEM41B, LPP, TAX1BP1, BARHL12) KO cells following infection with TGEV (MOI = 0.1) at 24 hpi. Scale bar, 200 μm. (D) Western blot assay to detect the TGEV N protein expressed in TMEM41B KO and WT cells following infection with TGEV (MOI = 1) at 24 hpi. GAPDH used as an internal control gene. (E) Viral quantification by plaque assays. TMEM41B KO and WT cells were seeded into 24-well culture plates and infected with TGEV at different MOIs (0, 0.001, 0.01, 0.1, 1 and 10). Plates stained with 1% crystal violet to view plaques. (F) TMEM41B KO and WT cells were infected with TGEV at different MOIs (0.001 and 1), TGEV N copy number was assessed by absolute quantitative real-time PCR. (G and H) Rescue assays for WT, TMEM41B-KO and TMEM41B-KO-rescue cells infected with TGEV (MOI = 1). RT-qPCR assay for determination of (G) relative mRNA level of TMEM41B and (H) absolute mRNA level of TGEV N gene. TMEM41B-KO-rescue: reconstituted TMEM41B in TMEM41B KO cells. (I and J) Overexpression of TMEM41B in PK-15 control cells following infection with TGEV (MOI = 1). RT-qPCR assay for the determination of (I) relative mRNA level of TMEM41B and (J) absolute mRNA level of the TGEV N gene. PK-15-NTC: Transfection of pcDNA3.1 empty vector in WT cells; PK-15-overexpression: Transfection of pcDNA3.1-TMEM41B vector in WT cells. WT, wild-type; KO, knockout; MOI, multiplicity of infection; kDa, kilodaltons; DAPI, 4’,6-diamidino-2-phenylindole. **P < 0.01; ***P < 0.001; ****P <0.0001. P values were determined by two-tailed Student’s t-tests. Data are represented as a percentage mean titer of triplicate samples relative to WT control cells ± S.D.

Fig 3. TMEM41B is an intrinsic negative regulator of innate immune responses.

(A-D) Gene Set Enrichment Analysis (GSEA) for differentially expressed genes from the RNA-seq data shows a distinct upregulation of the interferon signaling pathway across the four pairwise groups (A) WT-MOCK vs WT-TGEV; (B) TMEM41B-KO-MOCK vs TMEM41B-KO-TGEV; (C) WT-MOCK vs TMEM41B-KO-MOCK; (D) WT-TGEV vs TMEM41B-KO-TGEV. (E) RT-qPCR validation of the mRNA expression of ISGs selected by RNA-seq results. (F) RT-qPCR outcome reveal that the stable rescue of TMEM41B in TMEM41B KO cells led to the significant suppression of genes associated with the interferon signaling pathway. (G) Viral titers for WT, TMEM41B-KO and TMEM41B/IFNAR double KO cells infected with TGEV. The TMEM41B/IFNAR double KO cells still maintain the ability to resist TGEV replication compared with TMEM41B KO cells as determined by the TCID₅₀ assay. WT, wild-type; KO, knockout; hpi, hours post-infection; MOCK, uninfected cells; TGEV, Transmissible gastroenteritis virus infected cells. *P < 0.05; **P < 0.01; ***P < 0.001. P values were determined by two-sided Student’s t-test. Data are representative of at least three independent experiments.

Fig 4. The internalization and early-stage replication of TGEV were impaired on TMEM41B KO cells.

