ORP1L mediated PI(4)P signaling at ER-lysosome-mitochondrion three-way contact contributes to mitochondrial division

Abstract

Mitochondrial division is not an autonomous event but involves multiple organelles, including the endoplasmic reticulum (ER) and lysosomes. Whereas the ER drives the constriction of mitochondrial membranes, the role of lysosomes in mitochondrial division is not known. Here, using super-resolution live-cell imaging, we investigate the recruitment of lysosomes to the site of mitochondrial division. We find that the ER recruits lysosomes to the site of division through the interaction of VAMP-associated proteins (VAPs) with the lysosomal lipid transfer protein ORP1L to induce a three-way contact between the ER, lysosome, and the mitochondrion. We also show that ORP1L might transport phosphatidylinositol-4-phosphate (PI(4)P) from lysosomes to mitochondria, as inhibiting its transfer or depleting PI(4)P at the mitochondrial division site impairs fission, demonstrating a direct role for PI(4)P in the division process. Our findings support a model where the ER recruits lysosomes to act in concert at the fission site for the efficient division of mitochondria.

Fig. 1. Lysosomes, mitochondria, and the ER are involved in a three-way contact that participates in a late step of mitochondrial division.

a Representative image and three-dimensional reconstruction of an ER-lysosome-mitochondrion three-way contact in a COS-7 cell. Scale bars: 5 and 1 µm (inset). b HeLa, COS-7, and U-2 OS expressing mito-BFP, GFP-KDEL, and Lamp1-mCherry. Insets are the magnification of the area indicated by the red box showing a lysosome-mitochondria contacts marked by the ER. White arrow indicates a lysosome-mitochondria contact. Scale bars: 10 µm and 1 µm (inset). c Quantification of the percentage of lysosome-mitochondria contacts that are marked by the ER in HeLa, COS-7, and U-2 OS cells. The graph shows the mean ± SEM. Cells from three independent experiments. d Live-cell imaging of HeLa cells expressing mito-BFP, GFP-KDEL, and Lamp1-mCherry over 10 s. Yellow arrows indicate an ER-mediated mitochondrial constriction site and red arrows indicate a mitochondrial fission event. Scale bar: 1 µm. e Percentage of mitochondrial division events (cells from three independent experiments) marked by the ER or lysosomes and the percentage of lysosomes marked division events that were also marked by the ER in HeLa cells. f Sequence of recruitment of the ER and lysosomes at mitochondrial division site for events that were positive for both the ER and lysosome in e. g–i Live-cell imaging of HeLa cells expressing GFP-Drp1, mito-BFP, and Lamp1-mCherry (g). Scale bar: 1 µm. h Percentage of mitochondrial division events marked by Drp1 and by lysosomes (cells from two independent experiments). i Sequence of recruitment of Drp1 and lysosomes at mitochondrial division site for events that were positive for both Drp1 and lysosome. j–l Live-cell imaging of HeLa cells expressing mito-BFP, GFP-Arf1, and Lamp1-mCherry (j). Scale bar: 1 µm. k Percentage of mitochondrial division events marked by Arf1 and by lysosomes (cells from two independent experiments). l Sequence of recruitment of Arf1 and lysosomes at mitochondrial division site for events that were positive for both Arf1 and lysosome. f, i, l Division sites that were already positive for both markers at the beginning of the acquisition were excluded from the analysis.

Fig. 2. Rab7 regulates a subset of ER-lysosome-mitochondria three-way contacts.

a Representative image of Rab7 positive lysosomes-mitochondria contacts that are marked by the ER in HeLa cells expressing mRFP-Rab7, GFP-KDEL, and mito-BFP. White arrows indicate the lysosome-mitochondria contacts. Line-scan analysis of relative fluorescence intensities (left to right) from the red line shown in merge. Scale bar: 1 µm. b Quantification of the percentage of lysosome-mitochondria contacts marked by the ER in HeLa cells expressing either a wild-type Rab7 or a control vector, with mito-BFP, GFP-KDEL, and Lamp1-mCherry. The graphs show the mean ± SEM, cells from three independent experiments. Two-sided unpaired t-test. c Normalized Rab7 mRNA levels in cells transfected by a siRNA targeting specifically Rab7 or a control siRNA measured by RT-qPCR. The graphs show the mean ± SEM, cells from two independent experiments. d Quantification of the percentage of lysosome-mitochondria contacts marked by the ER when Rab7 expression level was downregulated by a siRNA targeting Rab7 compared to a control siRNA in cells expressing Lamp1-mCherry, mito-BFP, and GFP-KDEL. The graphs show the mean ± SEM, cells from three independent experiments. Two-sided unpaired t-test. e Representative images of HeLa cells expressing the wild-type (WT), constitutively active (Q67L), or dominant negative (T22N) mRFP-Rab7. Note that the dominant negative Rab7 (T22N) does not localize to lysosomes but is cytosolic. Scale bar: 10 µm. f Quantification of the percentage of lysosome-mitochondria contacts that are marked by the ER in HeLa cells expressing either the 3HA-Rab7 or its mutants with Lamp1-mCherry, mito-BFP, and GFP-KDEL. The graph shows the mean ± SEM, Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test.

