ReagentsCells

Madin-Darby canine kidney (MDCK)-2,6-sialtransferase (SIAT1) cells (Sigma-Aldrich, 05071502), MDCK.2 cells (ATCC strain CCL-34) and human embryonic kidney (HEK)293T cells (ATCC) were propagated in DMEM medium (Cytiva) supplemented with 10% FBS (Atlas Biologicals). Calu3 cells (ATCC), A549 cells (ATCC) and Caco2 (ATCC) cells were propagated in EMEM medium (Wisent) supplemented with 20% FBS. HEK293F cells, a gift from S. C. Harrison (Harvard Medical School), were propagated in FreeStyle 293 Expression Medium (Thermo Fisher) and were used for the expression of HC19, MEDI8852 and 14C2 antibodies.

Antibodies

Purified Fi6, CR9144, O19, W6/32, T2-5D and T2-7D antibodies, and hybridomas producing anti-influenza virus HA antibodies Sb H36-26, Sa Y8-1A6-6, Ca H2-4B1-14 and Cb H9-D3-4R2 and NA antibodies NA2-1c1 and NA2-10e10 were gifts from Jonathan Yewdell (NIH/NIAID/LVD/CBS). Hybridoma producing anti-influenza virus M2 monoclonal antibody 65 was a gift from Xavier Saelens (VIB-UGent Center for Medical Biotechnology). Hybridoma producing anti-influenza virus NP monoclonal antibody HB65 was obtained from ATCC (H16-L10-4R5, 58696953). The expression vectors (modified pVRC8400) for HC19 and MEDI8852 IgG heavy and light chains were a gift from S. C. Harrison (Harvard Medical School). The sequence for F045 was previously described24. For 14C2 IgG, the variable heavy and variable light chain sequences from a single-chain variable fragment25 were incorporated into a mouse IgG expression plasmid. Expressed antibodies were produced by transient transfection of HEK293F cells with polyethylenimine (0.4 μg heavy chain DNA, 0.6 μg light chain DNA, 1.5 μg polyethylenimine, per 1 × 106 cells) and collected from the cell culture supernatant at 4–7 days post transfection. Antibodies from hybridomas, HC19, MEDI8852 and 14C2 IgG were purified using protein G resin (Cytiva). TriHSB.2 plasmid was a gift from David Baker (University of Washington). TriHSB.2 was expressed by autoinduction in bacteria26. F045 and TriHSB.2 constructs contain C-terminal 6×His tags and were purified by passage over a HisTrapFF column (GE Healthcare), followed by size exclusion chromatography using a Superdex 200 increase 10/300 column (Cytiva). Antibodies were stored in PBS or PBS containing 10% glycerol.

Antibody labelling

Lyophilised DyLight550 (DL550, Thermo Fisher), JaneliaFluor646 (JF646, Janelia) or AlexaFluor488 (AF488, Thermo Fisher) NHS ester dyes were resuspended in anhydrous dimethylsulfoxide (Sigma-Aldrich). Labelling was performed with a 100:1 (HC19, Sb H36-26, T2-5D, or T2-7D:DL550 or Fi6:AF488) or 20:1 (HB65:AF488, DL550 or JF646) molar ratio of dye to antibody in 100 mM sodium bicarbonate buffer for 1 h at room temperature, followed by passage over a Macro SpinColumn packed with G25 packing material (Harvard Apparatus) in PBS. Labelled antibody concentrations were calculated using absorbance at 280 nm and 493 or 557 nm.

Primers

NP-gene segment primers used were oNB23: TACTGGGCCATAAGGACCA and oNB24: TCCTCTGCATTGTCTCCGA. 7sk snoRNA primers used were oEP1591: CTGGCTGCGACATCTGTCA and oEP1592: GGAGGTTCTAGCAGGGGA.

