PLSCR1 is a cell-autonomous defence factor against SARS-CoV-2 infection – Nature

Cell lines

Huh7.5 (human hepatocellular carcinoma, a gift from C.B.W.), A549-ACE2 (human alveolar basal epithelial carcinoma cells, BEI Resources NR-53821), Vero E6 (African green monkey kidney epithelial cells, ATCC CRL-1586), HEK293T (human embryonic kidney cells, ATCC CRL-3216), HeLa (human cervical adenocarcinoma cells, ATCC CCL-2), Tonsil (human tonsillar epithelial cells, UT-SCC-60A), HaCaT (immortalized human keratinocytes, a gift from D. DiMaio) and LET1 (mouse lung epithelial type I cells, BEI Resources NR-42941) cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin–streptomycin (pen-strep). hTEpiCs (ScienCell 3220) were cultured in bronchial epithelial cell medium (ScienCell 3211) supplemented with 1% bronchial epithelial cell growth supplement (ScienCell 3262). Calu-3 (human lung adenocarcinoma cells, ATCC HTB-55) were cultured in Eagle’s minimum essential medium (EMEM; ATCC 30-2003) with 10% FBS and 1% pen-strep. All cells, unless otherwise stated, were cultured at 37 °C and incubated with 5% CO₂.

HEK293T-ACE2, HeLa-ACE2, Tonsil-ACE2, HaCaT-ACE2 and LET1-ACE2 cells were generated by stably expressing human ACE2 in the aforementioned original cell lines. In brief, lentiviruses were packaged in HEK293T cells by transfecting the cells with pLV-EF1a-hACE2-Hygro, psPAX2 and VSVG. Forty-eight hours after transfection, the medium was collected, filtered through a 0.45-μm filter and added to target cells for 24 h. Cells were subsequently selected with hygromycin for 7 days before further treatments. ACE2 expression was tested by western blots as well as virus infection.

Antibodies and reagents

Rabbit anti-GAPDH monoclonal antibody (60004-1-Ig), mouse anti-GFP tag monoclonal antibody (66002-1-Ig), rabbit anti-PLSCR1 polyclonal antibody (11582-1-AP), mouse anti-Halo tag monoclonal antibody (28a8), rabbit anti-TMEM41B polyclonal antibody (29270-1-AP) and rabbit anti-IFITM3 polyclonal antibody (11714-1-AP) were obtained from Proteintech. Rabbit anti-Na,K-ATPase polyclonal antibody (3010S), rabbit anti-Flag tag monoclonal antibody (14793S) and rabbit anti-β-tubulin monoclonal antibody (2128S) were obtained from Cell Signaling. Goat anti-ACE2 polyclonal antibody (AF933) was purchased from R&D Systems. Mouse anti-PLSCR1 monoclonal antibody (MABS483), mouse anti-dsRNA monoclonal antibody (MABE1134), rabbit anti-TMEM16F polyclonal antibody (HPA038958), sheep anti-mouse IgG horseradish-peroxidase-conjugated secondary antibody (GENXA931-1ML) and sheep anti-rabbit IgG horseradish-peroxidase-conjugated secondary antibody (GENA934-1ML) were obtained from Sigma. Rabbit anti-SARS-CoV-2 nucleocapsid monoclonal antibody (40143-R019) and rabbit anti-SARS-CoV-2 spike S2 antibody (40590-T62) were purchased from Sino Biological. Mouse anti-EEA1 monoclonal antibody (610456) was obtained from BD Biosciences. Rabbit anti-LY6E polyclonal antibody (ab300399) was purchased from Abcam. Mouse monoclonal antibody against SARS-CoV-2 spike (GTX632604) was obtained from GeneTex. Donkey anti-goat IgG horseradish-peroxidase-conjugated secondary antibody (PA1-28664), donkey anti-mouse IgG Alexa Fluoro-488 (A21202), donkey anti-rabbit IgG Alexa Fluoro-488 (A21206), donkey anti-mouse IgG Alexa Fluoro-568 (A10037), donkey anti-rabbit IgG Alexa Fluoro-568 (A10042), donkey anti-mouse IgG Alexa Fluoro-647 (A32787) and donkey anti-rabbit IgG Alexa Fluoro-647 (A31573) were purchased from Thermo Fisher Scientific. Rabbit anti-SARS-CoV-2 spike (NR-53788) monoclonal antibody, mouse anti-SARS-CoV-2 nucleocapsid monoclonal antibody (NR-53792) and rabbit anti-SARS-CoV-2 nucleocapsid monoclonal antibody (NR-53791) were obtained through BEI Resources, NIAID, NIH. Goat anti-mouse Fab AF647 (115-607-003) was obtained from Jackson ImmunoResearch. Goat anti-rabbit IgG CF660C (20813) was purchased from Biotium.

