Competitive organelle-specific adaptors recruit Vps13 to membrane contact sites

Abstract

The regulated expansion of membrane contact sites, which mediate the nonvesicular exchange of lipids between organelles, requires the recruitment of additional contact site proteins. Yeast Vps13 dynamically localizes to membrane contacts that connect the ER, mitochondria, endosomes, and vacuoles and is recruited to the prospore membrane in meiosis, but its targeting mechanism is unclear. In this study, we identify the sorting nexin Ypt35 as a novel adaptor that recruits Vps13 to endosomal and vacuolar membranes. We characterize an interaction motif in the Ypt35 N terminus and identify related motifs in the prospore membrane adaptor Spo71 and the mitochondrial membrane protein Mcp1. We find that Mcp1 is a mitochondrial adaptor for Vps13, and the Vps13-Mcp1 interaction, but not Ypt35, is required when ER-mitochondria contacts are lost. All three adaptors compete for binding to a conserved six-repeat region of Vps13 implicated in human disease. Our results support a competition-based model for regulating Vps13 localization at cellular membranes.

Figure 1.

The sorting nexin Ypt35 recruits Vps13 to endosomes. (A) GFP-tagged sorting nexins were immunopurified from lysates cross-linked with 400 µg/ml DSP, and coprecipitating Vps13-HA was detected. All proteins were expressed at endogenous levels. IP, immunoprecipitation. (B) Endogenous Vps13-GFP (top) and Ypt35-GFP (bottom) colocalize with Snf7-RFP–labeled endosomes but not Sec7-RFP labeled Golgi. (C) Quantitation of colocalization. Two-tailed equal-variance t test; n = 3, cells/strain/replicate ≥ 674; ***, P < 0.001; ****, P < 0.0001. (D) Overexpression (OE) of YPT35-RFP from the ADH1 promoter recruits Vps13-GFP to bright colocalizing puncta. DIC, differential interference contrast. (E) Quantitation of bright Vps13-GFP puncta in cells with endogenous or overexpressed YPT35. Two-tailed unequal variance t test; n = 3, cells/strain/replicate ≥ 355; ****, P < 0.0001. (F) Vps13-GFP requires YPT35 for its punctate localization. (G) Quantitation of Vps13-GFP and Ypt35-GFP puncta per cell. Two-tailed equal variance t test; n = 3, cells/strain/replicate ≥ 604; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Bars, 2 µm. Error bars indicate SEM.

Figure 2.

Vps13 and Ypt35 are interdependent for recruitment to the NVJ. (A) Growth in acetate-based media shifts Vps13^GFP and Ypt35-GFP to the NVJ at the interface of the vacuole (FM4-64; red) and nucleus (Hoechst; blue). White arrowheads indicate NVJs. (B) Localization of Vps13^GFP to the NVJ in acetate-based media is Ypt35 dependent. (C) Quantitation of Vps13^GFP-labeled NVJs. Two-tailed equal variance t test; n = 3, cells/strain/replicate ≥ 1,149; ****, P < 0.0001. DIC, differential interference contrast. (D) Ypt35 is dependent on Vps13 for localization to the NVJ, labeled in this figure with Nvj1-RFP, in acetate-based media. (E) Quantitation of the average Ypt35-GFP intensity at Nvj1-RFP–marked NVJs relative to the cell background. Two-tailed equal variance t test; n = 3, cells/strain/replicate ≥ 1,672; **, P < 0.01; ***, P < 0.001. Bars, 2 µm. Error bars indicate SEM.

Figure 3.

A motif in the N terminus of Ypt35 is necessary and sufficient for Vps13 recruitment. (A) An alignment of Ypt35 with fungal species identifies conserved residues in the N terminus. (B) Deletion of the N-terminal 48 residues of Ypt35 blocks Ypt35-RFP–induced recruitment of Vps13^GFP to puncta. (C) Schematic of the Ypt35(1–48)-RFP-FYVE construct used for Vps13^GFP localization experiments and mutant versions. (D) Ypt35(1–48)-RFP-FYVE recruits Vps13^GFP to puncta. DIC, differential interference contrast. (E) Vps13^GFP recruitment to puncta by Ypt35(1–48)-RFP-FYVE in a ypt35 strain was reduced by mutation of conserved N-terminal Ypt35 residues. (F) Quantitation of Vps13^GFPrecruitment to puncta in cells expressing WT and mutant forms of Ypt35(1–48)-RFP-FYVE. n = 3, cells/strain/replicate ≥ 913. (G) WebLogo (Crooks et al., 2004) of the Ypt35 PxP motif. Residues that when mutated to alanine caused a >40% loss of puncta (*) and >60% loss of puncta (**) are indicated. (H) Ypt35(P10,12A) is unable to recruit Vps13^GFP to puncta or NVJs in dextrose or acetate-based media, respectively. Bars, 2 µm. Error bars indicate SEM.

Figure 4.

A central region of Vps13 binds Ypt35 in a PxP motif–dependent manner. (A) Schematic of Vps13 truncations. (B) Coimmunoprecipitation of Ypt35-3HA with full-length and truncated forms of Vps13-GFP. Both proteins were overexpressed (OE) from the ADH1 promoter. (C) Densitometry of B; n = 3. (D) Coimmunoprecipitation of WT or mutant Ypt35-3HA with Vps13-GFP fragments indicates that interactions with DUF1162-containing regions are proline dependent. IP, immunoprecipitation; WCL, whole-cell lysate. (E) Densitometry of D; n ≥ 2. (F) Schematic of S. cerevisiae Vps13, highlighting an extended DUF1162 domain defined by six repeats identified by using HHrepID (Biegert and Söding, 2008). (G) An alignment of the most conserved section of the S. cerevisiae DUF1162 repeats. Error bars indicate SEM.