(A) Evaluation of the adsorption activity of TGEV on TMEM41B KO and WT cells by absolute quantitative real-time PCR assay. WT and TMEM41B KO cells were infected with TGEV (MOI = 5) at 4°C for 1 h and assessed for TGEV adsorption. (B) Confocal microscopy assay for detection of TGEV N protein (Red) on WT and TMEM41B KO cells infected with TGEV (MOI = 5) at 4°C for 1 h. Scale bars = 5 μm. (C) Evaluation of the TGEV endocytosis stage in TMEM41B KO cells. WT and KO cells were infected with TGEV (MOI = 50) at 4°C for 1 h and transferred to 37°C for 30 min. Left: Confocal microscopy assay for detection of TGEV N protein (Red) expression in TMEM41B KO and WT cells. Right: Normalized mean fluorescence intensity was standardized by the ratio of each fluorescence intensity was divided by the maximum fluorescence intensity, n≧3. Scale bars = 10 μm. (D) RT-qPCR validation of mRNA expression of endocytic pathway genes enriched by RNA-seq. (E) Evaluation of the effects of TMEM41B KO cells on virus particle assembly by transmission electron microscope. Compared with WT cells, no virus-like particles (red arrows) wrapped in vesicles of varying sizes were found in TMEM41B KO cells. Scale bar, 2 μm or 500 nm as indicated. (F) Assessment of TGEV release stage of replication in TMEM41B KO and WT cells infected with TGEV (MOI = 5). Intracellular and extracellular viral titers were evaluated by virus TCID₅₀ assays at 24 hpi. (G and H) Confocal microscopy to evaluate early-stage TGEV replication by detecting (G) dsRNA formation and (H) TGEV N protein expression in WT and TMEM41B KO cells infected with TGEV (MOI = 5) at 3 hpi. WT, wild-type; KO, knock out; hpi, hours post-infection; MOI, multiplicity of infection; Mock, uninfected cells; TGEV, Transmissible gastroenteritis virus infected cells; N, Nucleus, dsRNA, double-stranded RNA; DAPI, 4’,6-diamidino-2-phenylindole. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, no significant. P values were determined by two-sided Student’s t-test. Data are representative of at least three independent experiments.

Fig 5. TMEM41B contributes to the formation of CoVs replication organelles.

(A-D) Representative images from transmission electron microscope of TGEV (MOI = 0.1, 24 h) induces membrane structures. There are typical DMVs (diameter ~200 nm) were observed (A and B). Electron micrograph of an area with abundant double-membrane spherule (DMSs) (C) and irregular coiled membrane formed by ERs expansion (D). Scale bar, 1μm or 500 nm. (E-F) Electron micrograph showing mock-infected and TGEV-infected (MOI = 0.1, 24 h) TMEM41B KO cells bearing several monolayer vesicles. Scale bar, 500 nm or 250 nm. (G) Confocal fluorescence microscopy analysis of the co-localization of TGEV-NSP3-HA (indicated in red) and restored expression of TMEM41B-FLAG (indicated in fuchsia) in TMEM41B KO cells. Scale bar, 5 μm. (H) Analysis of the subcellular location of ectopic expression TMEM41B in KO cells mock-infected (above) or infected with TGEV (MOI = 1, 12 h) (bottom). The restored expression of TMEM41B (indicated in red) were co-localization with dsRNA (indicated in green). Scale bar, 5 μm. (I) Assessment of the co-localization of dsRNA (indicated in green) and ectopic expression of TMEM41B-FLAG (indicated in red) in L929 cells during MHV (MOI = 1, 12 h) infection via confocal fluorescence microscopy. Scale bar, 5 μm. (J) A model for the role of TMEM41B in the formation of coronavirus replication organelles. WT, wild-type; KO, knock out; Mock, uninfected cells; TGEV, Transmissible gastroenteritis virus infected cells; MHV, Mouse hepatitis virus infected cells; M, Mitochondria; ER, Endoplasmic reticulum; DAPI, 4’,6-diamidino-2-phenylindole; HA, HA-Tag; FALG, 3×FLAG-Tag; dsRNA, double-stranded RNA.

Fig 6. TMEM41B is a host factor for the replication of multiple viruses.