Dulbecco’s phosphate-buffered saline (14190-144) and LB Miller Broth (BP1426-2) were purchased from Fisher Scientific. FBS (10438-026), DMEM (11965-092), Trypsin-EDTA (0.25%) with phenol red (25200072), pen-strep (15140122), Hanks’ balanced salt solution (HBSS) without phenol red (14025092), DMEM powder with high glucose (12100046), LysoSensor Green DND-189 (L7535), PicoPure DNA Extraction Kit (KIT01013), Lipofectamine 2000 Transfection Reagent (11668019), puromycin dihydrochloride (A1113803), blasticidin S HCl (A1113903), hygromycin B (10687010), Hoechst 33342 (H3570), avidin beads (53150), carbenicillin disodium salt (10177012), Annexin V–AF647 (A23204), dithiothreitol (DTT, R0861) and DiIC18 (D7757) were obtained from Thermo Fisher Scientific. EMEM (ATCC 30-2003) was obtained from ATCC. Paraformaldehyde (sc-253236) was obtained from Santa Cruz Biotechnology. Recombinant human IFNα2a (Cyt-204), IFNβ1a (Cyt-236) and IFNλ1 (Cyt-117) proteins were obtained from Prospec Bio. Recombinant human IFNγ (285-IF-100/CF), TNF (210-TA-005/CF) and IL1β (201-LB-010) were purchased from R&D Systems. Luciferase assay reagent (E4550) and Nano-Glo assay reagent (N2011) were purchased from Promega. Camostat mesylate (SML0057), E-64d (E8640), brefeldin A (B652), hydroxychloroquine sulfate (H0915), cellulose (435244), poly-l-lysine hydrobromide (P9155) and polybrene (TR-1003) were obtained from Sigma. In-Fusion snap assembly master mix (638948) and Stellar Competent Cells (636766) for cloning were purchased from Takara Bio. Four-well chambered cover glass (C4-1.5H-N) was obtained from Cellvis. Tween-20 (AB02038-00500) and dimethyl sulfoxide (AB03091-00100) were obtained from American Bio. Sulfo-NHS-SS-Biotin (A8005) and ionomycin calcium salt (B5165) were purchased from APExBio. Paraformaldehyde (PFA; 4%, SC-281692) was obtained from ChemCruz. Glutaraldehyde (50%, 16320) was obtained from Electron Microscopy Science. Alexa Fluor 488 ChromPure Human Transferrin (009-540-050) was obtained from Jackson ImmunoResearch. cOmplete Protease Inhibitor Cocktail (11697498001) was obtained from Roche. High-Fidelity 2X PCR Master Mix (M0541L) was obtained from NEB.

Plasmid constructs

The following constructs were obtained from Addgene: pLenti-hACE2-hygro (161758), HIV-1 Gag-mCherry (85390), pLV-EF1α-IRES-Hygro (85134), pLV-EF1a-IRES-Blast (85133), pMSCV-Blasticidin (75085), Lact-C2-GFP (22852), the spike protein expression plasmid for bat CoV-WIV1 (pTwist-WIV1-CoV Δ18) (164439) and the HCV glycoprotein expression plasmid (pD603 H77 E1E2) (86983). The following reagents were obtained through BEI Resources, NIAID, NIH: SARS-Related Coronavirus 2, Wuhan-Hu-1 spike D614G-Pseudotyped Lentiviral Kit (NR-53817) including pLenti-Luc2/ZsGreen, pHDM-gag/pol, pRC-rev1b, pHDM-tat1b and pHDM-spike-D614G. Plasmids encoding the spike proteins for SARS-CoV-2 (USA-WA1/2020), SARS-CoV, MERS-CoV and HcoV-NL63 were provided by C.B.W. Spike protein expression plasmids for SARS-CoV-2 Delta (B.1.617.2) and Omicron (B.1.1.529) variants were provided by S. Chen. pMX-PH-Halo-LgBiT was provided by M. Yamamoto and Z. Matsuda. CypA-HiBiT was provided by W.M. Expression plasmids of the glycoproteins for HcoV-229E (VG40605-UT), HcoV-OC43 (VG40607-UT), HcoV-HKU1 (VG40021-UT) and EboV (VG40304-CF) were purchased from Sino Biological. pLV-EF1α-Flag-IRES-Hygro was modified from pLV-EF1α-IRES-Hygro by inserting a Flag tag between the promoter region and the multiple cloning site.

Human ACE2 was subcloned into a pLV-EF1α-IRES-Hygro vector by using pLenti-hACE2-hygro (Addgene 161758) as a template. Full-length cDNAs encoding human PLSCR1, TMEM41B, LY6E and IFITM3 were obtained by PCR using cDNA from Huh7.5 or A549 cells. Full-length cDNA encoding mouse PLSCR1 was obtained by PCR from the cDNA of mouse liver. Full-length cDNA encoding R. sinicus PLSCR1 (NCBI reference sequence: XM_019748913.1) was directly synthesized from Azenta–GENEWIZ. Full-length cDNA encoding human CD74 p41 (NCBI reference sequence: NM_001025159) and NCOA7 isoform 4 (NCBI reference sequence: NM_001199622.1) were obtained from Origin.