Fig. 3. The Rab7 interactor ORP1L regulates the formation of ER-lysosome-mitochondria three-way contacts in Hela cells.

a Representative image of a cell expressing GFP-ORP1L and mApple-TOM20 acquired by SIM microscopy. White arrows show ORP1L positive lysosomes-mitochondria contacts. Scale bars: 10 µm and 1 µm (inset). b Representative image of two ORP1L positive lysosome-mitochondria contacts, indicated by white arrows, marked by the ER. Line-scan analysis of relative fluorescence intensities from the yellow line in merge is shown. Scale bar: 1 µm. c Representative time lapse image of ORP1L positive lysosomes forming stable contact with mitochondria indicated by white arrows on the first frame. Scale bar: 1 µm. d, e Percentage of Lamp1-mCherry and mCherry-ORP1L positive lysosomes in contact with mitochondria (d) and e minimum duration of these contacts in cells co-expressing mito-GFP. Cells from three independent experiments. Two-sided unpaired t-test. f, g Percentage of Lamp1-mCherry positive lysosomes in contact with mitochondria (mito-BFP) (f) and g minimum duration of these contacts in cells expressing GFP, GFP-ORP1L, or the D478A mutant. Cells from two independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. h–j Percentage of lysosome-mitochondria contacts that were marked by the ER: h when ORP1 expression was downregulated. Cells from three independent experiments, one-way ANOVA with Dunnett’s Multiple Comparison Test; i in ORP1L KO cells and control cells. Cells from three independent experiments, two-sided unpaired t-test; and j when VAPA and VAPB expression levels were downregulated. Data from three independent experiments, Two-sided unpaired t-test. k Cytosolic GAI-ΔANKORP1L can be recruited to GID1-Rab7 lysosomes upon GA₃-AM treatment. l Representative image of a cell expressing mTq2-GAI-ΔANKORP1L and iRFP-GID1-Rab7 before and after GA₃-AM treatment (10 µM). Scale bar: 10 µm. m Normalized fluorescence intensity of mTq2-GAI-ΔANKORP1L WT and D478A colocalizing with iRFP-GID1-Rab7 lysosomes at the indicated times before and after GA₃-AM treatment. Cells from four independent experiments. n Relative ER fluorescence intensity at lysosome-mitochondria contacts at the indicated times before and after GA₃-AM treatment (shaded area represent the area within one SEM). Cells from four independent experiments. Two-way ANOVA, Sidak’s multiple comparisons test. d–j, m, n All graphs show the mean ± SEM.

Fig. 4. The Rab7-ORP1L-VAPs interaction regulates mitochondrial division.

a, b Normalized rate of mitochondrial division (number of mitochondrial divisions normalized by time and volume) in HeLa cells (a) expressing wild-type or mutants 3HA-Rab7. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test; b when VAPs expression levels were downregulated using siRNAs. Cells from three independent experiments. Two-sided unpaired t-test. The graphs show the mean ± SEM. c Live-cell imaging of HeLa cells expressing GFP-ORP1L and mApple-TOM20. The white and yellow arrows show an ORP1L positive lysosome-mitochondria contact and a mitochondrial fission event. Scale bar: 1 µm. d Percentage of mitochondrial division events that were marked by GFP-ORP1L (cells from two independent experiments) in HeLa cells. e Representative maximum projection images of mitochondrial morphology in HeLa cells treated with indicated siRNAs. Scale bars: 10 µm and 5 µm (inset). f–h Mitochondrial morphology was quantified for f mean area per mitochondrion, g mitochondrial number per region of interest (ROI), and h number of junctions per mitochondria. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. The graphs show the mean ± SEM. i, j Normalized rate of mitochondrial division of i HeLa cells treated with indicated siRNAs. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test and of j ORP1L WT and KO HeLa cells. Cells from three independent experiments. Two-sided unpaired t-test. The graphs show the mean ± SEM. k ORP1L mutants used and the loss of function caused by the mutations or deletions. l Normalized rate of mitochondrial division in ORP1L KO HeLa cells expressing GFP or GFP-ORP1L constructs and mApple-TOM20. The graphs show the mean ± SEM, cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. m Normalized rate of mitochondrial division in ORP1L KO HeLa cells expressing mTq2-GAI-ΔANKORP1L, iRFP-GID1-Rab7, and mApple-TOM20 before and after GA₃-AM (10 µM) treatment. Cells were imaged from 5 to 20 min after GA₃-AM treatment. Cells from three independent experiments. Two-sided unpaired t-test. The graphs show the mean ± SEM.