Figure 5.

The Vps13 DUF1162 domain recognizes the meiosis-specific adaptor Spo71 through a PxP motif. (A) Coimmunoprecipitation of TEF1pr-driven Spo71-HA with ADH1pr-driven Vps13-GFP fragments indicates that fragments containing the DUF1162 domain are sufficient for the interaction. (B) Densitometry of selected strains from A; n = 3. (C) A schematic of Spo71 highlighting a putative Vps13 interaction motif identified by a FIMO search by using a MEME motif generated from Ypt35 homologues. (D) Vps13^GFP puncta in ypt35 strains expressing the indicated WT or mutant Spo71(359–411)-RFP-FYVE constructs. (E) Quantitation of Vps13^GFP puncta per cell; n = 3, cells/strain/replicate ≥ 1,473. (F) ADH1pr-driven Vps13-GFP was immunopurified and coprecipitating WT or mutant TEF1pr-expressed Spo71-HA was detected. IP, immunoprecipitation; WCL, whole-cell lysate. (G) Densitometry of F; n = 3. Bars, 2 µm. Error bars indicate SEM.

Figure 6.

Mcp1 is a mitochondrial PxP motif–containing Vps13 adaptor that interacts with Vps13 to suppress ERMES mutants. (A) Schematic of Mcp1 highlighting a putative PxP interaction motif identified by FIMO by using a MEME motif generated from Ypt35 and Spo71 homologues. (B) Mutation of the motif (4–12Δ, P9,11A) blocked the ability of the Mcp1(1–58)-RFP-FYVE chimera to induce Vps13^GFP puncta in a ypt35 strain. (C) Quantitation of Vps13^GFP puncta per cell; n = 3, cells/strain/replicate ≥ 1,410. (D) The motif is required for coprecipitation of Vps13 and Mcp1. Vps13-GFP and WT and mutant forms of Mcp1-HA were overexpressed (OE) from the ADH1 promoter. IP, immunoprecipitation; WCL, whole-cell lysate. (E) Densitometry of D; n = 3. (F) The interaction between ADH1pr-driven overexpression of WT Mcp1 but not Mcp1 lacking the PxP motif recruits Vps13^GFP to mitochondria marked by preCox4-RFP (red). (G) Loss of the Mcp1 PxP motif is synthetic lethal with mdm10Δ. (H) Overexpression of WT but not mutant Mcp1suppresses the mitochondrial morphology defect of a strain lacking the ERMES subunit Mdm10. In contrast, loss of YPT35 does not rescue the mdm10Δ mitochondrial morphology defect with or without expression of VPS13 from two copies of the gene, nor does it block rescue by the dominant suppressor Vps13(D716H). DAPI was used as a mitochondrial marker. Bars, 2 µm. Error bars indicate SEM.

Figure 7.

Vps13 adaptors compete to recruit Vps13 to different membranes. (A) Suppression of the mdm10Δ mitochondrial morphology defect by moderately overexpressed Mcp1 (driven by the ADH1 promoter; ↑) is partially blocked by strong overexpression of Spo71 (driven by the TEF1 promoter; ↑↑↑) and completely blocked by strong overexpression of Ypt35 (driven by the GPD1 promoter; ↑↑↑). (B) High levels of Spo71-RFP or Ypt35-RFP block the Mcp1-dependent mitochondrial targeting of Vps13^GFP, which is instead relocalized to the plasma membrane or to endosomes, respectively. DAPI (blue) marks mitochondria. Arrowheads indicate Vps13^GFP colocalization with RFP (red), mitochondria (blue), or neither (white). (C) Vps13^GFP recruitment to puncta by YPT35 (ADH1pr; ↑) is reduced by highly expressed (TEF1pr; ↑↑↑) soluble RFP chimeras with the PxP motif–containing regions of Ypt35, Mcp1, or Spo71. (D) Quantitation of Ypt35-induced Vps13^GFP puncta in RFP-expressing cells shows a significant PxP-dependent reduction of puncta when the soluble chimeras are expressed. Unpaired one-way ANOVA: n = 5, ≥890 cells/strain/replicate; P < 0.0001 overall; Holm-Sidak's multiple comparisons test, ****, P < 0.0001. (E) Vps13^GFP recruitment to puncta by moderately overexpressed MCP1 (ADH1pr; ↑) is similarly reduced by strong overexpression (TEF1pr; ↑↑↑) of soluble RFP chimeras with the PxP motif–containing regions of Ypt35, Mcp1, or Spo71. (F) Quantitation of the Mcp1-induced Vps13^GFP puncta in RFP-expressing cells shows a significant PxP-dependent reduction of puncta when the soluble chimeras are expressed. Unpaired one-way ANOVA: n = 5, ≥719 cells/strain/replicate; P < 0.0001 overall; Holm-Sidak's multiple comparisons test, ****, P < 0.0001. Error bars indicate SEM. Bars, 2 µm.

Figure 8.

Model of Vps13 recruitment by adaptor proteins with PxP motifs. Ypt35, Spo71, and Mcp1 all contain PxP motifs that compete to recruit Vps13 to endosomes/vacuole/NVJ, the prospore membrane, and mitochondria respectively. Each adaptor shows PxP-dependent binding to a central repeat region of Vps13 that overlaps the DUF1162 domain; we propose this domain be named the VPS13 adaptor binding (VAB) domain.