Influenza A virus exploits transferrin receptor recycling to enter host cells

Significance We still have an incomplete understanding of the proteins influenza A virus (IAV) uses to enter the host cell. Here we demonstrate that IAV entry is diminished in the absence of transferrin receptor 1 (TfR1). Consequently, pharmacological targeting of TfR1 efficiently interferes with IAV replication. While this process could be mediated by an interaction of IAV HA with glycosylated TfR1, a “headless” TfR1 mutant still functions as a host entry factor for IAV in trans. Our “revolving door” model provides a plausible explanation as to why IAV is capable of entering such a broad range of cell types in culture and why the adaptation of the zoonotic avian IAV to mammalian cells involves primarily an adjustment to attachment factors rather than the entry machinery.

120aa-flag, 120aa Y20C-flag or wt-BFP) using 2 μl/μg of DNA of Trans-IT LT1 (Mirus) according to the manufacturer's protocol. Twenty-four hours post transfection, 293T medium was replaced with target cell medium. Target cells (A549 TFRC KO, CHO Pro5 or Lec1) were seeded in a 6-well plate at subconfluent density (50%). At 48 h post transfection, the 293T cell supernatant containing lentiviruses was harvested with a syringe and pressed slowly through a 0.44-μm sterile filter. The filtered supernatant was complemented with 8 μg/ml of Polybrene. Target cells were washed one time with PBS before addition of 2 ml of the supernatant-Polybrene mixture. Four hours after infection with lentiviruses, the 293T cell supernatant was removed from target cells and replaced with the appropriate medium. Two days post infection, target cells were split and subjected to selection using puromycin at 2 μg/ml for A549, 6 μg/ml for CHO Lec1 or 8 μg/ml for CHO Pro5. The efficiency of overexpression was assessed by WB. Cells generated using a vector expressing TfR1 are named TFRC OE, and those using an empty vector are named CTRL.

Minigenome assay
Minigenome assays were performed as described previously (10). Briefly, cells were transfected with pCAGGS expression plasmids for PB2, PB1, PA and NP and a human Pol1driven negative sense Luciferase reporter flanked by IAV promoter regions of the NS segment (11). pCAGGS Renilla luciferase was contransfected as a control. Dual luciferase Reporter assays were performed 24h post transfection according to the manufacturers instructions (Promega).

VLP entry assay and intracellular β-lactamase detection
Aliquots of VLP were thawed on ice. VLPs were diluted in fresh OptiMEM media to a ratio of 1:5 or 1:2 for subsequent infection of 293T or CHO cells respectively. First, VLPs were treated with 5 µg/mL of TPCK-trypsin for 20 min at 37 °C in order to activate the HA at the surface of the VLPs. The TPCK-trypsin was inactivated using 10 µg/mL of trypsin inhibitor from Glycine max (Soybean) (Sigma #T6414) for 20 min at 37 °C. Then 1 % DEAE was added and ammonium chloride was added to a final concentration of 20 mM if needed. Prior to infection, the target cells (293T or CHO approximately 80 % confluent) were washed twice with PBS. For a 24-well plate, 400 µL of conditioned VLPs were added per wells. Cells were incubated 3 h at 37 °C.
Following infection, cells were detached by pipetting and centrifuged for 5 min at 300 x g.

LDL uptake assay
Cells were incubated with LDLDyLightTM 550 (Abcam) for four hours and LDL uptake was visualized by standard fluorescence microscopy, comparing to untreated cells. Cell density was monitored by brightfield imaging.