Viruses

A/Puerto Rico/8/1934 (PR8) wild-type influenza virus and A/Udorn/307/1972 containing the A/Aichi/1968 (X31) HA segment (XUdorn)20 were passaged in Calu3 cells at MOI 0.002 with 1 µg ml−1 TPCK-trypsin (Sigma-Aldrich) in OptiMEM (Thermo Fisher). The HA- and NA-gene segments were sequenced to ensure correctness in stock viruses. Viruses used here were passaged once from these stocks and behaviour was indistinguishable in our assays. Viruses were titred by haemagglutination (HA), infectivity assay and flow virometry. HA assays were performed with turkey red blood cells. Viruses A/Hong Kong/1/1968 (HK68) and A/California/07/2009 (Cal0709) were provided by Jonathan Yewdell (NIH/NIAID/LVD/CBS). These were derived from chicken eggs and used without propagation. Work as described with IAV strains A/Puerto Rico/8/1934, A/Udorn/307/1972 containing the A/Aichi/1968 (X31) HA-gene segment, A/Hong Kong/1/1968 and A/California/07/2009 was approved by the Institutional Biosafety Committee (IBC) at the Biosafety Level 2 Laboratory (BSL-2; Registration #RD-23-IV-10).

Infectivity assays

For each cell line–virus strain combination, infectivity (virions per IU) was calculated from the average result of infections at two MOIs, each performed in triplicate. Cells were infected at virion inputs ranging from 0.001–2 virions per cell (see ‘Infections without spread’ below). At 24 h.p.i., cells were fixed and permeabilized using 4% paraformaldehyde and 0.1% Triton X-100, and then stained with AF488- or JF646-labelled HB65, which recognizes NP. The number of infected cell singlets was determined by flow cytometry (NP deriving from input virions binding to cells at t0 was not detected under these conditions). Each replicate derived at least 2 inputs with a percent of cells infected of 8 or less, ensuring that infected cells resulted from a single infectious unit.

Virus purification

XUdorn viruses were purified through a 20% sucrose cushion (SC), then either by passage over a 20–60% sucrose gradient20 or by sequential centrifugation: viruses were processed for 8 cycles of centrifugation at 3,250 relative centrifugal force for 1.5 h at 4 °C9. Three fractions were collected from the sucrose gradient. ‘Spherical’ is a distinct band in the sucrose gradient, and ‘Filamentous 1’ and ‘Filamentous 2’ are virion samples deriving from diffuse-virus regions of the gradient of progressively higher sucrose density. From the sequential centrifugation, additional filament- (filamentous 3 and fil. other) and sphere- (sph. other) enriched fractions were collected from the pellet or the supernatant, respectively (Extended Data Fig. 1).

Direct (antibody-independent) virus labelling

Virus was fluorescently labelled by conjugating viral surface proteins to AF488 via an NHS ester reaction. PR8 virions (4 × 109) in 0.1 M NaHCO3 were combined with 5.6 nmol AF488 NHS ester dye in dimethylsulfoxide and incubated at room temperature for 1 h. Labelled viruses were purified using a PD-10 desalting column. Virus was quantified and confirmed to retain binding of Sb H36-26 IgG and MEDI8852 IgG by flow virometry.

Electron microscopy

For electron microscopy, ‘Filamentous 1’, ‘Filamentous 2’ and ‘Filamentous 3’ were adjusted to 2 × 104 HAU per ml and ‘All’ and ‘Spherical’ were adjusted to 1 × 105 HAU per ml. As a precaution, viruses were pretreated with 25 μM rimantadine (M2 inhibitor, Sigma-Aldrich) for 30 min at room temperature to prevent fragmentation of the filamentous virions27. Grids were stained with 2% phosphotungstic acid (pH 7.5–8)20. Images were taken using a Philips Morgagni v.3.0 transmission electron microscope (80 kV) using an AMT NanoSprint5 camera. Virion size measurements were performed using custom MATLAB codes20.

Time-course experiments

MDCK-SIAT1, Calu3, A549 or Caco2 cells were grown in 24-well plates until they formed confluent monolayers. The cells were washed twice with HBSS solution (Thermo Fisher) before attachment of PR8 or XUdorn virus in OptiMEM at MOI 0.006 or 6 for 1 h at room temperature. For Cal0709, titres could not reach MOI 6, so 35 µl undiluted virus was used for attachment. After attachment, unattached virus was removed, and the cells were washed twice with HBSS. Infection media (300 µl) with or without trypsin and with or without 20 mM ammonium chloride, were added. At 4 h.p.i., ammonium chloride was added to some wells. Samples (10 µl) were taken every 4 or 8 h until 48 h.p.i. Infected-cell supernatant aliquots were analysed by flow virometry. Samples were stored in PCR tubes at −80 °C until analysis.