pLV-Hg-PLSCR1, pLV-Hg-LY6E and pLV-Hg-TMEM41B were generated by cloning PCR fragments encoding the corresponding gene into a pLV-EF1α-IRES-Hygro vector by infusion cloning. pMSCV-PLSCR1 was generated by cloning the PCR fragment encoding PLSCR1 into pMSCV-Blasticidin. pLV-Flag-PLSCR1, pLV-Flag-mPlscr1, pLV-Flag-batPlscr1, pLV-Flag-TMEM41B, pLV-Flag-IFITM3, pLV-CD74-p41 and pLV-Flag-NCOA7 isoform 4 were generated by inserting PCR fragments encoding the corresponding gene into pLV-EF1α-Flag-IRES-Hygro by infusion cloning. pLV-GFP-PLSCR1 was generated by cloning the PCR fragments encoding PLSCR1 and EGFP into a pLV-EF1α-IRES-Hygro vector by infusion cloning. pLV-Lact-C2-GFP was generated by amplifying a Lact-C2-GFP fragment from the template plasmid purchased from Addgene and cloning it into pLV-EF1α-IRES-Hygro. pLV-PH-Halo-LgBiT was generated by amplifying a PH-Halo-LgBiT fragment from pMX-PH-Halo-LgBiT and then cloning it into pLV-EF1α-IRES-Hygro. pLV-Hg-PLSCR1-KKHA (K²⁵⁸K²⁶¹H²⁶²A); pLV-Hg-PLSCR1-H262Y, -H262Q, -H262A, -H262F, -H262W, -H262D, -H262E, -H262K, -H262R, -H262L and -H262V; pLV-Hg-PLSCR1-F281A; pLV-Hg-PLSCR1(5CA) (C¹⁸⁴C¹⁸⁵C¹⁸⁶PC¹⁸⁸C¹⁸⁹ to AAAPAA); PLV-PLSCR1-I105A, -I108A, -I279A, -L283A, -L285A, -M293A, -D242A, -D244A, -S260A, -W263A, -T264A and -N276A; pLV-Hg-mPlscr1-Q271Y; pLV-Hg-batPlscr1-Q286Y; MSCV-PLSCR1(5CA); MSCV-PLSCR1-F281A; and MSCV-PLSCR1-H262Y were generated by PCR-based site-directed mutagenesis. PLV-Hg-Flag-PLSCR1-86-CT, pLV-Hg-Flag-PLSCR1-Δ86-118 and pLV-Hg-Flag-PLSCR1-1-290 were generated by subcloning PCR fragments encoding the corresponding PLSCR1 truncations into pLV-EF1α-Flag-IRES-Hygro.

Virus strains

The following viruses were used in our study: SARS-CoV-2 USA-WA1/2020 (BEI Resources NR-52281), SARS-CoV-2-mNG (a gift from C.B.W.), SARS-CoV-2 Delta variant (B.1.617.2, a gift from C.B.W.), SARS-CoV-2 Omicron variant (B.1.1.529, a gift from C.B.W.), MHV-A59-GFP (BEI Resources NR-53716), Dengue (DENV-I, BEI Resources NR-82) and HSV-1 VP26-GFP (a gift from A. Iwasaki).

Genome-wide CRISPR–Cas9 knockout screen

The genome-wide CRISPR–Cas9 knockout screen was modified from a previous report¹¹. The LentiCRISPR-V2 pooled library (GeCKO v2) was amplified as described previously⁵². A total of 5 × 10⁷ Huh7.5 or A549-ACE2 cells were transduced with lentiviruses carrying the GeCKO v2 library (MOI = 0.3) followed by puromycin selection (2 μg ml⁻¹) for 5 days. Surviving cells were split into two groups (+ or −IFNγ) and seeded into 20 T-175 flasks at a density of 5 × 10⁶ per flask. After 24 h, IFNγ (R&D Systems) was added for an additional 20 h (10 U ml⁻¹ for Huh7.5 and 70 U ml⁻¹ for A549-ACE2). Cells were subsequently infected with icSARS-CoV-2-mNeonGreen (mNG) at MOI = 1 (for Huh7.5) or MOI = 0.3 (for A549-ACE2). At 24 hpi (A549-ACE2) or 48 hpi (Huh7.5), cells were trypsinized and fixed in 4% PFA for 30 min and analysed on a FACSAria (BD). Cells were sorted into two groups: mNG^high or mNG^low on the basis of the intensity of mNG. Cellular DNA was extracted using the PicoPure DNA Extraction Kit according to the manufacturer’s instruction. sgRNA sequences were amplified using High-Fidelity PCR master mix (NEB) and amplicons were purified from 2.5% agarose gel. Amplicons were sequenced using an Illumina HiSeq2500 (40 million reads per sample). The enrichment of genes in mNG^high versus mNG^low was ranked by the MAGeCK algorithm and the MAGeCK P value of each gene in both the resting and the IFNγ-activated condition was calculated for comparison.

FACS

Cells were fixed and collected in FACS buffer (1× PBS, 1% FBS, 5 mM EDTA) and filtered through a 40-µm cell strainer before FACS sorting using a BD FACSAria. For sorting of Huh7.5 and A549-ACE2 cells, gates were drawn to separate the cells into mNG^low and mNG^high populations in both IFNγ-untreated and IFNγ-treated conditions. GFP was excited by a 488-nm laser and detected with a 550-nm filter. Cells were sorted on the basis of their mNG intensity and collected into separate tubes for later processing. Data were analysed with FlowJo (BD Biosciences).

RNA-seq

Huh7.5 cells were treated with 100 U ml⁻¹ IFNγ for 24 h. Total RNA was isolated using the Rneasy Plus Mini Kit (Qiagen). mRNA libraries for sequencing were prepared according to the standard Illumina protocol. Sequencing (100 bp, paired-end) was performed using the Illumina NovaSeq sequencing system at the Genomics Core of Yale Stem Cell Center. The RNA-seq reads were mapped to the human genome (hg38) with STAR in local mode using default settings. The uniquely mapped reads (cut-off: mapping quality score (MAPQ) > 10) were counted to ENCODE gene annotation (v.24) using FeatureCounts. Differential gene expression was analysed with the R package DESeq2. Transcripts with a log₂-transformed fold change > 1 and adjusted P < 0.05 were considered as differentially expressed.