Flow cytometry
For detection of TfR1 on the cell surface after ferristatin II treatment, A549 cells were seeded in 12 well plates at a subconfluent density (80%). The cells were incubated for 4h with 1:1000 DMSO and 50um of Ferristatin II. After the incubation, cells were washed with PBS, detached using trypsin and centrifuged at 300g for 10 mins at 4°C. The supernatant was removed and the cells were washed once with PBS 1% BSA 0.01% NaN3. The cells were centrifuged at 300g for 10 mins at 4°C and the supernatant was removed. The cell pellets were stained with CD71 Mouse anti-Human, PE (BD Biosciences #561938) at 1:250 for 1h at 4°C in dark. After the staining, the cell pellets were washed twice with PBS 1% BSA 0.01% NaN3 followed by fixing with PBS 2% formaldehyde for 20 mins at 4°C. The cells were centrifuged at 300g for 10 mins and the supernatant removed. The cells resuspended in PBS and analysed at the UNIGE Flow Cytometry facility on a Beckman Coulter Cytoflex typically recording 10,000 events. Samples were gated on live cells and analysed for the surface expression of PE using FlowJo 10.7.1 software.
For uptake of fluorescent beads cells, A549 and CHO cells were seeded in 12 well plates at 80 % density. In line with the IAV infection experiments the cells were incubated with 10 or 1000 yellow-green FluoSpheres (Thermofisher #F8848)/cell. Beads were incubated for 45 min or 5 h on the cells. After the incubation, the cells were washed three times with PBS, detached using trypsin and centrifuged at 1500rpm for 5 mins. The supernatant was removed and the cells were resuspended in PBS and analysed at the CMU Flow Cytometry facility on a Beckman Coulter Cytoflex typically recording 10,000 events. Samples were gated for live cells and analysed for the expression of FITC using FlowJo 10.7.1 software. The gating strategies visualized in Supplementary Fig. 19-21.

Purified proteins
Trimeric HA of A/California/04/2009 was previously described (12). Based on this sequence we added the coding sequence for HRP on the 3' end of the T4foldon, connected via a GSGSGlinker and followed by a His10 tag (sequence provided in Supplementary Fig. 22-24). The Y98F mutant was derived by targeted mutagenesis. A trimerized HRP control was designed with the same T4foldon and His10 tag. The CDS for a recombinant TfR1 ectodomain was previously published (13). All constructs were synthesized by GeneArt and provided in a pFastBac1 vector for production of baculoviruses at the protein core facility of the CMU.
Briefly, recombinant secreted proteins were expressed in baculovirus infected Sf9 insect cells.
Baculovirus were generated using a modified pFastBac vector encoding C-terminally tagged i) WT Cal9, ii) mutated Cal9 or iii) no protein. All three proteins had a C-terminal tag composed of fused HA epitope -HRP and a 10-histidine tag. The gene encoding the transferrin receptor was placed following a gp67 secretion signal. All proteins were purified following a similar protocol.
Media containing the proteins of interest was clarified by centrifugation at 4000 g for 15 min at 4 °C and filtered using 0.22 micrometer filters. Media was concentrated to 50 ml final, adjusted to 10 mM imidazole concentration and applied to a 5 ml His-trap FF column (Cytiva).
Column was washed with 100 ml of PBS supplemented with 1 M NaCl and 10 mM Imidazole and eluted with 15 ml of elution buffer (1 x PBS, 200 mM NaCl, 450 mM imidazole). Eluted protein was concentrated to 1 ml using AMICON 30 MWCO concentrators and loaded on a Size Exclusion Chromatography Superdex 200 10/300 column at 4 °C equilibrated in PBS.
Fractions containing pure protein were pooled, concentrated and flash frozen in liquid nitrogen.

