Recycling endosome tubule morphogenesis from sorting endosomes requires the kinesin motor KIF13A

Abstract

Early endosomes consist of vacuolar sorting and tubular recycling domains that segregate components fated for degradation in lysosomes or reuse by recycling to the plasma membrane or Golgi. The tubular transport intermediates that constitute recycling endosomes function in cell polarity, migration, and cytokinesis. Endosomal tubulation and fission require both actin and intact microtubules, but although factors that stabilize recycling endosomal tubules have been identified, those required for tubule generation from vacuolar sorting endosomes (SEs) remain unknown. We show that the microtubule motor KIF13A associates with recycling endosome tubules and controls their morphogenesis. Interfering with KIF13A function impairs the formation of endosomal tubules from SEs with consequent defects in endosome homeostasis and cargo recycling. Moreover, KIF13A interacts and cooperates with RAB11 to generate endosomal tubules. Our data illustrate how a microtubule motor couples early endosome morphogenesis to its motility and function.

Figure 1. KIF13A localizes to recycling endosomes and controls their distributions in a microtubule-dependent manner

(A–B) IFM on KIF13A-expressing cells that internalized Tf-A555 (30 min) (A) before or after labeling for RAB11A (B). KIF13A co-distributes with RE (arrows) and generates RE tubules (B, arrows; and Movies 1–2). (C–D) IFM on KIF13A-expressing cells incubated with 10 µM nocodazole (NZ) and co-labeled for α–tubulin and TGN46 (C) or for TfR after Tf-A647 internalization (30 min) (D). NZ treatment disrupts the microtubule network and disperses TGN46 labeling (C), and prevents KIF13A-dependent RE tubulation (C–D, 1^st panel) without affecting KIF13A endosomal localization (D, arrows). (E–F) IFM on KIF13A-ST-expressing cells that internalized Tf-A555 (30 min) (E) or after labeling for RAB11A (F). KIF13A-ST co-distributes with RE (arrows) without generating tubules (arrows; and Movie 3). Bars, 10 µm. (See also Figures S1–S2–S3 and Movies 1–2–3).

Figure 2. KIF13A controls the formation and function of recycling endosome tubules

(A) Western Blot (WB) of control (Ctrl)- or KIF13A-inactivated cell lysates (two different siRNAs -KIF13A#1 or #5-) probed for KIF13A or calnexin antibodies (loading control). (B) IFM on Ctrl- and KIF13A-inactivated cells pulsed with Tf-A555 and EGF-A647 (30 min), chased (30 min) and labeled for EEA1. EGF compartments were distinct from Tf- or EEA1-SE (arrowheads). (C) Conventional EM on Ctrl- or KIF13A-inactivated cells that internalized Tf-HRP and processed for DAB/H₂O₂ cytochemistry. Electron-dense Tf-HRP localized to tubulovesicular structures in Ctrl cells (arrows) while accumulated within MVEs in KIF13A-depleted cells (arrowheads). (D) Tf-HRP-positive MVEs (reported as a fraction of total MVEs) increased up to 85±4 % in KIF13A-depleted cells (n=238) as compared to control (36±2 %; n=96). Data were presented as a mean ± SD. ***, p<0.001. (E) The recycling of Tf was measured in Ctrl- (black star) or KIF13A-siRNA-treated cells (KIF13A#1, white star; #5, black triangle). Intracellular Tf percentage was plotted according to the time of chase. Both siKIF13A inhibited 15% of the Tf recycling by 6 min of chase. Data are presented as the average of 3 independent experiments, normalized to control and presented as a mean ± SD. (F) Ctrl- and KIF13A-inactivated cells that internalized Tf-A555 were captured by TIRF microscopy. Mobile Tf-positive structures were quantified, normalized to their total number and decreased upon KIF13A inactivation as compared to Ctrl. Data are presented as a mean ± SD. **; p< 0.005. (G) Ctrl- and KIF13A-inactivated cells were processed as in B, chased for the indicated time, analyzed by WB probed for TfR, EGFR (positive control), Calnexin (loading control) or KIF13A (depletion control) antibodies. Data are from one experiment representative out of three. Molecular masses, kDa. Bars: IFM, 10µm; EM, 500 nm. (See also Figure S4).

