Fission of SNX-BAR-coated endosomal retrograde transport carriers is promoted by the dynamin-related protein Vps1

Abstract

Retromer is an endosomal sorting device that orchestrates capture and packaging of cargo into transport carriers coated with sorting nexin BAR domain proteins (SNX-BARs). We report that fission of retromer SNX-BAR-coated tubules from yeast endosomes is promoted by Vps1, a dynamin-related protein that localizes to endosomes decorated by retromer SNX-BARs and Mvp1, a SNX-BAR that is homologous to human SNX8. Mvp1 exhibits potent membrane remodeling activity in vitro, and it promotes association of Vps1 with the endosome in vivo. Retrograde transport carriers bud from the endosome coated by retromer and Mvp1, and cargo export is deficient in mvp1- and vps1-null cells, but with distinct endpoints; cargo export is delayed in mvp1-null cells, but cargo export completely fails in vps1-null cells. The results indicate that Mvp1 promotes Vps1-mediated fission of retromer- and Mvp1-coated tubules that bud from the endosome, revealing a functional link between the endosomal sorting and fission machineries to produce retrograde transport carriers.

Figure 1.

The Vps10 sorting receptor accumulates in retromer-decorated endosomes in mvp1Δ-null mutant cells. (A) Micrographs showing Vps10 (Vps10-GFP) coexpressed with either Sec7-tomato (Sec7-tom) or Vps17-tomato in wild-type and mvp1 Δ cells are shown. Bar, 1 µm. (B) Distribution of Vps10-GFP between the Golgi and endosome in wild-type and mvp1Δ cells. The proportion of Vps10-GFP that localizes to Golgi compartments decorated by Sec7-tomato and endosomes decorated by Vps17-tomato in wild-type and mvp1Δ cells is shown.

Figure 2.

Mvp1 and Vps17 retromer SNX-BAR coat tubules that bud from the endosome. (A) Micrographs of Mvp1-GFP–, Vps17-GFP–, or Vps26-GFP–decorated endosomes captured by time-lapse fluorescence deconvolution microscopy. Arrowheads point to a tubule of interest and open arrows (<) indicate a presumptive fission event. Images were captured at a single focal plane within cells and acquired at 270-ms intervals. (B) Micrographs showing tubules emanating from Vps17-GFP– and Mvp1-2xmRFP–decorated endosomes were captured by time-lapse fluorescence microscopy. An example of a tubule decorated by both Vps17-GFP and Mvp1-2xmRFP (“Mixed”) is shown on the top, a tubule decorated only by Vps17-GFP in the middle, and a tubule decorated by only Mvp1-GFP on the bottom. (C) Quantitation of the different types of SNX-BAR–coated tubules. (D) Micrographs showing Vps17-GFP–coated (top) or Mvp1-GFP– coated (bottom) tubules emanating from an endosome also decorated by Did2-mKate2. Image capture speeds for each strain was 50 ms for the GFP channel and 200 ms for the RFP channel. Bar, 500 nm.

Figure 3.

Retromer- and Mvp1-coated SNX-BAR–coated tubules contain retrograde cargo. (A) Galleries of representative time-lapse micrographs showing endosomes from cells coexpressing Vps10-2xmRFP and either Mvp1-GFP or Vps17-GFP and cells coexpressing Vps10-2xGFP with Did2-mKate2. A single tubule colocalized for both Mvp1 and Vps10 is shown in the top row. The middle row shows a Vps17-mGFP–decorated endosome and three types of tubules; a tubule decorated by both Vps17-GFP and Vps10-2xmRFP (filled arrowhead), a tubule containing Vps10-2xmRFP but not Mvp1-GFP (<), and a tubule containing Vps10-2xmRFP but not Vps17-GFP (*). The bottom row shows time-lapse micrographs of a Did2-mKate2 endosome from which a tubule containing Vps10-GFP, but not decorated by Did2-mKate2, is budding (arrowhead). The end point (milliseconds) of each exposure is indicated on the top. Bar, 500 nm. (B) The proportion of tubules decorated by Vps17-GFP and containing Vps10-2xmRFP in wild-type and mvp1Δ cells is plotted. (C) The proportion of tubules in cells expressing Mvp1-GFP and Vps10-2xmRFP is plotted.