Fig. 5. Lysosomal PI(4)P contributes to mitochondrial division.

a Representative image of a mitochondrial division marked by a GFP-2xP4M positive lysosome (yellow arrow) in HeLa cells expressing the indicated markers. PI4KA inhibitor GSK-A1 treatment allowed to better visualize PI(4)P at lysosomes. Scale bar: 1 µm. b Percentage of mitochondrial division events marked by lysosomes (cells from two independent experiments) and of lysosomes marked by GFP-2xP4M at mitochondrial division events. c In ORPSAC1, the ORD domain of ORP1L was replaced by the catalytic domain of Sac1 allowing the dephosphorylation of lysosomal PI(4)P. d Representative maximum projection images of mitochondrial morphology in HeLa cells overexpressing the indicated constructs. Scale bars: 10 µm and 5 µm (inset). e–g Mitochondrial morphology was quantified for e mean area per mitochondrion, f mitochondrial number per region of interest (ROI), and g number of junctions per mitochondria. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. h Normalized rate of mitochondrial division in HeLa cells overexpressing the indicated constructs. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. i Cytosolic GAI-Sac1 can be recruited to GID1-Rab7 lysosomes upon GA₃-AM treatment to specifically deplete lysosomal PI(4)P. j, k Representative images of HeLa cells expressing CFP-GAI-Sac1 or the catalytic dead C392S mutant, iRFP-GID1-Rab7, and mCherry-2xP4M after GA₃-AM treatment (j). Scale bars: 10 µm and 1 µm (inset). k Quantification of the lysosomal levels, normalized by the plasmalemmal levels, of 2xP4M. Cells from three independent experiments. Two-sided unpaired t-test. l Normalized rate of mitochondrial division before or after GA₃-AM treatment in HeLa cells overexpressing iRFP-GID1-Rab7 and the indicated constructs. Cells from three independent experiments. Two-way ANOVA, Sidak’s multiple comparisons test. m Representative image of a HeLa cell expressing GFP-PI4K2B and Lamp1-mCherry with zoomed insert of white box in merge panel. Scale bar: 10 µm. n Normalized rate of mitochondrial division in HeLa cells treated with the indicated siRNAs. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test. e–h, k, l, n All graphs show the mean ± SEM.

Fig. 6. Targeting the PI(4)P transfer protein OSBP to lysosomes rescues the mitochondrial morphology alterations caused by ORP1L depletion.

a ORPOSBP is a chimeric protein made of the amino acids 1 to 335 of ORP1L fused to amino acids 184 to 809 of OSBP from Oryctalogus cuniculus. Domains are not to scale. b Representative image of a Hela cell expressing GFP-ORPOSBP and mRFP-Rab7. Zoomed images of dashed box shown on right. Scale bars: 10 µm and 5 µm (inset). c Time-lapse images of HeLa cells expressing GFP-ORPOSBP and mApple-TOM20. The yellow arrow shows a mitochondrial fission event marked by GFP-ORPOSBP. Scale bar: 1 µm. d, e Representative images of HeLa cells expressing the wild-type GFP-ORPOSBP or the lipid binding deficient HH/AA mutant with the PI(4)P probe mCherry-2xP4M (d). Scale bars: 10 µm and 1 µm (inset). e Quantification of the lysosomal levels of 2xP4M normalized by the plasmalemmal levels of the probe. The graphs show the mean ± SEM, cells from three independent experiments. Two-sided unpaired t-test. f Representative maximum projection images of mitochondrial morphology in HeLa cells treated with siRNA downregulating ORP1L and overexpressing GFP, GFP-ORPOSBP or the HH/AA ORPOSBP mutant and mApple-TOM20. Note that ORPOSBP wild-type expression, but not of the HH/AA mutant rescue the mitochondrial elongation and hyperfusion of the mitochondrial network caused by ORP1L depletion. Scale bars: 10 µm and 5 µm (inset). g–i Mitochondrial morphology was quantified for g mean area per mitochondrion, h mitochondrial number per region of interest (ROI), and i number of junctions per mitochondria. Cells from three independent experiments. One-way ANOVA with Dunnett’s Multiple Comparison Test, ns non statistically significant: g p = 0.6437, h p = 0.7205, and i p = 0.3046. The graphs show the mean ± SEM. j–l Mitochondrial morphology of Hela cells expressing GFP or GFP-ORPOSBP was quantified for j mean area per mitochondrion, k mitochondrial number per region of interest (ROI), and l number of junctions per mitochondria. Cells from three independent experiments. Two-sided unpaired t-test, ns non statistically significant. j p = 0.8011, k p = 0.6381 and l p = 0.9266. The graphs show the mean ± SEM.

