Apolipoproteins L1 and L3 control mitochondrial membrane dynamics

Abstract

Apolipoproteins L1 and L3 (APOLs) are associated at the Golgi with the membrane fission factors phosphatidylinositol 4-kinase-IIIB (PI4KB) and non-muscular myosin 2A. Either APOL1 C-terminal truncation (APOL1Δ) or APOL3 deletion (APOL3-KO [knockout]) reduces PI4KB activity and triggers actomyosin reorganization. We report that APOL3, but not APOL1, controls PI4KB activity through interaction with PI4KB and neuronal calcium sensor-1 or calneuron-1. Both APOLs are present in Golgi-derived autophagy-related protein 9A vesicles, which are involved in PI4KB trafficking. Like APOL3-KO, APOL1Δ induces PI4KB dissociation from APOL3, linked to reduction of mitophagy flux and production of mitochondrial reactive oxygen species. APOL1 and APOL3, respectively, can interact with the mitophagy receptor prohibitin-2 and the mitophagosome membrane fusion factor vesicle-associated membrane protein-8 (VAMP8). While APOL1 conditions PI4KB and APOL3 involvement in mitochondrion fission and mitophagy, APOL3-VAMP8 interaction promotes fusion between mitophagosomal and endolysosomal membranes. We propose that APOL3 controls mitochondrial membrane dynamics through interactions with the fission factor PI4KB and the fusion factor VAMP8.

Keywords: APOL1 risk variants; COVAN; COVID-19-associated nephropathy; CP: Cell biology; HIV-associated nephropathy; HIVAN; inflammation; interferon 1; kidney disease; mitochondrion fission/fusion; mitophagy.

Graphical abstract

Figure 1

Interactions between PI4KB, APOLs, NCS1, and CALN1 (A) BLI measurements of immobilized PI4KB interaction with various recombinant APOLs, with or without calcium as indicated (n = 3). (B) BLI measurements of immobilized PI4KB interaction with recombinant NCS1 with or without calcium (n = 3). (C) BLI measurements of immobilized PI4KB sequential interaction with recombinant APOL3 and NCS1 as indicated (n = 3). (D) BLI measurements of immobilized NCS1 (left) or CALN1 (right) interaction with recombinant APOL1 or APOL3 with or without calcium (n = 3). (E) PI(4)P synthesis by recombinant PI4KB (50 ng) with or without addition of different recombinant proteins, expressed as percentage of activity without addition. The activity without PI substrate or without PI4KB (<0.1%) was subtracted (n = 3). PI4KB was mixed with a 3-fold stoichiometric amount of CALN1 as shown on the right. Error bars represent the standard deviation.

Figure 2

Cis- and trans-interactions of APOL1, APOL3, PI4KB, and ARF1 (A) Representative SPR sensorgrams of immobilized PI4KB H1-H3 interactions with recombinant APOL1, APOL1Δ, APOL3, NCS1, or NCS1 L89K (n = 3). As in (B)–(I), interaction was performed in 50 μM CaCl₂. (B) SPR sensorgrams of immobilized APOL1 CTD interaction with the APOL1 NTD peptide, measured at different pHs (n = 3). (C) BLI measurements of immobilized PI4KB sequential interactions with recombinant APOL1Δ or APOL3 as indicated (n = 3). (D) Representative SPR sensorgrams of immobilized ARF1p interaction with PI4KB H1-H3 or APOL1/3 HC1 peptides (n = 3). (E) Representative SPR sensorgrams of immobilized ARF1p interaction with recombinant APOL3, APOL1, APOL3Δ, or APOL1Δ (n = 3). Notice the different binding scales between the two graphs. (F) BLI measurement of immobilized recombinant ARF1 interaction with recombinant APOL3 (n = 3). (G) Representative SPR sensorgrams of immobilized ARF1p sequential interactions with recombinant APOL3 or PI4KB as indicated (n = 3). (H) BLI measurements of immobilized PI4KB sequential interactions with recombinant APOL3 or ARF1 as indicated (n = 3). (I) Immunofluorescence analysis of APOL3-V5 co-localization with PI4KB in WT podocytes treated or not with poly(I:C). Bar: 20 μm; blue: nucleus (DAPI). n = 3; Mann-Whitney test, ^∗p < 0.05. Error bars represent the standard deviation.

Figure 3

Cellular changes linked to APOL1/-3 gene editing (A) Immunofluorescence analysis of APOL3-V5 and APOL1 co-localization with ATG9A vesicles in various podocyte cell lines. The arrowheads point to coalesced vesicles. Bar: 20 μm; blue: nucleus (DAPI). n = 3; Mann-Whitney test, ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001. Error bars represent the standard deviation. (B) Co-localization of APOL3-V5 and APOL1 with ATG9A in coalesced ATG9A vesicles of WT podocytes incubated with poly(I:C). Arrowhead: co-localization of ATG9A with APOL1 occurring only in non-coalesced vesicle. Bar: 10 μm. (C) Immunodetection of the MERCS marker PANX2 in different podocyte cell lines and co-localization of PANX2 with MFN2. Bar: 20 μm; blue: nucleus (DAPI). n = 3; Mann-Whitney test, ^∗∗p < 0.01; ^∗∗∗∗p < 0.0001. Error bars represent the standard deviation. (D) Co-localization of PI4KB with FIS1 in various podocyte cell lines. Bar: 10 μm; blue: nucleus (DAPI). n = 3; Mann-Whitney test, ^∗∗∗∗p < 0.0001. Error bars represent the standard deviation.