Figure 4.

Mvp1 possesses potent membrane remodeling activity. (A) Gallery of micrographs of negative-stained liposomes incubated with SNX-BAR Mvp1 at indicated concentrations. Bar, 200 nm. Montages of micrographs for each condition showing representative vesicles or tubules generated by Mvp1 are shown. (B) Histogram of the diameter of vesicles generated by Mvp1 (n = 129 vesicles) in 5-nm bins. (C) Box-plot compares the distribution of vesicle diameters generated by Mvp1. Average vesicle diameters of liposomes alone are 171.6 ± 96.8 nm, n = 55, and with Mvp1 are 54.4 ± 12.3 nm, P < 0.0001 (unpaired t test). (D) Histogram of the diameter of tubules generated by 10 or 20 µM Mvp1 in 5-nm bins. 10 µM Mvp1 (n = 29) generated average tubule diameters of 63.3 ± 15.3 nm, and 20 µM Mvp1 (n = 9) generated average tubule diameters of 35.8 ± 18.2 nm, P < 0.0001 (unpaired t test).

Figure 5.

Vps1 dynamin family protein localizes to SNX-BAR–coated endosomes. The top row shows diploid cells expressing one untagged copy of Vps1 and one GFP-tagged copy of Vps1 (Vps1-GFP) in cells coexpressing Mvp1-2xRFP (top) or Vps17-2xRFP (middle). Vacuoles (blue) were visualized by staining with the vital dye, CMAC. An expanded view of the boxed region in the “merge” column is shown in the “inset” column. The bottom row shows Vps1-GFP in homozygous diploid mvp1Δ mvp1Δ cells. One deconvolved image of a z-series containing the endosome or vacuole marker is shown. Bar: (main panels) 5 µm; (insets) 2.5 µm. Two methods for representing colocalization of Vps1-GFP with Vps17-2xRFP are depicted in B and C. Refer to text for a detailed description of the quantitation methods. (B) Distribution of the proportion of Vps1-GFP and Vps17-2xRFP double-positive endosomes. Wild-type (dark blue) and mvp1Δ mvp1Δ (light blue) cells (n = 20) were binned according to the proportion of Vps1-GFP and Vps17-2xRFP double-positive endosomes. (C) The proportion of endosomes (wild type: n = 64 of endosomes; mvp1Δ: n = 82 endosomes derived from cells analyzed in B) decorated by both Vps1-GFP and Vps17-2xRFP are shown. Standard deviations are indicated; P < 0.0001 (unpaired t test).

Figure 6.

Loss of Vps1 dynamin family GTPase impacts endosome number and morphology. (A) The indicated GFP-tagged endosomal proteins and Vps10 were visualized in wild-type and vps1Δ cells. Maximum projections of deconvolved z-stacks are shown. Bar, 5 µm. (B) Quantitation of SNX-BAR–, retromer-, and Sec7-decorated (Golgi) compartments in wild-type and vps1Δ cells. The number of puncta per cell was determined by masking each new punctum in each z-section. The average number of puncta per cell, wild-type vs. vps1Δ, and standard deviation, is shown for each of the indicated proteins. Every endosome marker showed significantly more puncta per cell: Vps17, 8.2 ± 2.4 vs. 14.8 ± 6.8; Mvp1, 7 ± 2.5 vs 11.1 ± 1.7; Vps26, 12 ± 2.5 vs. 21 ± 2.5; Sec7 (Golgi), 10.1 ± 1.4 vs. 10.9 ± 1.9; n = 30 for each strain, P < 0.0001 (unpaired t test), wild-type and vps1Δ, respectively. (C) To better visualize vacuole localization of Vps10 in vps1Δ cells, a single plane of the z-section containing vacuoles from the maximum projection shown in A is shown. Bar, 5 µm. (D and E) Thin-section electron micrographs of multivesicular endosomes in wild-type and vps1Δ cells (F and G) with endosomes highlighted in reference images at low magnification (D and F; bar, 500 nm) and shown at high magnification (E and G; bar, 100 nm). (H) Multivesicular bodies in vps1Δ cells are larger than in wild-type cells (unpaired t test, P = 0.0084; vps1Δ 193.5 nm ± 7.267, n = 28; wild-type 158.3 nm ± 9.179, n = 12).