Analysis of transcription-factor binding profiles

The promoter region of the human PLSCR1 gene was analysed using the JASPAR website⁵³ (https://jaspar.genereg.net/). The sequence of the 2-kb region upstream from the transcription initiation site of PLSCR1 was downloaded from NCBI and scanned by JASPAR using the ISRE and GAS profile. A relative score 0.85 was set as the threshold.

Virus infection

For fluorescent reporter assays using SARS-CoV-2-mNG, P3 stocks were used for infection. Cells were seeded at 40% confluency in 96-well plates 2 days before infection. The following day, cells were either left untreated or primed with IFNγ (Huh7.5: 8 U ml⁻¹; A549-ACE2: 70 U ml⁻¹) 18 h before infection. IFNγ was kept in the medium during infection. Huh7.5 cells and A549-ACE2 cells were infected at MOI = 1 and MOI = 0.2, respectively. Subsequently, at 1 dpi (A549-ACE2) and 2 dpi (Huh7.5), the medium was removed from the wells and cells were washed once with PBS before being fixed with 4% PFA for 30 min and stained with Hoechst 33342 for an additional 20 min. Then, high-content imaging (Cytation 5, BioTek) of the cells was performed to measure mNG expression. The average intensity of mNG per cell as well as the percentage of infected cells were quantified and analysed by Gen5 software. Infection of additional cell lines was performed under the following conditions: Calu-3 (MOI = 1, 24 hpi), HeLa-ACE2 (MOI = 0.2, 24 hpi), Tonsil-ACE2 (MOI = 1, 24 hpi), HaCaT-ACE2 (MOI = 1, 24 hpi) and LET1-ACE2 (MOI = 0.1, 24 hpi).

For viral RNA experiments with SARS-CoV-2 USA-WA1/2020, the Delta variant (B.1.617.2) and the Omicron variant (BA.1), Huh7.5 cells were seeded at 40% confluency in 12-well plates 2 days before infection. On the day of infection, Huh7.5 cells were infected at MOI = 1 (2 dpi), MOI = 0.5 (1 dpi) and MOI = 0.5 (1 dpi), respectively. All infection assays above were performed in a Biosafety Level 3 (BSL-3) facility.

For infection with additional viruses, the experimental conditions are as follows: HeLa cells were infected with HSV-1 VP26-GFP at MOI = 0.2 for 48 h. LET1 cells were infected with MHV-A59-GFP at MOI = 0.1 for 48 h. Huh7.5 cells were infected with DENV-I at MOI = 0.1 for 24 h. Subsequently, at 1 dpi (A549-ACE2) and 2 dpi (Huh7.5), the medium was removed from the wells and cells were washed once with PBS before being fixed with 4% PFA for 30 min and stained with Hoechst 33342 for an additional 20 min. Then, high-content imaging (ImageXpress Pico, Molecular Devices) of the cells was performed to measure GFP expression. The average intensity of GFP per cell as well as the percentage of infected cells were quantified and analysed by the CellReporterXpress software.

For confocal imaging, A549-ACE2 cells were spinfected with SARS-CoV-2 USA-WA1/2020 at 1,000g, 37 °C for 30 min to synchronize the infection, followed by washing twice with pre-chilled PBS. Pre-warmed DMEM was added to cells to initiate the virus entry. The cells were incubated at 37 °C for various time periods. The MOIs used for experiments are indicated in the figure legends.

SARS-CoV-2 plaque assay

Vero E6 cells were seeded at 9 × 10⁵ cells per well in 6-well plates for infection with SARS-CoV-2 USA-WA1/2020 the following day. First, the medium was removed, and each well was washed once with PBS. Then, 200 µl of 10-fold serial dilutions of virus was added to the corresponding wells, and cells were incubated at 37 °C for 1 h with gentle rocking every 10 min. Afterwards, 2 ml of overlay medium (DMEM, 2% FBS, 0.6% methylcellulose) was added to each well. At 2 dpi, the medium was removed, and cells were washed once with PBS. Then, cells were fixed with 4% PFA for 30 min before staining with 0.5% crystal violet solution for 15 min. Finally, cells were washed three times with PBS and then dried before counting the number of plaque-forming units (PFU).

Measurement of viral RNA by qPCR with reverse transcription

Cells grown in 12-well plates were washed twice before total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific; 13778030) and subsequently purified with the Direct-zol RNA Miniprep kit (Zymo Research; R2050). Then, the RNA was reverse-transcribed using PrimeScript RT Master Mix (Takara Bio; RR036B). The cDNA was diluted 1:5 before qPCR with reverse transcription (qRT–PCR) was performed using PowerUp SYBR Green Master Mix (Thermo Fisher Scientific; A25776). SARS-CoV-2 (US-WA1/2020, Delta variant and Omicron variant) replication was quantified by using primers specific to nucleocapsid (N) mRNA (forward 5′-GGGGAACTTCTCCTGCTAGAAT-3′; reverse 5′-CAGACATTTTGCTCTCAAGCTG-3′). DENV-I replication was quantified by using primers specific to non-structural protein 1 (NS1) mRNA (forward 5′-GCATATTGACGCTGGGAGAGAC-3′; reverse 5′-TTCTGTGCCTGGAATGATGCTG-3′). All viral mRNA levels were normalized to β-actin (forward 5′-CACCATTGGCAATGAGCGGTTC-3′; reverse 5′-AGGTCTTTGCGGATGTCCACGT-3′). Reactions were performed on the QuantStudio Real-Time PCR system (Thermo Fisher Scientific, Applied Biosystems). For relative quantification of mRNA levels, the cycle threshold (Ct) values were compared using the ΔΔCt method.

