Interferon-induced transmembrane protein 3 blocks fusion of sensitive but not resistant viruses by partitioning into virus-carrying endosomes

Suddala, Krishna C.; Lee, Christine C.; Meraner, Paul; Marin, Mariana; Markosyan, Ruben M.; Desai, Tanay M.; Cohen, Fredric S.; Brass, Abraham L.; Melikyan, Gregory B.

Lipid mixing between two membranes is a necessary but not sufficient condition for complete fusion; lipids can diffuse through a hemifusion intermediate, without opening of a fusion pore HeLa cells were transduced with both IFITM3-iEGFP (green) and N-terminally tagged myc-IFITM3 (red). Cells were fixed, permeabilized, immunostained with anti-myc tag antibody, and counterstained with DAPI. Co-localization of EGFP-and myc-tagged IFITM3 appears as yellow punctae. (C) IFITM3-iEGFP expressed in A549 cells inhibits viral infection of varied doses (dilutions of virus inoculum) of Influenza A/WSN/33 compared to an empty vector control. (D) A549 cells were transduced with empty vector (left) or IFITM3-iEGFP (three panels to the right), infected with influenza A/WSN/33, and immunostained 16 hours later for expression of HA antigen on the cell surface, which indicates productive virus infection. IFITM3-iEGFP expressing cells are protected from virus infection. (E) IAV pseudovirus fusion is inhibited in IFITM3 expressing A549 cells. HIV-1 pseudoviruses carrying BlaM-Vpr and IAV A/WSN/33 HA/NA were bound in the cold to A549 cells transduced with Vector, IFITM3, IFITM3-imNeonGreen (IFITM3-imNG), IFITM3-imTFP1 or the inactive mutant IFITM3 tagged with mTFP1, 2M-IFITM3-imTFP1. Fusion was initiated by shifting to 37˚C and (e.g., [11, 30, 45, 46] ). We have previously shown that IFITM3 expression does not inhibit lipid mixing (hemifusion) between IAV and host endosomal membranes, but rather interferes with the formation of fusion pores [30] . The ability to visualize the distribution of functional fluorescent IFITM3 in living cells enables spatiotemporal analysis of viral fusion restriction. To test lipid mixing activity in the context of endosomes containing fluorescent IFITM3, we colabeled infectious IAV (A/PR/8/34 H1N1) with a self-quenching concentration of the lipophilic dye SP-DiI 18 and with the amine-reactive far-red Alexa Fluor647 dye (AF647) that labels surface glycoproteins of the virus [30] . As shown in the schematic diagram, single-virus hemifusion with an endosome can be detected by the appearance of bright SP-DiI 18 spots resulting from dilution of this dye within an endosomal membrane (Fig 2A) . Double-labeled IAV were incubated with cells at 4˚C for 20 min, followed by the addition of pre-warmed Live Cell Imaging Buffer (LCIB), and imaging continued at 37˚C. Representative snapshots from the timelapse movie illustrate the single IAV SP-DiI 18 (colored green) dequenching event around 25 min post-infection of Vector cells, indicating redistribution of the dye into an endosomal membrane ( Fig 2B, S1 Movie). Fluorescent traces obtained by single IAV tracking show an increase in SP-DiI 18 intensity over time, whereas the reference AF647 signal (red) remains relatively constant (Fig 2C) . From these traces, the time required for complete dequenching (Δt) and the extent of dequenching (ratio of the initial and final mean intensities I f /I i ) can be determined ( Fig 2C) .