Infections in chicken eggs

The allantoic fluid of day-10 SPF premium eggs (AVSbio) was injected with 106 Cal0709 virions. At 24 h.p.i., eggs were moved to 4 °C for at least 2 h. The allantoic fluid was collected, centrifuged at 4,000 g for 10 min to remove debris, and the supernatant was analysed by flow virometry.

Infections without spreadGeneral

Cells were grown in 24-well plates until they formed confluent monolayers. After washing twice in HBSS, virus was added in 35 µl OptiMEM and incubated at room temperature for 1 h with frequent shaking. After attachment, cells were washed twice with HBSS and 300 µl OptiMEM was added. Trypsin was omitted and 20 mM ammonium chloride was added at 4 h.p.i. to inhibit spread. Infections were incubated at 34 °C with 5% CO2 and 100% humidity. Infected-cell supernatant aliquots were analysed by flow virometry. Samples were analysed freshly or stored in PCR tubes at −80 °C until analysis.

MOI experiments

MDCK-SIAT1, Calu3, A549 or Caco2 cells were infected with PR8, XUdorn, HK68 or Cal0709 virus at MOI 0.0156–2, except for Cal0709, where 0.0078–0.979 was used because of limited titre. Infected-cell supernatants were collected at 24 h.p.i.

Antibody-sweep experiments

MDCK-SIAT1 cells were used for PR8, and Calu3 cells for PR8 and XUdorn. PR8 or XUdorn virus and antibody dilutions were made in OptiMEM. MEDI8852, Fi6, CR9114, HC19, monoclonalAb65, 14c2 and O19 antibodies were used for both viruses. Antibodies used only for PR8 were Sa Y8-1A6-6, Sb H36-26, Ca H2-4B1-15, Cb H9-D3-4R2, NA2-1c1 and NA2-10e10. XUdorn-specific antibodies were F045 and TriHSB.2. Cells were infected at MOI 0.6. For PRE and ALL, a range of the indicated treatment was included during the room-temperature incubation and first 4 h of infection. At 4 h.p.i., the media were removed and the cells washed with HBSS. New infection media containing untreated OptiMEM for PRE or OptiMEM with treatment for POST and ALL were added to the cells. Infected-cell supernatants were collected at 24 h.p.i. Ammonium chloride was not included in these experiments, but effects of spread were expected to be minimal.

Inhibitor experiments

For Fig. 3b–e, MDCK-SIAT1 cells were infected with PR8 virus at MOI 3. When included, Sb H36-26 and MEDI8852 were included during the room-temperature incubation and until 4 h.p.i. At 4 h.p.i., the media were removed and the cells washed with HBSS. When included, baloxavir marboxil (Medchemexpress) was added to the indicated concentration from an 875 µM stock in cell culture dimethylsulfoxide (Sigma-Aldrich) at the end of the room-temperature incubation and after washing with HBSS at 4 h.p.i. Infected-cell supernatants were collected at 24 h.p.i. For Fig. 3f,g and Extended Data Fig. 5, MDCK-SIAT1 cells were infected with PR8 and Calu3 cells were infected with PR8 or XUdorn at MOI 0.3 or 10. Baloxavir was added at the end of the room-temperature incubation. Infected-cell supernatants were collected at 24 h.p.i.

Transient antibody treatment experiment

MDCK-SIAT1 cells were infected with PR8 virus at MOI 0.6. At 20 h.p.i., cells were washed into a range of concentrations of MEDI8852. Samples of the supernatant were collected at 1, 2 and 4 h post treatment with MEDI8852 (Fig. 5a,b) or 2 h post treatment with MEDI8852 (Extended Data Fig. 10a–c). Cal0709 transient antibody experiment (Extended Data Fig. 7b) was performed in eggs (see ‘Infections in chicken eggs’). At 3 h.p.i., PBS or antibody was injected into the allantoic fluid. Concentrations were based on an assumed 50 ml per egg.

Transient antibody genome quantification experiment

MDCK or Calu3 cells were infected with PR8 or XUdorn at MOI 0.6 (Extended Data Fig. 9b). At 20 h.p.i., cells were washed into 250 nM MEDI8852 and incubated for 2 h. Resulting supernatants were analysed by flow virometry and RT–qPCR (see ‘Viral genome quantification’ below).