Figure 3. KIF13A associates dynamically and interacts with active RAB11

(A) Concomitant movements of KIF13A-YFP and mCherry-RAB11A were captured by spinning-disc microscopy. KIF13A induced an extensive tubulation and peripheral redistribution of RAB11A (see Movie 2). Magnified insets (of boxed area) of consecutive time-lapse images (image/6 sec) showed that KIF13A localized to the tip of RAB11A-positive tubule (arrows). (B) Kif13a^−/− MEF cells were co-transfected for mCherry-RAB11A together with GFP, KIF13A-YFP or GFP-KIFF13A-ST. RAB11A localized to vesicles apposed to nuclei of GFP- or KIF13A-ST-cells, while KIF13A redistributed RAB11A towards the cell periphery (white contours). (C) Domain structure of KIF13A revealed the motor domain (1–351), a Stalk region (352–1307) and a C-terminal Tail domain (1308–1770). (D) Y2H was performed to detect KIF13A/ RAB11 interactions. KIF13A Stalk and Tail interact with the GTPase-deficient (Q/L)-RAB11A, -RAB11B and wild-type RAB25. (E) Cell lysates incubated with GST, GST-RAB6A, -RAB4A or -RAB11A preloaded with GDP or GTPγs were analyzed by WB using KIF13A antibodies (top panel) or stained with coomassie blue (loading control; bottom panel). Lysates of Ctrl- and KIF13A-inactivated cells and experimental input showed KIF13A enrichment in RAB11A-pulled-down extracts. (F) Anti-Flag immunoprecipitations on lysates of cells co-expressing KIF13A-YFP, GFP-KIF13A-ST, GFP-KIF13A-T or GFP together with Flag-RAB11A or KIF13A-YFP with and empty Flag vector were analyzed by WB using GFP (upper panel) or Flag antibodies (lower panel). Inputs correspond to 10% of the whole cell lysates. (G) YFP fluorescence intensity (left) and FLIM (right) images of KIF13A-YFP-positive cells co-transfected (bottom) or not (top) with mCherry-RAB11A. The color-coding bar of FLIM images indicates YFP fluorescence lifetime value. (H) Fluorescence lifetime decreased in co-transfected cells indicating FRET between KIF13A and RAB11A in Hela (*; p<0.05) and MNT1 cells (***; p<0.001) as compared to control - no FRET was measured between KIF13A and RAB30 in HeLa cells. Molecular masses, kDa. Bars, 10 µm. (See also Movies 2 and 4).

Figure 4. KIF13A and RAB11 cooperate to generate endosomal tubules

(A–B) Ctrl- and triple RAB11 (A/B/25)-depleted cells transfected (A) or not (B) with KIF13A were pulsed with Tf-A647 (30 min), chased (20 min), fixed and labeled for RAB11A (A) or EEA1 (B). RAB11-depleted cells (A, white star) accumulated either enlarged KIF13A- and Tf-positive endosomal compartments (arrows; p<0.001) or enlarged EEA1- and Tf-positive SE without KIF13A overexpression (B, arrows) as compared to control. (C) Proposed model for KIF13A function. The TfR bound to Tf is internalized via clathrin-coated pits (1) that uncoat and fuse with vacuolar sorting endosomes (SE) (2). TfR segregates rapidly within tubular domains for recycling to the plasma membrane by two distinct routes: a fast RAB4-dependent pathway (3) or a slow RAB11-dependent route mediated by tubular transport intermediates (4). KIF13A functions at the SE and drives the generation, extension and peripheral transport of RAB11-dependent recycling endosome (RE) tubules in a microtubule- (MT) dependent manner. The formation of RE likely requires the interaction of KIF13A with RAB11 active form at specific SE subdomains, allowing the coordination between recycling of cargoes and proper maturation of SE. Bars, 10 µm.