(A) K562 cells were stably transduced to express shRNA targeting IFITM1, 2, or 3 or non-targeting control shRNA (scrambled). IFITM protein expression was measured by Western blot, as in Fig. 1A. (B) shRNA-transduced K562 cells characterized in panel A were infected with the indicated pseudoviruses as described in Fig. 1C. (C) shRNA-transduced K562 cells characterized in panel A were infected with infectious DENV2 NGC strain as described in Fig. 1D. Error bars indicate standard error. Single and double asterisks indicate statistically significant (P<0.05 and P<0.005, respectively) differences between cells expressing IFITM1 and control shRNA for corresponding infection conditions. Experiment is representative of three with similar results.

(A) K562 cells were treated with the indicated concentrations of human interferon-α or mock treated, and lysed for Western blot analysis 48 h later, using an antibody that recognizes IFITM1 alone, or one that recognizes both IFITM2 and IFITM3. (B) K562 cells were stably transduced to express human IFITM1, 2, or 3 or with the control vector pQCXIP. IFITM protein expression was measured by Western blot using an anti-c-myc antibody. (C) Stably transduced K562 cells characterized in panel B were infected with an MLV-based retrovirus expressing EGFP and pseudotyped with MACV or IAV H5 entry proteins, as previously described [9]. Infection was determined by flow cytometry, and normalized to K562 cells transduced with vector alone. (D) K562 cells characterized in panel B were infected with infectious DENV2 NGC strain (labeled “DENV2” for virus only) at the indicated MOI, or the same amount of infectious DENV2 pre-opsonized with enhancing titers of antibodies against DENV structural proteins prM or E (“+2H2” and “+4G2” respectively) [8], [21]. Cells were washed after 1.5 h and incubated for ∼24 h. Intracellular staining of DENV antigen was performed with a DyLight-649-conjugated antibody against prM and infection was determined by flow cytometry. Experiment is representative of three with similar results. (E) An experiment similar to that shown in panel D was performed except that infection was assayed by measuring viral loads in the supernatant by plaque assays using BHK cells. (F) An experiment similar to that in panel D was performed, except that J774A.1 murine macrophage cells were stably transduced to express murine orthologs of IFITM1, 2, or 3 or with the control vector pQCXIP. Stably transduced cells were infected with infectious DENV2 NGC strain at ∼MOI 5 and incubated for ∼2 days before harvesting for flow cytometry. Error bars indicate standard error. Single and double asterisks indicate statistically significant (P<0.05 and P<0.005, respectively) differences between IFITM protein expressing and control cells for corresponding infection conditions.

(A) K562 cells stably transduced to express human IFITM1, 2, or 3 or with control vector were infected as in Fig. 1D, except that an MOI of 3.0 was used for non-opsonized DENV, and an MOI of 1.0 was used for DENV pre-opsonized with the antibodies 2H2 or 4G2. Different MOIs were used to achieve comparable infection levels between direct and ADE-mediated infection. (B) K562 cells stably transduced to express control shRNA or shRNA targeting IFITM1 were infected as described in Fig. 2C, except that an MOI of 1.6 was used for non-opsonized DENV, and an MOI of 1.0 was used for DENV pre-opsonized with the indicated antibodies. Standard error bars are shown. For each transduction condition, differences between direct infection and ADE-mediated infection were not statistically significant.