RUFY3 regulates endolysosomes perinuclear positioning, antigen presentation and migration in activated phagocytes

Abstract

Endo-lysosomes transport along microtubules and clustering in the perinuclear area are two necessary steps for microbes to activate specialized phagocyte functions. We report that RUN and FYVE domain-containing protein 3 (RUFY3) exists as two alternative isoforms distinguishable by the presence of a C-terminal FYVE domain and by their affinity for phosphatidylinositol 3-phosphate on endosomal membranes. The FYVE domain-bearing isoform (iRUFY3) is preferentially expressed in primary immune cells and up-regulated upon activation by microbes and Interferons. iRUFY3 is necessary for ARL8b + /LAMP1+ endo-lysosomes positioning in the pericentriolar organelles cloud of LPS-activated macrophages. We show that iRUFY3 controls macrophages migration, MHC II presentation and responses to Interferon-γ, while being important for intracellular Salmonella replication. Specific inactivation of rufy3 in phagocytes leads to aggravated pathologies in mouse upon LPS injection or bacterial pneumonia. This study highlights the role of iRUFY3 in controlling endo-lysosomal dynamics, which contributes to phagocyte activation and immune response regulation.

Fig. 1. The FYVE domain bearing RUFY3 isoform is preferentially expressed in activated immune cells.

a Schematic representation of RUFY3 functional domains organization. RUN: RPIP8, UNC-14 and NESCA domain, CC1: coiled-coil 1 domain, CC2: coiled-coil 2 domain, FYVE: Fab1, YOTB, Vac1 and EEA1 domain. An amino acids alignment of the iRUFY3 FYVE domain with other known FYVE domains is displayed showing the absence of normally conserved Histidine tandem from the iRUFY3 and RUFY4 Zinc finger 1 domains. Color code: cyan for hydrophobic positions (A, V, I, L, M), turquoise for aromatic positions (F, Y, W, H), red for basic residues (K, R), purple for acidic residues (D, E), green for non-polar charged (N, Q, S, T), salmon for cysteine (C), orange for glycine (G) and yellow for proline (P). Br Brain, Sp Spleen and Bm Bone marrow. b Comparative expression of the different rufy3 isoforms in mouse tissues (RNA seq Immgen database). The shorter rufy3 transcript (nrufy3) is mostly expressed in brain, while the FYVE-bearing larger transcript (irufy3) is enriched in bone marrow, lymphoid organs and immune cells. c, d Immunoblot showing the expression of RUFY3 protein isoforms in brain, spleen, mesenteric (MLN) and iliac (ILN) lymph nodes with or without LPS stimulation in vivo (c) and (d) in bone marrow-derived DCs and Raw264.7 (RAW) macrophages in vitro. Actin is not detectable in HeLa i/nRUFY3 over-expressing control due to minimal sample loading (100 ng). Blots are representative of two independent experiments. e Immunoblot after non-denaturing PAGE revealing iRUFY3 dimerization in RAW Wild Type (WT) and rufy3 knock out (KO). Blot is representative of two independent experiments. f, g Quantification by RT-qPCR of the rufy genes family transcripts in Flt3L-bmDCs after 6 h exposure to TLR ligands and IFN-α (f) and of rufy3 over 6 h (g). For f, the boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. Each dot represents one independent experiment. Statistical significance was established using two-way ANOVA with Dunnett’s multiple comparisons test. For g, each dot represents one independent experiment (n = 3) and data are presented as mean values +/- SD. Statistical significance was established using one-way ANOVA with Tukey’s multiple comparisons test. h, i Immunoblot detection of iRUFY3 levels in Flt3L-bmDCs (h), and (i) quantification by RT-qPCR of the rufy3 mRNA and iRUFY3 protein by immunoblot in RAW stimulated or not with LPS. All data are presented as mean values +/− SD. n = 3 independent experiments except for IFNα condition where n = 4. Statistical significance was established using one-way ANOVA with Dunnett’s multiple comparisons test. For all panels (*p < 0,05; **p < 0,01; ***p < 0,001; ****p < 0.0001).