Western blot

Cell lysates were prepared in 1.2× SDS–PAGE sample loading buffer. The cell lysates were fractionated on SDS–PAGE (12% gel) and transferred onto polyvinylidene fluoride (PVDF) membrane (Millipore; IPVH00010). Membranes were blocked with 5% milk in 1× TBST (1× Tris-buffered saline, 0.1% Tween-20) and then incubated with primary antibody at 4 °C overnight in 5% BSA. Subsequently, membranes were washed three times with 1× TBST and then incubated with horseradish-peroxidase-conjugated secondary antibodies. The membranes were exposed using Clarity Normal/Max Western ECL substrate (BioRad; 1705062), and the readout was detected using the BioRad ChemiDoc MP system.

Production of pseudovirus particles

HIV-1-based PsV was produced in HEK293T cells plated on a 10-cm plate. Cells were transfected with lentiviral backbone (9 μg pLenti-Luc2/ZsGreen) and helper plasmids (2 μg pHDM-gag/pol, 2 μg pRC-rev1b and 2 μg pHDM-tat1b), along with an expression plasmid encoding the glycoprotein gene of the virus of interest (3 μg). After incubation for 4 h at 37 °C, the medium was replaced with fresh medium (DMEM, 10% FBS, supplemented with pen-strep). PsV particles were collected 48 h after transfection, clarified by centrifugation (1,000g × 5 min), filtered through a 0.45-µm filter and then aliquoted for storage at −80 °C.

mCherry-labelled HIV-based pseudoviral particles were prepared by transfecting 293T cells plated on a 10-cm dish with 9 μg pLV-EF1a-IRES-Blast, 4.5 μg psPAX2, 1.5 μg HIV-gag-mCherry and an expression plasmid encoding spike protein from SARS-CoV-2 Omicron variant (3 μg) by Lipofectamine 2000. At 48 h after transfection, the supernatant was filtered through a 0.45-μm filter, laid onto a 20% sucrose (w/v in 1× HBSS) cushion and centrifuged using a Sorvall TH-641 rotor at 100,000g for 2 h. The supernatant was discarded and the pseudoviral particles concentrated in the pellet were resuspended with 500 μl of DMEM cell culture medium.

Pseudovirus entry infection assay

A total of 1.2 × 10⁴ Huh7.5 cells were seeded in each well of a clear bottom 96-well plate. Two days later, 100 µl of PsV was added to each well, and the plate was centrifuged at 1,000g for 30 min at room temperature. Infected cells were incubated for two days before being lysed with 1× passive lysis buffer (Promega; E1941) for 20 min at 4 °C. Then, 80 µl luciferase assay reagent (Promega; E4550) was added to the lysate and relative luminescence was measured using a microplate reader (BioTek).

Protease inhibitor assay

Huh7.5 and A549-ACE2 cells were seeded at 2.5 × 10⁴ cells per well in 96-well plates. The next day, cells were primed with protease inhibitors at the following concentrations or dilutions: mock DMSO (1:200), E-64d (25 μM), camostat (20 μM), brefeldin A (5 μg ml⁻¹) and HCQ (10 μM). Three hours later, Huh7.5 and A549-ACE2 cells were infected with SARS-CoV-2-mNG at MOI = 1 and MOI = 0.2, respectively. Inhibitors were kept in the medium during infection. At 1 dpi (A549-ACE2) and 2 dpi (Huh7.5), cells were washed with PBS and then fixed with 4% PFA. Nuclei were stained with Hoechst and mNG expression was measured by high-content imaging.

For inhibitor assays in Calu-3, cells were seeded at 2.75 × 10⁵ cells per well in 12-well plates. When they reached 90–95% confluency, cells were primed with protease inhibitors at the following concentrations or dilutions: mock DMSO (1:800), E-64d (25 μM) and camostat (25 μM). Three hours later, cells were infected with SARS-CoV-2 USA-WA1/2020 at MOI = 1. Inhibitors were kept in the medium during infection, and cells were incubated at 37 °C for 1 h with gentle rocking every 10 min. Afterwards, the infection medium was removed, and cells were washed twice with PBS. Then, fresh medium with inhibitors was added to each well. At 1 dpi., cells were washed twice before total RNA was extracted using TRIzol reagent and then purified using Direct-zol RNA Miniprep kit. Subsequently, viral RNA was measured by qRT–PCR using primers specific to SARS-CoV-2 nucleocapsid mRNA.

Biotinylated ACE2 pulldown assay

A549-ACE2 cells were seeded at 2 × 10⁶ cells per well in 6-well plates. Two days later, cell-surface proteins were labelled with biotin. First, cells were washed twice with DBPS⁺ solution (DPBS supplemented with 0.9 mM CaCl₂ and 0.49 mM MgCl₂, pH 7.4). Next, cells were incubated with 2.5 mg ml⁻¹ biotin (EZ-LinkTM Sulfo-NHS-LC-LC-Biotin, Thermo Fisher Scientific) in DBPS⁺ solution at 4 °C for 30 min. Afterwards, cells were washed three times in 100 mM glycine for 5 min, followed by two additional wash cycles in 20 mM glycine for 5 min. Subsequently, cells were lysed with lysis buffer (1% Triton X-100, 50 mM Tris/HCl pH 7.4, 150 mM NaCl, 1 mM EDTA and Complete Protease Inhibitor Cocktail), and a portion of the whole-cell extract was aliquoted for input. The remaining extract was incubated with avidin beads on a rocker at 4 °C overnight. Finally, mixtures of the whole-cell extract and beads (biotin pulldown) were washed six times with lysis buffer and then boiled for immunoblotting.