Figure 7.

Loss of Vps1 dynamin family GTPase results in deficient fission of SNX-BAR–coated tubules. (A) Subcellular fractionation analysis of Vps10-myc. Spheroplasts derived from the indicated strains were lysed and a low-speed membrane pellet (13,000 g) and a high-speed membrane pellet (100,000 g) were generated. Equivalent amounts of each fraction were separated by SDS-PAGE and immunoblotted with anti-myc antibody. The position of full-length Vps10-myc is indicated, as well as degradative products that arise as a result of proteolysis by vacuolar proteases. The positions of molecular mass (kD) markers are indicated. (B) Time-lapse gallery of Mvp1-GFP– and Vps17-coated endosomes in wild-type and vps1Δ cells by time-lapse fluorescence deconvolution microscopy. Arrows highlight a tubule of interest in each time-lapse series and open arrow (<) indicates a fission event. (C) Tubule lifetimes were calculated as described in the text for Mvp1-GFP– or Vps17-GFP–coated tubules. Vps17 tubule lifetimes were significantly increased in vps1Δ cells (P = 0.0005) and in mvp1Δ cells (P = 0.043). (D) Vps17 tubule frequencies were calculated as described in the text. The frequency of Vps17-GFP tubules increased approximately twofold in mvp1Δ cells (P < 0.0001).

Figure 8.

Tomographic representations of vps1Δ cells reveal aberrant endosome morphology. (A and B) Multivesicular bodies observed by electron tomography (ET) are spherical (A and B, yellow models, n = 7; average diameter = 131 nm, similar to wild-type) and tubular (B and C, arrows and purple models, n = 3; average diameter, ∼50-nm diameter, variable length). Spherical MVBs observed by ET are smaller than those observed by thin-section EM quantitation. Small cisternal endosomes are also observed associated with endosomes (B and C, orange models). Tubular endosomes had fewer ILVs and were associated with other endosomes or vacuoles (B and C). (D) Intraluminal vesicle diameters from vps1Δ cells are significantly larger than wild-type (unpaired t test, P = 0.0001; wild type 25.55 nm ± 0.2307, n = 333; vps1Δ 30.65 nm ± 0.9362, n = 102). (E) Multivesicular bodies in vps1Δ cells have increased limiting membrane surface area and decreased ILV surface area compared with MVBs in wild-type cells. The distribution of endosomal membrane and internalized intraluminal vesicle membrane in wild-type and vps1Δ cells are plotted. In wild-type cells the surface areas of MVB and ILV are nearly equal (paired t test, P = 0.7434). In vps1Δ cells there is a loss in ILV surface area compared with the MVB surface area (paired t test, P = 0.0002). The mean ILV surface area generated in vps1Δ cells compared with wild-type (unpaired t test, P = 0.0049; wild-type 58924 ± 8591 nm², n = 12; vps1Δ 27751 ± 6069 nm², n = 20). The mean surface areas of endosomal limiting membrane of wild-type and vps1Δ cells were not significantly different (unpaired t test, P = 0.947; wild-type 57034 ± 6607 nm², n = 12; vps1Δ 57690 ± 6459 nm², n = 20).

Figure 9.

Mvp1 promotes Vps1-mediated fission of SNX-BAR–coated retrograde carriers from the endosome. A retrograde tubule budding from an endosome is depicted. The tubule contains integral membrane cargo, Vps10, and is coated by retromer, retromer SNX-BARs (Vps5–Vps17 heterodimer), and Mvp1. Based on the narrow diameter of membrane tubes produced in vitro by human SNX8 (van Weering et al., 2012a), the orthologue of yeast Mvp1, the coated tubule will be constricted at sites where it is wrapped by Mvp1. The constrictions are preferential sites for Vps1 membrane association and oligomerization, leading to fission of the tubule at these sites and the production of retrograde transport carriers of heterogeneous size.