Fig. 7. The recruitment of the PI(4)P phosphatase Sac1 to mitochondria impairs their division.

a Soluble GAI-Sac1 can be recruited to mitochondrial fission site using GID1-MFF upon GA₃-AM treatment, leading to dephosphorylation of PI(4)P at the mitochondrial division site. b Representative image of a HeLa cell expressing CFP-GAI-Sac1, iRFP-GID1-MFF (not imaged) and mApple-TOM20 before and after GA₃-AM treatment (10 µM). Scale bar: 10 µm. c Normalized rate of mitochondrial division before and after GA₃-AM (10 µM) treatment in HeLa cells overexpressing the wild-type CFP-GAI-Sac1 or the inactive C392S mutant, iRFP-GID1-MFF and mApple-TOM20. When treated with GA₃-AM (10 µM) cells were imaged between 5 and 25 min of treatment. Cells from three independent experiments. Two-way ANOVA, Sidak’s multiple comparisons test. d In GFP-Sac1-MFF, Sac1 was fused to MFF to anchor it directly to the outer mitochondrial membrane leading to dephosphorylation of PI(4)P at the mitochondrial division site. e Representative maximum projection images of HeLa cells expressing GFP, GFP-MFF, GFP-Sac1-MFF or the catalytic inactive GFP-Sac1 C292S-MFF and mApple-TOM20. Inset shows the morphology of the mitochondrial network. Scale bars: 10 µm and 5 µm (inset). f–h Mitochondrial morphology was quantified for f mean area per mitochondrion, g mitochondrial number per region of interest (ROI), and h number of junctions per mitochondria. Cells from three independent experiments. One-way ANOVA with Tukey’s Multiple Comparison Test. i Representative images of ORP1L KO HeLa cells expressing the indicated constructs before and after GA₃-AM (10 µM) treatment. Scale bar: 10 µm. j Normalized rate of mitochondrial division in ORP1L KO HeLa cells expressing the cytosolic mTq²-GAI-ΔANKORP1L D478A, mApple-TOM20, GFP-ORP1LΔORD, and iRFP-GID1-Rab7 before and after GA₃-AM (10 µM) treatment. Cells from three independent experiments. Two-sided unpaired t-test. k Normalized rate of mitochondrial division in ORP1L KO HeLa cells expressing the cytosolic mTq²-GAI-ΔANKORP1L D478A, mApple-TOM20 and iRFP-GID1-Rab7 before and after GA₃-AM (10 µM) treatment. Cells from three independent experiments. Two-sided unpaired t-test. c, f–h, j, k All graphs show the mean ± SEM. ns non-significant: f p = 0.7632, g p = 0.3592, h p = 0.7935, and k p = 0.5192.

Fig. 8. Proposed model for ORP1L functions at the mitochondrial division site.

Mitochondrial division initiates at contacts with the ER, where the ER drives the constriction of mitochondrial membranes. This implicates an actin machinery that involves the ER protein INF2 and the mitochondrial Spire1C. The constriction allows the recruitment of Drp1 by adapters such as MFF and its oligomerization that further constrict mitochondrial membranes (non-represented in the cartoon). Lysosomes are then recruited to the division site in a process mediated by the Rab7-ORP1L-VAPs interaction that establishs contact sites with the ER (1). This allows the formation of a three-way contact between the ER, the lysosome and mitochondria at the division site that brings the lysosome in contact with mitochondria. Lysosome-mitochondria tethers are still unknown. At the Lysosome-mitochondria contact (2), we propose that ORP1L mediates the transfer of PI(4)P from lysosome to the division site. This model is supported by the impaired mitochondrial division when this transfer is inhibited or when PI(4)P is depleted at the mitochondrial division site.