Figure 4

Mitochondrial structures in WT and APOL1Δ podocytes (A) Representative images of the cytoplasm of WT podocytes in independent sections of the same cell volume imaged by FIB-SEM. N, nucleus, CM, cytoplasmic membrane; L, lysosome; G, Golgi; M, normal mitochondrion. Bars: 1 μm. (B) Same as (A) but in APOL1Δ podocytes. Arrows point to mitophagosomes at several stages of maturation. Notice the decrease in cytoplasmic membrane complexity due to actomyosin cytoskeleton changes. Bars: 2 μm. (C) Zoom in on some of the mitophagosomes in (B). Bars: 250 nm. (D) Model of mitophagosome formation in the different podocyte cell lines. In APOL1Δ cells, the delay in mitophagosome resolution leads to its increase in size and complexity, with the emergence of multilamellar bodies attached to the mitochondrion.

Figure 5

Effects of APOL1/-3 gene editing on the morphology and activity of mitochondria (A) FIB-SEM images of reconstructed 3D cell segments in various podocyte cell lines. Blue, mitochondria; orange, mitophagosomes; light gray, unresolved fission structures; dark gray, cytoplasm or cytoplasmic membrane. Videos are provided at https://owncloud.ulb.ac.be/index.php/s/pkny92qwROGq1df. (B) Number of mitophagosomes per 10 μm³ for the segments shown in (A). (C) Volumes of mitochondria and mitophagosomes for the segments shown in (A), expressed as percentages of cell volumes. (D) Number of structures showing unresolved mitochondrial fission for the segments shown in (A). (E) Live measurement of mitochondrial ROS production in various podocyte cell lines, as determined by co-localization of MitoSox and MitoTracker staining. Bars: 20 μm. n = 3; Mann-Whitney test, ^∗p < 0.05; ^∗∗∗p < 0.001; ^∗∗∗∗p < 0.0001. Error bars represent the standard deviation.

Figure 6

APOL1- and APOL3-specific interactions (A) Interaction of APOL1, APOL1 mutants, or APOL3 with PHB2, as measured in the E. coli pDUET co-expression system. Nickel binding of S-tagged proteins (yellow frame) reflects their relative association with His-tagged partners (green frame). These data are representative of at least 3 independent experiments. A1, WT APOL1; Δ, APOL1Δ; A3, WT APOL3; H4Q, APOL1 mutHel4Q; H5Q, APOL1 mutHel5Q; H5P, APOL1 mutHel5P; H5mS, APOL1 mutHel5S; H5sw, APOL1 Hel5 swap; BH3E, APOL1 BH3 mutant E. (B) Representative SPR sensorgrams of immobilized L1Hel5 (left) or L3Hel5 (right) peptide interaction with recombinant PHB2 with or without calcium (n = 3). (C) Representative SPR sensorgrams of immobilized L1Hel5 (left) or L3Hel5 (right) peptide interaction with APOL3 Hel5, NTD, or CTD peptide (n = 3). (D) SPR measurements of bound sVAMP8 (left) or TbVAMP7Bp (right) interaction with various APOL3 versions as indicated (n = 3). (E) SPR measurements of bound sVAMP8 (left) or TbVAMP7Bp (right) interaction with different APOL3 peptides as indicated (n = 3). (F) Immunodetection of recombinant sVAMP8 binding to various lipids spotted on membrane strips. LPA, lysophosphatidic acid; LPC, lysophosphocholine; PI, phosphatidylinositol; PE, phosphatidylethanolamine; PC, phosphatidylcholine; S1P, sphingosine-1-phosphate; PA, phosphatidic acid; PS, phosphatidylserine; TG, triglyceride; DAG, diacylglycerol; PG, phosphatidylglycerol; CL, cardiolipin; Chol, cholesterol; SM, sphingomyelin. Blank, no lipid.

Figure 7

Vesicular membrane fusion induced by APOL3 and VAMP8 (A) Scheme of the lipid mixing assay. The fluorescence of the donor dye (NBD, dark green star) quenched due to the proximity to the acceptor dye (rhodamine, pink star) is increased upon membrane fusion with non-fluorescent partner SUVs. (B) Protein and lipid composition of small unilamellar vesicles (SUVs) involved in the fusion assays. P:L, protein:lipid ratio; N-PE, NBD-PE; R-PE, rhodamine-PE (see STAR Methods for lipid abbreviations). In red, alternative lipid composition (“-CL” condition). (C) Fusion assays between APOL3 or APOL1 SUVs and VAMP8 N/R (N-PE/R-PE) SUVs (n = number of replicates; SD = standard deviation at t = 90). (D) Same as in (C) but including SUVs with APOL3 mutants. (E) Same as in (C) but including SUVs with STX1A + SNAP25.