Immunostaining

For immunostaining, cells cultured on coverslips were washed twice with PBS, fixed with 4% PFA for 30 min and permeabilized with 0.2% Triton X-100 for 3 min. Cells were blocked with 1% BSA in PBS for 1 h and incubated with primary antibodies (1:50–1:200 diluted) in PBS supplemented with 1% BSA overnight at 4 °C, followed by incubation with secondary antibodies (1:500 diluted) for 1 h at room temperature. Nuclei were stained with Hoechst 33342 (1:4,000 diluted) for 5 min.

Fluorescent imaging

All confocal images were acquired using a Leica SP8 laser scanning confocal microscope with 405-nm and 488-nm lasers and a pulsed supercontinuum white light source (470 nm–670 nm). For analysing the subcellular localization of proteins, images (1,024 × 1,024) were taken under a HC PL APO 100× oil immersion objective (N.A. 1.44) with 4 frames average. For determining the number of dsRNA foci or the distribution of SARS-CoV-2 nucleocapsid, images (1,024 × 1,024) were taken under a HC PL APO CS2 63× oil immersion objective (N.A. 1.40) with 4 frames average. For determining the number of spike and nucleocapsid double-positive particles, Z-stack images (512 × 512) were acquired under a HC PL APO CS2 63× oil immersion objective (N.A. 1.40) with 0.5 µm per optical section.

For time-lapse microscopy of the formation of PLSCR1-wrapped vesicles, A549-ACE2 cells were seeded on 4-well chambered cover glass (1.5) a day before infection. Cells were spinfected with mCherry-labelled pseudoviruses containing SARS-CoV-2 spike at 1,000g for 30 min at 4 °C. Images (512 × 512) were captured on a DeltaVision OMX SR microscopy system under a 63× oil immersion objective (N.A. 1.40) with 1-min intervals for around 2 h at 37 °C, 5% CO₂ and 80% humidity.

4Pi-SMS nanoscopy

Two-colour 4Pi-SMS⁴¹ was performed on a custom-built microscope with two opposing objectives in 4Pi configuration⁵⁴. A549 cells were seeded on 30-mm-diameter no. 1.5H round coverslips (Thorlabs) and grown for 1 day before infection. Infected cells were fixed with 4% PFA for 30 min and permeabilized with 0.2% Triton X-100 for 3 min. Cells were blocked with 10% goat serum in PBS for 1 h and subsequently incubated with mouse anti-PLSCR1 and rabbit anti-spike antibodies overnight at 4 °C. Primary antibodies were detected by goat anti-mouse Fab AF647 (Jackson ImmunoResearch) and goat anti-rabbit IgG CF660C (Biotium) at a 1:200 dilution (2 h at room temperature). Samples were post-fixed in 3% PFA + 0.1% glutaraldehyde for 10 min and stored in PBS at 4 °C. Sample mounting, image acquisition and data processing were mostly performed as previously described⁵⁴ except that imaging speed was 200 Hz with a 642-nm laser intensity of around 12.5 kW cm⁻². Typically, 3,000 × 100~200 frames were recorded. DME were used for drift correction. All 4Pi-SMS images and videos were rendered using Point Splatting mode (20-nm particle size) with Vutara SRX 7.0.06 software (Bruker).

Image processing and analysis

Images were processed in LAS X (Leica), SoftWoRx (v.7.0), Fiji or Imaris 9.8 (Oxford Instruments) software. Deconvolution of fluorescent images was performed in LAS X using the default settings. Fluorescence colocalization analysis was performed using Imaris software. The percentage of PLSCR1 fluorescent signal colocalized with LAMP1 or CD63 as well as the Mander’s overlap coefficient were calculated by setting proper thresholds for both channels to avoid background signal. The PLSCR1-positive foci were automatically detected by surface reconstitution using Imaris. The average number or fluorescent intensity of PLSCR1-positive foci per cell was calculated by dividing the total number or fluorescent intensity of PLSCR1-positive foci by the total number of cells in each randomly selected image, respectively. Cells with large dsRNA foci were identified manually using Fiji and the percentage of cells with large dsRNA foci was calculated by dividing the number of cells with dsRNA foci by the total number of cells in each image. Cells with dispersed or endosomal nucleocapsid signal were manually identified using Fiji. The percentage of cells with the indicated nucleocapsid distribution was normalized by the total cell count in each image. Spike and nucleocapsid double-positive particles were automatically identified using Imaris and the average number of double-positive particles per cell was normalized by the total cell number in each image. For all the image analysis, 10–13 images with 120–250 cells in total (as indicated in the figure legends) were randomly captured. The percentage of SARS-CoV-2-containing vesicles coated with PLSCR1 in each cell was quantified manually using Fiji; around 25 cells were counted in each condition.

Virus binding and internalization assay

For the virus binding assay, A549-ACE2 cells were pre-chilled to 4 °C for 15 min followed by incubation with SARS-CoV-2 (MOI = 20) at 4 °C for 1 h. Unbound viral particles were removed by washing with pre-chilled PBS three times. The relative amount of bound virus normalized to β-actin was quantified by qPCR.