Fig. 2. iRUFY3 co-localizes with perinuclear LAMP1+ endosomes upon LPS activation or nutrients starvation.

a Airyscan Immunofluorescence confocal microscopy (AICM) panels showing iRUFY3 intracellular distribution in RAW rufy3^-/- cells stably expressing a control empty vector (left, RAW rufy3^-/-), nRUFY3 (middle, RAW-nRUFY3) and iRUFY3 (right, RAW-iRUFY3) at steady state (top), after 16 h of LPS stimulation (middle) and 6 h of nutrient starvation (EBSS, bottom), dotted lines indicate cell boundaries. Scale bar is 5 µm. These results are representatives of n = 3 independent experiments with >100 cells observed by experiments. b–d AICM images of RAW-iRUFY3 showing RUFY3 intracellular (myc) distribution compared to endocytic markers EEA1, Rab11A, Syntaxin-6 (STX6) and LAMP1 at steady state (b), after 16 h LPS stimulation (100 ng/mL) (c) or 6 h starvation (EBSS) (d). White arrows indicate co-localization (LAMP1) or absence of co-localization (Rab11a) with RUFY3. Pearson co-localization and Mander’s overlap coefficients (MOC) were calculated using Image J for b-d panels. A highly significant co-localization score is considered above 0.7 for Pearson’s. Statistical significance was established using two-way ANOVA with Tukey’s multiple comparisons test (*p < 0,05; **p < 0.01; ****p < 0,0001). Each dot represents the mean off all Z-stack from one region of interest. Scale bar 1 µm.

Fig. 3. iRUFY3 associates with perinuclear LAMP1+ endosomes upon LPS activation or nutrient starvation.

a AICM showing LAMP1+ vesicles and RUFY3 distribution in Raw264.7 macrophages (WT), rufy3 depleted (rufy3 -/-) and stably complemented with nRUFY3 or iRUFY3. Dotted lines indicate cell boundaries. Scale bar 2 µm. These results are representatives of n = 3 independent experiments with >100 cells observed by experiments. b 3D reconstruction showing iRUFY3 at steady state (top), after 16 h LPS stimulation (middle) or 6 h nutrient starvation (bottom). Scale bar 5 µm. c 3D display showing iRUFY3 intracellular distribution compared to LAMP1 at steady state, after 16 h LPS stimulation and 6 h starvation (EBSS). Scale bar 2 µm. X, Y and Z-axis are defined at the bottom right. d AICM showing iRUFY3 distribution and the microtubule network (ß-tubulin) in LPS-activated RAW-iRUFY3 cells. Scale bar 2 µm. This result is representatives of n = 2 independent experiments with >100 cells observed by experiments. e Analysis of cytosolic (Cy) and membrane (Mb) fractions from post-nuclear supernatants (PNS) of RAW and RAW rufy3^-/-. After LPS treatment, RUFY3 is enriched in Mb fractions that contain the type-II transmembrane Invariant chain p31/CD74 used as fractionation control marker. This blot is representative of n = 2 independent experiments.

Fig. 4. iRufy3 promotes LAMP1 + -lysosomes clustering and associates with ARL8b.

a Detection of myc-RUFY3 and LAMP1 + EL distribution by AICM in RAW-iRUFY3 quantified with centroid based method and 360-ASOD. Dotted lines indicate cell boundaries. Scale bar 2 µm. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. Statistical relevance was established using unpaired t-test (*p < 0,05; **p < 0.01; ***p < 0.001; ****p < 0,0001; ****p < 0,0001). b Distribution of iRUFY3 and ARL8b monitored by AICM in RAW-iRUFY3 and RAW rufy3 ^-/- with centroid based method. Colocalization score and clustering index are shown. A clear co-localization score is considered above 0.7 for Pearson’s. Scale bar 2 µm. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. Statistical relevance was calculated with one-way ANOVA with Tukey’s multiple comparison test (****p < 0,0001; ****p < 0,0001). c Immunofluorescence Proximity Ligation Assay (PLA) was performed for iRUFY3 and ARL8b with 3D image reconstitution in RAW rufy3 ^-/- (top), RAW-nRUFY3 (middle) and RAW-iRUFY3 (bottom) at steady state (left), 16 h LPS stimulation (center) or 6 h nutrient starvation (EBSS, right). Scale bar 5 µm. PLA spot quantification and distribution are shown. Clustering was quantified by nearest neighbor’s distance calculation for each PLA spot. Data are presented as mean values +/- SD. For all panels numbers (n) of cells analyzed are indicated. Statistical relevance was calculated by Welch’s t-test for clustering indexes, two-way ANOVA with Tukey’s multiple comparisons test for co-localization and nearest neighbor distance (*p < 0,05; ***p < 0.001; ****p < 0,0001; ****p < 0,0001).