AF488-virus reuptake control

MDCK-SIAT1 cells were infected with PR8 virus at MOI 3. At 20 h.p.i., cells were washed into a range of concentrations of MEDI8852 along with a fixed concentration (~4,000 per µl) of AF488-labelled PR8 virus. Samples of the supernatant were collected at 2 h post treatment. We confirmed that the produced virus had the expected shape effects.

Antibody internalization experiment

For Fig. 5c, MDCK-SIAT1 cells were infected with PR8 virus at MOI 0.6. At 21 h.p.i., cells were washed into 500 nM MEDI8852 and incubated for 2 h. The cells were then collected, fixed, permeabilized and stained with 300 nM Dapi and 13.3 nM AF647-αMouse IgG secondary antibody. Samples were imaged on a Leica SP8 (690) DMI6000 confocal microscope using an HC PL APO CS2 ×40/1.30 oil objective. Imaging and image processing were performed in LAS X and Fiji, respectively. For Fig. 5d, MDCK-SIAT1 cells were infected with PR8 virus at MOI 0.3. At 22 h.p.i., cells were washed into 300 nM AF488-Fi6 and incubated for 1 h. The cells were then collected, fixed, permeabilized and stained with 300 nM Dapi. Samples were imaged as above.

Osmolarity experiments

MDCK-SIAT1 cells were infected with PR8 virus at MOI 0.6 (Fig. 5e) or Calu3 cells with PR8 or XUdorn virus at MOI 0.3 (Extended Data Fig. 8i,j). At 20 h.p.i., cells were washed into either OptiMEM diluted with ultrapure H2O (0.4×, 0.6×, 0.8×) or supplemented with NaCl (50, 100, 150 mM). Supernatants were collected at 30 min post treatment.

Osmolarity experiments

Calu3 cells were infected with HK68, and A549 and Caco2 cells were infected with PR8, at MOI 1 (Extended Data Fig. 8e,f). At 20 h.p.i., cells were washed into either OptiMEM diluted with ultrapure H2O (0.4×) or supplemented with NaCl (150 mM). Supernatants were collected at 60 min post treatment.

Osmolarity reversibility experiments

MDCK-SIAT1 cells were used for PR8, and Calu3 cells for PR8 and XUdorn. Cells were infected with virus at MOI 0.6. At 20.5, 21.5 and 22.5 h.p.i., cells were washed into either OptiMEM diluted with ultrapure H2O (0.4×, 0.6×, 0.8×) or supplemented with NaCl (50, 100, 150 mM), and the supernatant was collected at 30 min post treatment. At 21 and 22 h.p.i., cells were washed into 1× OptiMEM and the supernatant was collected at 30 min post treatment.

Reinfection experiments

MDCK-SIAT1 cells were infected with PR8 supernatants from the osmolarity or antibody treatment experiments diluted to 3 virions per cell in 35 µl. After 1 h attachment, cells were washed into fresh OptiMEM. Samples of the supernatant were collected at 24 h.p.i.

Clumping experiments

An untreated supernatant from the antibody-sweep experiments (see above) was mixed with antibodies at the highest concentrations used in the antibody experiments. Mixtures were incubated at 34 °C for 1 h and then analysed by flow virometry.

Flow virometry

DyLight550-labelled Sb H36-26 IgG, HC19 IgG, T2-5D IgG and T2-7D IgG stock solutions were diluted to 11.85 nM, 50 nM, 25 nM and 25 nM, respectively, in 0.2% BSA and HNE20 (20 mM HEPES NaOH pH 7.4, 150 mM NaCl and 0.2 mM EDTA). Infected-cell supernatants were undiluted or diluted up to 1:30 in HNE20 and combined 1:1 with antibody dilution in BSA. Sb H36-26 IgG was used to label PR8, HC19 IgG was used to label XUdorn and HK68, and T2-5D IgG or T2-7D IgG was used to label Cal0709. Binding reactions were incubated at room temperature for 30 min to 1 h, then diluted 1:250 in HNE20. Flow virometry was performed using the CytoFLEX S platform (Beckman Coulter). Laser powers were 70 mW for violet and 50 mW for yellow. Gain values were set to 300 for VSSC and 1,000 for RFP. Samples were triggered on violet side-scatter area (1,000–3,500 a.u. threshold) and RFP (300–500 a.u. threshold) and acquired for 600 s or 25,000 particles in the virion gate. Thresholds were optimized for each instrument. Unlabelled concentrated virus preparations were triggered on VSSC and acquired after 2 min until 500,000 particles in the virion gate (Fig. 1b and Supplementary Fig. 1). Virus samples at variable dilutions (1:10 to 1:40,000) were mixed with FluoSphere (Invitrogen) beads (170 nm, 505/515) at 1:300 and acquired by flow cytometry (Supplementary Fig. 1). All flow virometry samples were prepared in HNE20. Analysis was performed in Cytexpert 2.5 or FlowJo 10.9.0.