For the virus internalization assay, cells were incubated with SARS-CoV-2 using the same condition described above. Cells were then transferred to 37 °C for 30 min to allow internalization of bound virus. Uninternalized viral particles were removed by treating cells with 0.25% trypsin for 15 min at 4 °C. As a negative control, another set of cells was directly treated with trypsin after virus binding to digest all the bound but uninternalized viruses. The relative amount of internalized virus normalized to β-actin was quantified by qPCR.

Virus–cell membrane fusion assay

The virus–cell fusion assay was performed according to the methods described previously^40,55 with modification. HIV-based pseudoviral particles containing CypA-HiBiT were prepared by transfecting 293T cells plated on a 10-cm dish with 9 μg pLenti-Luc2/ZsGreen, 4.5 μg psPAX2, 2 μg CypA-HiBiT and 3 μg pVP40-spike (encoding the spike protein from SARS-CoV-2 Delta) by Lipofectamine 2000. At 48 h after transfection, the supernatant was filtered through a 0.45-μm filter, laid onto a 20% sucrose (w/v in 1× HBSS) cushion and centrifuged using a Sorvall TH-641 rotor at 100,000g for 2 h. The supernatant was discarded and the pseudoviral particles concentrated in the pellet were resuspended with 500 μl of DMEM + 10% FBS medium.

293T-ACE2 cells were transfected with pMX-PH-Halo-LgBiT a day before infection. Huh7.5 cells were stably transduced with pLV-PH-Halo-LgBiT, followed by hygromycin (350 μg ml⁻¹) selection for 7 days.

Target cells expressing the LgBiT fragment were plated in a white opaque 96-well dish. One day after plating, each well of cells was spinfected with 100 μl of pseudoviruses containing CypA-HiBiT at 1,000g and 4 °C for 30 min, followed by incubation at 37 °C for 1 h (293T-ACE2) or 2 h (Huh7.5). The medium was removed and Nano-Glo assay reagent was added to the target cells. The activity of complemented NanoLuc was measured by Spectramax i3x microplate reader (Molecular Devices).

Cell–cell membrane fusion assay

The virus–cell fusion assay was performed according to the methods described previously²⁵ with modification. Huh7.5 cells (acceptor cells) of indicated genotypes were stably transduced with the construct encoding the complementary fragment of the split-NanoLuc (PH-Halo-LgBiT) as described above. 293T cells (donor cells) were transfected with plasmids encoding the spike protein from SARS-CoV-2 Omicron together with a split-NanoLuc construct encoding the HiBiT fragment. Twenty-four hours after transfection, the donor and acceptor cells (at ratio 1:1) were mixed and co-cultured in a 96-well plate at 37 °C for 18 h before assay. The medium was removed and Nano-Glo assay reagent was added to the target cells. The medium was removed and Nano-Glo assay reagent was added to the cells. The activity of complemented NanoLuc was measured by Spectramax i3x microplate reader (Molecular Devices).

For visualization of the syncytia formed after cell–cell fusion, 293T cells were transduced with plasmids encoding SARS-CoV-2 spike protein and EGFP. Twenty-four hours after transfection, Huh7.5 cells of the indicated genotypes were co-cultured with the 293T cells at 37 °C for 18 h. The cells were washed with PBS, fixed by 4% PFA and stained with Hoechst. Images were obtained with high-content imaging (ImageXpress Pico, Molecular Devices) as described above.

Spike protein cleavage assay

The spike cleavage assay was modified according to a previous report⁵⁶. A549-ACE2 cells seeded on a 24-well plate were spinfected with pseudovirus particles containing SARS-CoV-2 spike at 1,000g for 30 min at 4 °C. Cells were then washed twice with DPBS, incubated with pre-warmed culture medium and shifted to 37 °C. Cells were collected at different time points followed by western blot analysis. Both the intact S2 domain and the processed S2’ fragment was detected by an antibody specifically recognizing the S2 domain. Cells treated with Cathepsin inhibitor E-64d were used as a negative control.

PS externalization assay

PS externalization was detected using Annexin V conjugated with Alexa647 (A23204, Thermo Fisher Scientific) according to the manufacturer’s instruction. In brief, A549-ACE2 cells were digested by trypsin, spun down at 200g and washed twice with PBS. Cell pellets were resuspended in 100 μl of 1× binding buffer (prepared from 5× stock solution provided by the manufacturer) at a density of 5 × 10⁶ cells per ml, and treated with DMSO or 10 μM ionomycin for 10 min. Cells were subsequently incubated with 5 µl Annexin V–AF647 for 20 min at room temperature followed by the addition of 400 µl 1× binding buffer. Cells were then analysed using a Beckman CytoFLEX S flow cytometer (APC filter).

MDS analysis

A coarse-grained simulation using the AlphaFold⁵⁷ structure of PLSCR1 with the N-terminal region truncated was assembled using the insane.py Python script⁵⁸, memembed (ref. ⁵⁹) and martinize2 (ref. ⁶⁰) in an asymmetric membrane comprised of phosphatidylcholine (PC):phosphatidylethanolamine (PE) (8:2) in the lower leaflet and PC:PE:PS:PtdIns(4,5)P₂ (2:5:2:1) in the upper leaflet. A 5-µs simulation was performed to equilibrate the system. This was then back-mapped to atomistic resolution using the cg2at tool⁶¹. The mutation H262Y was made at this point using PyMOL, followed by an energy minimization step using the steepest descents algorithm. Separately, the atomistic system with the palmitoyl tails was assembled using CHARMM-GUI (refs. ^62,63) in the same membrane as previously described. The protein had palmitoyl chains added at residues C184, C185, C186, C188 and C189.

