Fig 2: The antiviral function of RABGAP1L relies on its catalytically active TBC domain and residues implicated in endosomal trafficking(A) Schematic representation of RG1L WT and the 421 mutant (RG1L 421) which lacks the C-terminal region downstream of the kin domain.(B) Immunofluorescence analysis of A549 cells stably expressing EV, RG1L WT, or RG1L 421. Cells were fixed and stained with the indicated antibodies. Scale bar represents 25 µm.(C) A549 cells stably expressing GFP, RG1L WT, or RG1L 421 were stimulated with IFNa2 (1,000 U/mL or mock) for 16 h prior to infection with WSN/33 (MOI 0.001 PFU/cell). Supernatants were collected after 48 h and titrated on MDCK cells.(D) Schematic representation of the TBC domain of RABGAP1L and the localization of mutants R584A (Rmut), Q621A (Qmut), R584A-Q621A (RQmut), and KK784EE (KKmut). KKmut has previously been shown to prevent interaction with the AnkB death domain (DD).(E) Western blot validation of RABGAP1L expression in A549 cells stably expressing RG1L WT or the indicated mutants.(F) Immunofluorescence analysis of cells described in (E) (here, EV was used as a control), fixed and stained with the indicated antibodies. Scale bar represents 25 µm.(G) Cells described in (E) were infected with WSN/33-Renilla (MOI 1 PFU/cell) following treatment with IFNa2 (1,000 U/mL or mock) for 16 h. The AUC was calculated from RLU values taken up to 11 h p.i.For (B), (E), and (F), representative data from three biologically independent experiments are shown. For (C) and (G), mean values from three biologically independent experiments are plotted, with error bars representing SDs. Individual data points are shown. Statistical significance was determined using one-way ANOVA following log transformation (*p < 0.05, **p < 0.002, ****p < 0.0001).See also Figure S3.

Fig 3: Hypothetical model of viral restriction by RABGAP1LIn WT cells, IAVs enter via endocytosis. Early endosomes mature into late endosomes where the low pH induces HA-mediated fusion between the viral and endosomal membranes, leading to the release of the viral genome into the cytoplasm and its transport to the nucleus for replication and transcription. Active GTP-bound RAB22A and RAB5, as well as EEA1, are known to localize to early endosomes and have been found to interact with each other (Kauppi et al., 2002; Zhu et al., 2009). The CORVET complex, including its core subunit VPS33A, associates with RAB5 on early endosomes (Spang, 2016; Wartosch et al., 2015). Maturation to late endosomes is marked by a RAB5 to RAB7 switch and dissociation of EEA1 and RAB22A (Rink et al., 2005; Zhu et al., 2009). LBPA is a marker found on late endo and endo-lysosomes (Gruenberg, 2020). The HOPS complex, including its core subunit VPS33A, is known to associate with RAB7 on late endosomes and to mediate fusion with lysosomes (Balderhaar and Ungermann, 2013; Poteryaev et al., 2010). In cells overexpressing RABGAP1L, RABGAP1L may associate with VPS33A at IAV-containing early endosomes and potentially promotes a more rapid maturation from early to late endosomes by deactivation of RAB22A and/or RAB5 (Itoh et al., 2006). Dissociation of the two RABs and EEA1 from early endosomes drives the maturation of endosomes, which fuse with lysosomes to form endo-lysosomes (Bright et al., 2016; Luzio et al., 2014). Association of RABGAP1L with VPS33A of the HOPS complex might enhance this process. Maturation to lysosomes could initiate the degradation of the endocytosed viral particle by lysosomal proteases, ultimately limiting infection. LAMP1 is found mainly on lysosomes. Proteins marked in blue are known to colocalize with IAV particles (Liu et al., 2014; Sieczkarski and Whittaker, 2003).