Fig. 5. Silencing of ARL8b alters iRUFY3 expression and endosomal localization.

a Immunoblot detection and quantification of iRUFY3 and ARL8b expression upon Arl8b RNAi silencing in RAW-iRUFY3. N = 3 independent experiments. Data are presented as mean values +/- SD. b Immunoblot detection of iRUFY3 and ARL8b expression in WT and rufy3 ^-/- RAW. The blot is representatives of two independent experiments. Number below actin represents ARL8B intensity normalized on WT. c Detection of iRUFY3, ARL8b and LAMP1 by AICM upon Arl8b silencing in RAW-iRUFY3 at steady state, upon starvation (EBSS, 6 h) or after treatment with PIKfyve inhibitor (YM201636, 45 min, 5 µM). Dotted lines indicate cell boundaries and red stars indicate cells with particularly efficient ARL8b silencing. Scale bar is 1 µm. Mander’s Overlay Coeficient (MOC) was calculated using Image J. One dot represents the mean of all z stacks from one cell. Statistical relevance was calculated by unpaired t-test (a) and two-way ANOVA with Tukey’s multiple tests for co-localization (c). For (c) numbers (n) of cells analyzed are indicated. For all panels: *p < 0,05; **p < 0.01; ***p < 0.001; ****p < 0,0001.

Fig. 6. PtdIns(3)P contributes to iRUFY3 endosomal localization.

a, b Detection of Myc-RUFY3 with EL markers LAMP1 after 6 h of starvation (a) or 6 h with VPS34 inhibitor (VPS34i, 5 µM). Dotted lines indicate cell boundaries. Pearson co-localization coefficients were calculated using Image J. A clear co-localization score is considered above 0.7. Numbers (n) of cells analyzed are indicated. Statistical significance was established using one-way ANOVA test (****p < 0,0001). Each dot represents the mean of all stacks from one region of interest. c GFP-2X-FYVE transfection and iRUFY3 detection by AICM in RAW-iRUFY3 at steady state or nutrient starvation (EBSS) with or without VPS34 inhibition (VPS34i, 5 µM for 6 h). Scale bar 2 µm. These results are representatives of n = 2 independent experiments with >100 cells observed by experiments.

Fig. 7. RUFY3 is required for intracellular Salmonella replication in macrophages.

a Salmonella enterica replication was monitored by Colony Forming Unit (CFU) assay in RAW WT, RAW rufy3^-/-, RAW-iRUFY3 and RAW-nRUFY3 cells. After gentamicin treatment, infected cells were lysed 2 h or 16 h post-infection (p.i) and colonies counted on LB agar plates. Each dot represents one independent experiment. One independent experiment represents the mean of three technical replicates (three independents agar plate). Data are presented as mean values +/- SD. Statistical relevance was established using one-way ANOVA with Holm-Šídák multiple comparisons test (*p < 0.05; ***p < 0,001; ****p < 0,0001). b Number of bacteria at 2 h p.i indicates no difference in the phagocytic rate among the different cell lines tested. Each dot represents one infection and the count of bacteria in one agar plate. For RAW WT + Salmonella ∆sifA n = 12 and n = 15 for other conditions. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. c, d S. enterica was imaged by AICM together with iRUFY3, LAMP1 (c) or PipB2 (d) either at 16 h p.i (left) or 16 h p.i with 6 h of EBSS starvation (right). Dotted lines indicate cell boundaries. Scale bar 2 µm. These results are representatives of n = 5 independent experiments with >100 cells observed per experiments. Infected cells (Inf) and non-Infected (Ni) are indicated in the same panels.

Fig. 8. iRUFY3 regulates migration of macrophages and DCs.

a Velocity assay for WT and CD11c^cre rufy3^fl/fl GM-CSF-bmDCs. Data represents instantaneous mean velocities with SD of migration in 4 by 5 µm fibronectin-coated microchannels of at least 100 cells per condition. N = 2 independent experiments were performed and each dot represents the average speed of one cell. Statistical relevance was established using unpaired Man-Whitney test (****p < 0.0001). b Wound healing assay on WT, RAW rufy3 ^-/-, RAW-iRUFY3- and RAW -nRUFY3 at steady state (left) or over 24 h LPS stimulation (100 ng/mL). For each condition, simple linear regression was performed with 99% of confidence bands of the best-fit line. Statistical relevance was established using unpaired multiple t-test (*p < 0.05; **p < 0.01; ***p < 0.001). N = 3 independent experiments were performed, each one in triplicate. c RT-qPCR quantification of Fascin-1 from mouse brain extract and RAW264.7 macrophages stimulated or not with LPS for 16 h. N = 3 independent experiments were performed. Data are presented as mean values +/− SD. Statistical relevance was established using one-way ANOVA with Tukey’s multiple comparisons test (*p < 0.05).

