Microtubule motors mediate endosomal sorting by maintaining functional domain organization

Abstract

Many microtubule motors have been shown to couple to endosomal membranes. These motors include dynein in addition to many different kinesin family members. Sorting nexins (SNXs) are central to the organization and function of endosomes. These proteins can actively shape endosomal membranes and couple directly or indirectly to the minus-end microtubule motor dynein. Motor proteins acting on endosomes drive their motility, dictate their morphology and affect cargo segregation. We have used well-characterized members of the SNX family to elucidate motor coupling using high-resolution light microscopy coupled with depletion of specific microtubule motors. Endosomal domains labelled with SNX1, SNX4 and SNX8 couple to discrete combinations of dynein and kinesin motors. These specific combinations govern the structure and motility of each SNX-coated membrane in addition to the segregation of distinct functional endosomal subdomains. Taken together, our data show that these key features of endosome dynamics are governed by the same set of opposing microtubule motors. Thus, microtubule motors help to define the mosaic layout of endosomes that underpins cargo sorting.

Keywords: Cargo sorting; Endosome; Microtubule motor; Tubulation.

Fig. 1.

Microtubule motors driving motility of SNX-labelled endosomal domains. (A–F) Frames depicting the movement of SNX1-coated endosomes, in control cells (A) and in cells depleted for different microtubule motor subunits (B–F). (A′–F′) Colour-coded representation of movement: immobile vesicles appear white, and moving vesicles appear with coloured tracks. Scale bars: 5 µm.

Fig. 2.

Quantification of motility of SNX-labelled endosome domains. (A–C) These histograms represent the different subpopulations of SNX decorated structures according to their motility. Long range movements are characterized by a range greater than 500 nm, short range less than 500 nm. The siRNA GL2 targeting the firefly luciferase (Elbashir et al., 2001) was used to transfect control cells throughout the present study. When depleted, dynein-1 heavy chain (DHC1, DYNC1H1), kinesin-1 (KIF5B), kinesin-2 (KAP3) and dynein light intermediate chains 1 and 2 (LIC1 and LIC2 respectively) impaired the motility of SNXs. Note that GFP-SNX1 endosomes couple to dynein-1 containing LIC2 and kinesin-1, GFP-SNX4 endosomes couple to dynein-1 containing LIC2 and kinesin-2, and GFP-SNX8 endosomes couple to dynein-1 containing LIC1 and kinesin-1. Note that experiments with defined outcomes were repeated three times (independently); error bars show s.d. Non-effective treatments were repeated twice. This is consistent throughout the present study. Asterisks indicate statistical significance from Mann–Whitney test comparing with GL2 controls (**P<0.01, ***P<0.001).

Fig. 3.

Tubulation of endosomal domains in the absence of coupled motor proteins. (A) Highly mobile tubulo-vesicular SNX-decorated structures are moving linearly and bidirectionally in control cells (GL2). Depletion of different motor subunits induced the generation of long tubules (arrows) and enlarged endosomes (asterisks) for SNX1- SNX4- and SNX8-coated endosomes. Scale bars: 5 µm. (B) This diagram represents the different phenotypes that could be observed in hTERT-RPE1 cells expressing GFP-tagged SNX. In non-treated cells, the percentage of tubules and their length is low as shown in case no. 1. The two parameters monitored, percentage tubules and tubule length could be affected by motor depletion as shown in cases 2–4. A tubulation index (T.I.) is calculated, where T.I. = % tubules×length of tubule, and is colour coded for clarity. (C) The proportion of tubules was counted against the number of SNX-labelled structures. (D) The length of tubules in µm measured in each case. (E) The tubulation index was calculated as in B, and bars are coloured for greater clarity. Red indicates case no. 4 with a high percentage of tubules and a greater length. Mirroring our other data, these results highlight the role of motors in the vesicle scission from endosomes. Asterisks indicate statistical significance from Mann–Whitney test comparing with GL2 controls (*P<0.05, ***P<0.001).

Fig. 4.

Microtubule motors direct segregation of endosomes into distinct functional domains. We determined the colocalization of different SNX proteins following microtubule motor depletion as a measure of functional coupling of motors to these domains. (A) The number of SNX4-labelled structures also labelled with SNX1 is increased following depletion of DHC1, KIF5B, or LIC2. (B) Colocalization of SNX8 and SNX1 following depletion of DHC1, KIF5B, or LIC2. Taken together, A and B indicate that DHC1, KIF5B, or LIC2 apply force to SNX1-labelled domains. (C,D) The number of (C) SNX1- or (D) SNX8-labelled endosomal domains also labelled with SNX4 is greatly enhanced following depletion of DHC1, KAP3, or LIC2. These data indicate that DHC1, KAP3, and LIC2 apply force to SNX4-positive structures. (E,F) The number of (E) SNX1- or (F) SNX4-labelled endosomal domains also labelled with SNX8 is greatly enhanced following depletion of DHC1, kinesin-1, or LIC1. These data indicate that DHC1, kinesin-1, and LIC1 apply force to SNX8-positive structures. Asterisks indicate statistical significance from Mann–Whitney test comparing with GL2 controls (***P<0.001).