Fig 4: RNAi screening identifies RABGAP1L as an IAV restriction factor(A) Schematic representation of recombinant IAV WSN/33 in which the coding region for the hemagglutinin (HA) glycoprotein has been replaced by Renilla luciferase (WSN/33-Renilla).(B) RNAi-screening experimental workflow.(C) MRC-5-HA cells were transfected for 30 h with individual siRNAs targeting MX1 or IFITM3 or with a non-targeting (NT) control siRNA. Following stimulation with IFNa2 (1,000 U/mL or mock) for 16 h, cells were infected with WSN/33-Renilla (MOI 5 PFU/cell) in the presence of the live-cell substrate EnduRen. Luciferase activity was monitored up to 12 h post-infection (p.i.), and the area under the curve (AUC) was calculated as indicated. Mean values from 50 technical replicates across two independent biological experiments are plotted, with error bars representing SDs.(D) Hit criteria for RNAi screening. In a primary screen following the workflow in (B), 100 putative ISGs were silenced with four individual siRNAs each. Twenty-two genes met the threshold, and 20 were re-tested in a confirmation screen. Applying the same hit criteria, a total of 8 putative ISGs were confirmed in both screening rounds.(E) Heatmap showing Z scores of positive controls (MX1 and IFITM3) and the top 8 hits from the two RNAi-screening rounds. Columns represent individual siRNAs targeting genes listed in rows.See also Table S1.

Fig 5: RABGAP1L expression impacts host endosomal function and IAV uptake(A–C) A549 cells stably expressing RABGAP1L WT, the 421-truncation mutant, or EV were infected with WSN/33 (MOI 5 PFU/cell) for 1 h on ice. Three hours after incubation at 37°C, cells were fixed and stained with antibodies against RABGAP1L (red) and NP (green) (A). Nuclei were stained with DAPI (blue). Scale bar represents 25 µm.(B and C) Green mean fluorescent intensities (MFIs) of nuclear NP signals were quantified from fluorescent-microscopy images from (A) using ImageJ software. Individual cells are represented by single dots (B). Mean values of data from three biologically independent experiments in (B), normalized to EV, are shown in (C).(D) MDCK cells, expressing the constructs described in (A), were infected for 4 h at 37°C with WSN/33-pseudotyped ß-lactamase-matrix protein (BlaM1) fusion protein VLPs prior to quantification of entry-positive cells via flow cytometry. Data represent means, with error bars showing SDs, from three biologically independent experiments. Individual data points are shown.(E) Experimental setup for immunofluorescence-based confocal microscopy to track early stages during IAV entry. Following infection with WSN/33 (MOI 25 PFU/cell or mock) for 1 h at 4°C, cells were fixed at the indicated timepoints.(F) A549 cells stably expressing RABGAP1L (RG1L) or EV were subjected to the experimental setup described in (E). The MFI of HA signals (green) at 0 min p.i. were quantified from confocal-microscopy images shown in Figure S6A using ImageJ. Individual cells are represented by single dots.(G) Quantification of co-localizations between EEA1 and HA from confocal images shown in (H) and Figure S6A using Imaris. Individual cells are represented by single dots.(H) Immunofluorescence analysis of RABGAP1L or EV-expressing A549 cells treated as described in (E). Cells were stained for early endosomes (EEA1, magenta), viral proteins (HA, green), and nuclei (DAPI, blue). Scale bar represents 25 µm. White arrows indicate co-localizations between EEA1 and HA. Representative images of at least nine analyzed cells per time point from at least two biologically independent experiments.(I) Quantification of co-localizations between EEA1 and HA from confocal images shown in Figure S6B. Individual cells are represented by single dots.(J and K) Cells described in (A) were serum starved for 2 h prior to treatment with Dynasore (Dyn.; 100 µm) or DMSO for 1 h at 37°C. Cells were then incubated with Alexa-Fluor-488-conjugated transferrin (Tf-488) for 1 h at 4°C followed by a 10-min incubation at 37°C prior to fixation.(J) MFI quantification of Tf-488 signals from confocal-microscopy images shown in (K) using ImageJ software. Individual cells are represented by single dots.(K) Cells were stained with anti-transferrin receptor (TfR) antibody (magenta) and DAPI (blue) prior to analysis by confocal microscopy. Scale bar represents 25 µm. Representative images of at least 25 analyzed cells from two biologically independent experiments.Statistical significance was determined using unpaired nonparametric t test (B, F, G, and J), unpaired one-way ANOVA (C and D), or ordinary two-way ANOVA (I) (*p < 0.05, **p < 0.002, ***p < 0.001, ****p < 0.0001; ns, non-significant).See also Figures S5–S7.