Palmitoylome profiling reveals S-palmitoylation-dependent antiviral activity of IFITM3

Abstract

Identification of immune effectors and the post-translational modifications that control their activity is essential for dissecting mechanisms of immunity. Here we demonstrate that the antiviral activity of interferon-induced transmembrane protein 3 (IFITM3) is post-translationally regulated by S-palmitoylation. Large-scale profiling of palmitoylated proteins in a dendritic cell line using a chemical reporter strategy revealed over 150 lipid-modified proteins with diverse cellular functions, including innate immunity. We discovered that S-palmitoylation of IFITM3 on membrane-proximal cysteines controls its clustering in membrane compartments and its antiviral activity against influenza virus. The sites of S-palmitoylation are highly conserved among the IFITM family of proteins in vertebrates, which suggests that S-palmitoylation of these immune effectors may be an ancient post-translational modification that is crucial for host resistance to viral infections. The S-palmitoylation and clustering of IFITM3 will be important for elucidating its mechanism of action and for the design of antiviral therapeutics.

Figure 1. Visualization and identification of palmitoylated proteins in DC2.4 cells

(a) Metabolic labeling of cells with alk-16 palmitate reporter and subsequent CuAAC ligation with bioorthogonal detection tags for imaging or proteomics. (b,c) DC2.4 cells were incubated for 2 h with 50 mM alk-16 or DMSO as a control. In b, cell lysates were reacted with az-rho by CuACC, and proteins were separated by SDS-PAGE for visualization by fluorescence gel scanning. Coomassie blue staining demonstrates comparable loading. In c, cell lysates were reacted with az-diazo-biotin by CuAAC for enrichment of alk-16–labeled proteins with streptavidin beads and identification by mass spectrometry. For each identified protein, the ratio of peptide spectral counts from the alk-16 and DMSO samples was plotted. Several known palmitoylated proteins are shown in black. IFITM3, the highest-ranked candidate palmitoylated protein, is shown in red.

Figure 2. IFITM3 is S-palmitoylated on membrane-proximal cysteine residues

(a) Ifitm3 mRnA expression in DC2.4 cells after 4 h IFn-α treatment as measured by qRt-PCR. NT, non-treated. (b) predicted topology of IFITM3 showing the location of cysteine residues (gray shading) with respect to transmembrane domains. (c,d) HeLa cells were transfected with 2 μg vector, pCMV-HA-IFITM3 wild-type (WT) or cysteine mutants in six-well plates overnight and labeled with 50 μM alk-16 for 2 h. Cell lysates were subjected to anti-HA immunoprecipitation, reacted with az-rho by CuAAC, separated by SDS-PAGE and visualized by fluorescence gel scanning. Comparable protein loading was confirmed by anti-HA western blotting. In c, half of the indicated immunoprecipitations were treated with 2.5% (w/v) neutral NH₂OH before boiling and gel loading to hydrolyze protein–alk-16 thioester linkages.

Figure 3. Palmitoylation-dependent clustering of IFITM3 in the ER

(a,b) HeLa cells grown on coverslips in 12-well plates were transfected with 1 μg pCMV-HA, pCMV-HA-IFITM3 or pCMV-HA-IFITM3-palmΔ and stained with anti-HA (red) and TOPRO-3 (blue). Insets are enlargements of the white-squared regions. Scale bars represent 10 μm. (b) Cells were also stained with antibodies against cellular markers, calreticulin and LAMP1 (green), or were co-transfected with GFP-CD9 (green).

Figure 4. Antiviral activity of IFITM3 is regulated by palmitoylation

(a) HeLa cells grown in six-well plates were transfected with 2 μg pCMV-HA, pCMV-HA-IFITM3 or pCMV-HA-IFITM3-palmΔ and infected with influenza virus (PR8 strain) with multiplicity of infection (MOI) of 1 for 6 h. ND, not detected; NI, non-infected. Influenza virus NP mRNA levels and Ifitm3 expression levels were examined by qRT-PCR. *P = 0.027 by Student’s t-test; error represents s.d., n = 3. (b) HeLa cells grown in 12-well plates were transfected with 1 μg pCMV-HA, pCMV-HA-IFITM3, pCMV-HA-IFITM3-palmΔ or pEGFP-C1-CD9 and infected with influenza virus at an MOI of 1 for 6 h. Virus NP and IFITM3 protein levels were examined by flow cytometry using anti-NP and anti-HA staining, respectively. (c) HEK293T cells grown in 12-well plates were transfected with 1 μg pCMV-HA, pCMV-HA-IFITM3, pCMV-HA-IFITM3-palmΔ or pEGFP-C1-CD9 and infected with influenza virus at an MOI of 1 for 6 h. Non-transfected and transfected cells expressing the proteins of interest from the same culture were gated on (labeled “low” and “high,” respectively) and analyzed for the percentage of these cells that were infected (Supplementary Fig. 8). Data from flow cytometric analysis of cells staining positive for influenza NP or NS1. (d) percentages in c were normalized such that the difference in infection rates for vector control and HA-IFITM3 high was set at 100% antiviral activity. Error in c and d represents s.d., n = 4. Note: the HA-tag epitope is derived from an H3 influenza virus strain and is not present in the PR8 strain of H1N1 influenza virus.