Fig. 9. iRufy3 deletion alters antigen processing and presentation in APCs.

a Immunoblot quantification of RUFY3 protein levels in RAW cells stimulated with IFN-γ (50 ng/mL) during 24 h. N = 5 independent experiments. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. b Surface MHC-II quantification by flow cytometry after IFN-γ exposure for indicated time (50 ng/mL) in RAW and RAW rufy3 ^-/-. N = 5 independent experiments for all time point in WT condition. For RAW rufy3^-/- condition, n = 5 for 0 h, n = 3 for 6 h and n = 4 for 12 h, 18 h and 24 h. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. c Percentage of cells expressing total detectable MHC-II signal (MHC-II +) by microscopy after 24 h of IFN-γ stimulation. Data are presented as mean values +/- SD. d RT-qPCR quantification of MHC-II α and β chains, CIITA and IFN-γ receptor-1 transcript from RAW and RAW rufy3 ^-/- stimulated with IFN-γ for 24 h. Each dot is one independent experiment. e Distribution of iRUFY3, MHC-II and LAMP1 by AICM in RAW rufy3 ^-/- (right) and RAW-iRUFY3 (left) with or without IFN-γ stimulation. Scale bar 2 µm. Pearson co-localization coefficients were calculated using Image J. f Monitoring of OVA-DQ endosomal degradation using a fluorescence dequenching assay in response to IFN-γ stimulation. N = 4 with two duplicates. Data are presented as mean values +/- SEM. g RAW and RAW rufy3 ^-/- stimulated with IFN-γ for indicated times (h) were treated with the cysteine protease inhibitor LHVS (1 µM) for 6 h prior immunoblot detection of Ii chain (p41/p31) and associated proteolytic fragments (p10 and p22). N = 3 independent experiments. h ELISA dosage of Interleukin-2 released after ovalbumin antigen processing and presentation to DO.11.10 T cells by RAW rufy3^-/- cells and RAW after IFN-γ stimulation. N = 3 independent experiment with two technical replicates. Statistical relevance was calculated using in: a and e, unpaired t-test, b and d, multiple unpaired t-test, and c, f and g, two-way ANOVA with Tukey’s multiple comparisons test. For all panels: *p < 0,05; **p < 0,01; ***p < 0,001; ****p < 0.0001. WT means Wild Type.

Fig. 10. Rufy3 deletion in the CD11C+ cell compartment is pro-inflammatory.

a and c Immunophenotyping of myeloid (a) and lymphoid (c) splenocyte populations from CD11c^cre and CD11c^cre rufy3 ^fl/fl mice by flow cytometry after IP injection of PBS (left) or LPS (1,5 ng/mg) (right), shows signs of mild spleen inflammation. Red line represents the average cell numbers in the spleens of CD11c^cre rufy3 ^fl/fl relative to control animals (blue line). For PBS condition, 7 CD11c ^cre mice and 8 CD11c^cre rufy3 ^fl/fl were used. For low LPS, 3 CD11c ^cre mice and 4 CD11c^cre rufy3 ^fl/fl were used. N = 4 for PBS injection and N = 3 for LPS injection. b Splenomegaly was revealed by monitoring spleens weight from control and CD11c^cre rufy3 ^fl/fl animals after LPS IP injection (1,5 ng/g or 10 µg). N was added directly on graph. Each dot represents the ratio between the weight of the spleen and the total weight of the corresponding mouse. Statistical relevance was established using two-way ANOVA with Šídák’s multiple comparisons test (*p < 0,05; **p < 0,01). The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. d Graphical abstract of the acute pneumonia model. e Disease scores of WT and CD11c^cre rufy3 ^fl/fl mice upon E.coli inhalation and primary pneumonia (N = 10) recovery over 7 days. Kaplan-Meyer survival curve (left), daily weight monitoring (middle) and overall pain score (right) graphs show increased severity of the disease in infected CD11c^cre rufy3 ^fl/fl animals. N = 10. f Flow cytometry profiling of total lung immune cells indicates increased counts in IFN-γ-producing resident macrophages, CD8 + T and NK cells in infected CD11c^cre rufy3 ^fl/fl compare to control animals. N = 7 for CD11c^cre mice and n = 6 for CD11c^cre rufy3 ^fl/fl mice. The boxplot data represent medians, interquartile ranges and spikes to upper and lower adjacent values. Statistical relevance was established by multiple unpaired Mann–Withney test (*p < 0,05; **p < 0,01).