TfR1-HA pull down in 293T
Subconfluent 293T cells were transfected with either 1 μg of pDisplay.Flag-TfR1ecto or 250 ng of pCAGGS.Flag-ZsGreen as indicated. 48 h post transfection, cells were washed twice and lysed in 500 μL of IP Lysis buffer (50 mM Tris HCl pH 7.5, 150 mM NaCl, 0.5 % v/v NP40, 5 mM EDTA, with protease inhibitors). The lysates were incubated on ice for 15 min, sonicated twice for 10 sec using microtip, output 6-9W and cleared by centrifugation at 10000 xg at 4 °C for 20 min. Whole cell lysates were prepared by mixing 50 μL of lusate with 50 μL of 2X Protein Lysis buffer (Tris-glycerol-SDS-DTT). The anti-FLAG M1 agarose affinity gel (Sigma # A4596) slurry (30 μL per condition) was washed thrice with IP lysis buffer and resuspended in 600 μL of IP lysis buffer (per condition). For immunoprecipitation, 400 μL of lysate and 600 μL of beads in buffer were incubated over night at 4 °C on a rotating wheel. The next day, beads were washer 3 times with IP lysis buffer without protease inhibitors, resuspended in 500 ul IP lysis buffer with protease inhibitor and 1 μg of either HA wt or Tf was added as indicated.
Tubes were incubated over night at 4 °C on a rotating wheel. The next day, beads were washer 3 times with IP lysis buffer without protease inhibitors, all buffer was removed and and 50 μL of 1X Protein Lysis buffer (0.05M Tris, pH6.8, 10%v/v glycerol, 0.05M DTT, 2% w/v SDS) added.
Samples were boiled at 95 °C for 5 min, spun at 10000 xg for 10 min and 15 ul were loaded on a SDS-PAGE and western blotting analysis.

TIRF microscopy
Microscopic analysis of interaction between VLP and TfR1-BFP or BFP expressing CHO cells was performed on a Nikon Eclipse Ti with perfect focus system with a 100x/1.49 NA oil immersion objective and a Hamamatsu Orca-Fusion C15440-20UP CMOS camera. Cells were cultured on glass-bottom petri dishes for 4 h before moving them to the microscope. Cells

TIRF data analysis
VLP were tracked with the Fiji software package (14) and the plugin TrackMate (15). Before tracking, movies were drift corrected with the 'Descriptor-based registration' plugin based on VLP that adsorbed to the glass surface around cells and that did not move during the time course of the movie. Tracking was performed on VLP imaged with TIRF in TrackMate. Only tracks with the following conditions were considered for further analysis: 1) following definition of the cell edges, we used the LoG Detector with an estimated object diameter of 6 pixels, 2) the initial threshold was left untouched while filters on spots was adjusted depending on the experimental background, 3) then as a tracker, we used the simple LAP tracker with a linking max distance and a gap-closing distance of 10 pixel, the gap-closing max frame gap was set to 2, 4) all tracks shorter than 3 frames were filtered out, 5) the tracks and branch hierarchy analisis were export as .cvs files for further analysis. Intensity of TfR-BFP or surface expressed BFP was corrected for bleaching by measuring TfR intensity at the beginning and at the end of movies and by correcting for the loss in fluorescence in all time frames.

TIRF Software availability
Relevant data was extracted from TrackMate-exported files via custom-written code in a Jupyter notebook. The notebook is available from https://github.com/Mitchzw/viral-tracking

2-step-Enzyme-linked lectin assay
Recombinant N-terminal human IgG1 Fc-tagged transferrin (Cayman, produced in HEK293T cells) was coated o/n at 4°C onto Nunc MaxiSorb plates (1µg/ml, 100µl/well). All washing steps were done using 200µl of buffer. Plates were washed 3x with PBS + 1% BSA. PNGase (or PBS) was added (1U/well) and incubated at 37°C for 45min. Plates were washed 3x with PBS + 1% BSA and blocked for 1h at 37°C with PBS + 1% BSA. Plates were washed 1x with PBS + 1% BSA. 5x 10e8 pfu of PR8 was added per well or an equivalent dilution of allantoic fluid in PBS in presence or absence of 0.4ng/ml of holotransferin (Sigma). Plates were incubated 2h at 4°C. Plates were washed 3x with PBS + 1% BSA and 3x with PBS.Fetuin was added at 5 µg/ml  Only tracks longer than 6 frames were considered. Tracks which resulted in entry before frame 30 (entry, n=199) were compared to those, which did not enter until frame 30 (no entry, n=113). Only the first three and the last three frames of each track were considered.                   1174. Supplementary Figure 22 The plots show the GC content in a 40 bp window centered at the indicated nucleotide position. 1174. 2692.

In vivo infection experiments
The plots show the GC content in a 40 bp window centered at the indicated nucleotide position. 1174. Supplementary Figure 24