Flow cytometry

Flow cytometry was performed using the CytoFLEX S platform (Beckman Coulter). Laser powers were 70 mW for violet and 50 mW for yellow. Gain values were set to 85–103 for FSC and 92–333 for SSC. For infectivity assays, gain values were set to 100 for FITC when using AF488-HB65 and 565 for APC when using JF646-HB65. For antibody internalization experiments, gain values were set to 3,000 for APC and 50 for RFP. Samples were triggered on FSC and acquired for 50,000 particles in the singlet-cell gate. All flow cytometry samples were prepared in PBS. Analysis was performed in Cytexpert 2.5 or FlowJo 10.9.0.

Viral genome quantification

Infected-cell supernatants were combined 1:1 with 100 ng ml−1 RNase A and incubated for 30 min at 37 °C before addition of 40 U of RNasin ribonuclease inhibitor. RNase-treated supernatants were purified using a New England Biolabs Luna Cell Ready One-Step RT–qPCR kit. Treated samples (1 µl) were added to 9 µl lysis reactions for 10 min at 37 °C, followed by addition of 1 µl stop buffer. Lysate mix (2 µl) was combined with 18 µl qPCR master mix, with primers amplifying a region of the NP-gene segment conserved across all our influenza model strains (oNB23 and oNB24). The genome content of experimental samples was determined by comparing cycle threshold (Ct) for detection with a standard curve of sucrose cushion-purified virions. The packaged genome per virion was determined by comparing genome titre by qPCR to virion titre by flow virometry.

Cellular vRNA quantificationRNA extraction

Cell pellets containing ~5 × 105 cells were resuspended in 500 µl of TRIzol, followed by 5 min incubation at room temperature. Chloroform (100 µl) was added to each of the lysis mixtures. After 2 min, samples were centrifuged at 12,000 g for 15 min. Aqueous phases were collected and mixed with 250 µl of isopropanol. After 10 min at 4 °C, samples were centrifuged for 10 min at 12,000 g. Supernatant was removed and RNA was washed in 500 µl of 75% ethanol, followed by 5 min of centrifugation at 7,500 g. Supernatant was removed and pellets were air dried before dissolving in 50 µl of water.

Two-step qPCR

To amplify the (−) sense genomic vRNA within cells, a two-step RT–qPCR was performed on RNA extractions of infected cells. One microlitre of extracted RNA was combined with RT–qPCR master mix (New England Biolabs Luna Universal RT–qPCR kit). A primer specific to the (−) sense NP strand (oNB23) was added to the reaction for 10 min at 50 °C. Samples were placed on ice and the (+) sense NP primer (oNB24) was added. Samples were then PCR amplified. Viral genome content in each sample was compared to total RNA via a separate qPCR reaction using primers (oEP1591and oEP1592) against 7sk (a non-mRNA control). All sample Ct values were normalized to a model sample to obtain a relative quantification, with a standard curve constructed by 10-fold serial dilutions of the model sample.

Statistics and reproducibility

No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those of ref. 9. The only measurements excluded from analysis were shape measurements of Cal0709 at <20,000 particles per µl or other viruses at <1,000 particles per µl, where noise measured in negative samples reached ~10% of the virion count, skewing shape measurements. Data distribution was assumed to be normal, but this was not formally tested. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment. Except for qualitative observations in immunofluorescence (Fig. 5b,c), all analysis was automated and independent of human interpretation. For example, gates in flow virometry were set using a control and applied to all samples.

Reporting summary

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