(A) Alignment of TMEM41B amino acid sequences from pigs, human and mice. The accession Numbers of TMEM41B from NCBI protein database for Sus scrofa is XP_003129438.1 (pTMEM41B), for Homo sapiens is NP_055827.1 (hTMEM41B), and for Mus musculus is NP_705745.3 (mTMEM41B). Protein sequences were aligned using Constraint-based Multiple Alignment Tool (COBALT). The dark star indicates the regions of sequence conservation amongst the three species (letters indicated in red); the blue letters indicate the regions with sequence variation among the three species; the box indicates the VTT domain. (B) Rescue assays for WT, TMEM41B-KO and TMEM41B-KO-rescue L929 cells infected with MHV (MOI = 0.1, 24 hpi). Cell supernatant was collected and assessed for MHV titers using TCID₅₀. (C-F) Rescue assays for WT, TMEM41B-KO and TMEM41B-KO-rescue PK-15 cells infected with (C) PDCoV (MOI = 0.01, 18 hpi), (D) JEV (MOI = 0.1, 24hpi), (E) IAV-PR8 (MOI = 0.01, 24 hpi) and (F) VSV (MOI = 0.01, 18 hpi). The cell supernatant was collected and tittered using TCID₅₀ assays and immunofluorescence assays to detect NP positive cells. (G) Viral titers for TMEM41B-KO, TMEM41B/IFNAR double KO and TMEM41B/IFNAR double KO-TMEM41B-rescue cells infected with VSV (MOI = 0.1, 18 hpi) and cell supernatants was collected and titrated using TCID₅₀. (H) Immunofluorescence assays for TMEM41B-KO, TMEM41B/IFNAR double KO and TMEM41B/IFNAR double KO-TMEM41B-rescue cells infected with IAV (MOI = 0.1, 18 hpi) and detection the NP positive cells. WT, wild-type; KO, knockout; hpi, hours post-infection; MHV, Mouse hepatitis virus; PDCoV, porcine deltacoronavirus; JEV, Japanese encephalitis virus; IAV-PR8, PR8 influenza A virus; VSV, Vesicular stomatitis virus. **P < 0.01; ***P < 0.001; ****P <0.0001. P values were determined by two-sided Student’s t-test. Data are representative of at least three independent experiments.

Fig 7. TMEM41B is required for CoVs infection in vivo.

(A) Line graphs showing the body weight of TMEM41B^+/− and WT mice during the 3 days post-inoculation with MHV A59 (n = 6 mice per genotype). *P < 0.05; **P < 0.01; ***P < 0.001; ns, no significant. P-values were determined by two-way measure. (B) The total weight change of TMEM41B^+/− and WT mice during the 3 days post-inoculation of MHV A59 (n = 6 mice per genotype). **P < 0.01; ns, no significant. P-values were determined by one-way measure. (C) Survival curves for TMEM41B^+/− and WT mice infected with MHV A59 (n = 6 mice per genotype). (D and E) Histopathological analysis of the degree of liver damage in TMEM41B^+/− and WT mice infection with MHV A59 at 3 dpi. (D) Gross postmortem examination of liver tissue; (E) histological examination of Hematoxylin-eosin (H&E) stained liver tissue. (F) The virus titers of half livers of TMEM41B^+/− and WT mice were measured TCID₅₀ assay with L929 cells (n = 3 mice per genotype). **P < 0.01. P values were determined by two-sided Student’s t-test. (G) A model illustrating the roles of TMEM41B in the CoV replication cycle. In WT cells, CoV binds to its receptor and enters through clathrin- and caveolin-mediated endocytosis before releasing their genome into the cytoplasm after membrane fusion in the early/late endosome. Subsequently, CoVs Nsps (NSP3, NSP4, and NSP6) act with TMEM41B and other transmembrane host proteins in the formation of ROs. In TMEM41B KO cells, endocytosis is impaired, and the internalization of virions was reduced. Moreover, the formation of DMVs was blocked during membrane elongation leading to the inhibition of TGEV replication so that almost no virions were produced. DMVs, double-membrane vesicles; ER, Endoplasmic reticulum; Mock, uninfected mice; MHV-A549, A549 Mouse hepatitis virus infected mice. WT, wild-type; KO, knockout; TMEM41B^+/− mice used in Fig 7 were the F2 generation. Data are representative of at least three independent experiments. Source data are provided as a Source Data file. The experiments were repeated three times with similar results and representative results shown. Scale bar, 200 nm.