Three repeats were performed for each system (palmitoylated wild type and H262Y mutant). Each simulation was 50 ns using a time step of 2 fs at 310 K, using the charmm36m (ref. ⁶⁴) forcefield and TIP3P water. The water bond angles and distances were constrained by SETTLE (ref. ⁶⁵). Hydrogen covalent bonds were constrained using the LINCS algorithm⁶⁶. The velocity rescale⁶⁷ and Parrinello-Rahman⁶⁸ coupling methods were used with the time constants τ_p = 1.0 ps and τ_T = 0.1 ps for pressure and temperature, respectively. The protein, lipids and water and ions were groups separately for temperature coupling. Simulations were run using GROMACS v.2021.3 (ref. ⁶⁹). A single cut-off of 1.2 nm was used for the van der Waals interaction. The particle mesh Ewald (PME) method was used for electrostatic interactions with a cut-off of 1.2 nm.

The RMSF was calculated using the gmx rmsf tool over all repeats and visualized in PyMOL. The hydrogen bond analysis was performed with VMD.

Generation of GPMVs

The preparation of GPMVs was performed according to the methods described previously^70,71. In brief, A549-ACE2-PLSCR1-KO cells stably overexpressing GFP-PLSCR1-WT or -5CA were plated on T-25 plates coated with poly-l-lysine. After 24 h, the cells were stained with 1 µg ml⁻¹ Dil-C18 for 20 min, washed four times with GPMV buffer (20 mM HEPES pH 7.4, 150 mM NaCl and 2 mM CaCl₂) and incubated with 1 ml GPMV buffer containing 1.9 mM DTT and 27.6 mM formaldehyde. Sixteen hours after induction, GPMVs were collected from the supernatant and used freshly for analysis. The quality of GPMVs was checked using a confocal microscope.

Measurement of membrane bending rigidity

The membrane bending modulus of a GPMVs was estimated by aspirating the GPMV using a micropipette, and pulling a thin membrane tether using a spherical bead trapped by a focused infrared laser beam. The force acting on the tether is (f=2{rm{pi }}sqrt{2{kappa }{sigma }}), where κ is the membrane bending modulus and σ is the membrane tension⁷². The membrane tension is varied by changing the aspiration pressure ΔP, σ = ΔPR_p/[2(1 − R_p/R_v)], where R_p and R_v are the radii of the micropipette and the GPMV, respectively⁷³. For every value of σ, the resulting f is measured using the displacement of the trapped bead from the centre of the optical trap. The bending modulus is estimated from the slope of a line fit to the plot of f² as a function of σ. The procedure is repeated for different GPMVs.

Our home-built optical tweezers setup is combined with a spinning-disc confocal system⁷⁴ which allowed us to verify that wild-type PLSCR1–GFP was localized to the GPMV membranes whereas the 5CA mutant was distributed diffusely inside the GMPVs. The trap stiffness was calibrated using a hydrodynamic flow method⁷⁵ and was 283 pN µm⁻¹ for the polystyrene beads (diameter = 3.15 µm, PP-30-10, Spherotech).

For micropipette aspiration, we used a micropipette holder attached to a programmable three-axis piezo stage (100 µm range, P-611.3 NanoCube, with controller E-727 and Mikromove software, Physik Intrumente), mounted on a manual manipulator for coarse movement (Newport M-462, Newport). Both the micropipette and the glass chamber were coated with 1% BSA to minimize GPMV adhesion and facilitate the free flow of the GPMV membrane inside the micropipette. Aspiration pressure was controlled by vertically moving a water reservoir connected to the micropipette. Before aspirating a GPMV, pressure was zeroed by bringing a bead near the tip of the micropipette (3–6 µm) and adjusting the reservoir height until the bead stopped moving. Subsequently, a bead was trapped, and its zero-force position (x₀, y₀) was recorded. Then, the GPMV was brought in contact with the bead briefly (around 1–3 seconds) before being pulled away to form a membrane tether. Bead positions (x, y) were determined from analyses of image stacks. The presence of a tether was confirmed either visually or by releasing the trap and observing the retraction of the bead toward the GPMV. The tether force was calculated from the deviation of the bead’s position from its load-free value and the trap stiffness using a custom-written MATLAB program⁷⁴. All the plots and fitting were done using Origin 2023 (OriginLab).

Statistics and reproducibility

Data were subjected to statistical analysis and plotted using Microsoft Excel 2010 or GraphPad Prism v.9.0. D’Agostino and Pearson omnibus normality test or Kolmogorov–Smirnov test were used to determine the normal distribution of data. For data with a normal distribution, single comparisons were performed using the two-sided Student’s t-test for groups with equal variances, and Welch’s correction was used for groups with unequal variances. Datasets that did not follow a normal distribution were analysed using a nonparametric test (Mann–Whitney test). Multiple comparisons were assessed by one-way ANOVA with Tukey’s post-hoc test or two-way ANOVA with Tukey’s post-hoc test, whereas Brown–Forsythe and Welch ANOVA with Dunnett’s post-hoc test was used for groups with unequal variances. For all analyses, P < 0.05 was considered statistically significant. Results were reported as either mean ± s.e.m. or mean ± s.d. as indicated. The methods for calculating P values are indicated in the figure legends. All of the P values obtained from statistical analysis are listed in the graphs or in the source data files. All experiments were repeated at least three times with similar results unless otherwise mentioned in the figure legends.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.