Readdressing the Localization of Apolipoprotein E (APOE) in Mitochondria-Associated Endoplasmic Reticulum (ER) Membranes (MAMs): An Investigation of the Hepatic Protein-Protein Interactions of APOE with the Mitochondrial Proteins Lon Protease (LONP1), Mitochondrial Import Receptor Subunit TOM40 (TOMM40) and Voltage-Dependent Anion-Selective Channel 1 (VDAC1)

Abstract

As a component of circulating lipoproteins, APOE binds to cell surface receptors mediating lipoprotein metabolism and cholesterol transport. A growing body of evidence, including the identification of a broad variety of cellular proteins interacting with APOE, suggests additional independent functions. Investigating cellular localization and protein-protein interactions in cultured human hepatocytes, we aimed to contribute to the elucidation of hitherto unnoted cellular functions of APOE. We observed a strong accumulation of APOE in MAMs, equally evident for the two major isoforms APOE3 and APOE4. Using mass spectrometry proteome analyses, novel and previously noted APOE interactors were identified, including the mitochondrial proteins TOMM40, LONP1 and VDAC1. All three interactors were present in MAM fractions, which we think initially facilitates interactions with APOE. LONP1 is a protease with chaperone activity, which migrated to MAMs in response to ER stress, displaying a reinforced interaction with APOE. We therefore hypothesize that APOE may help in the unfolded protein response (UPR) by acting as a co-chaperone in cooperation with LONP1 at the interface of mitochondria and ER membranes. The interaction of APOE with the integral proteins TOMM40 and VDAC1 may point to the formation of bridging complexes connecting mitochondria with other organelles.

Keywords: APOE4; chaperone; co-immunoprecipitation; liver; mitochondria–ER contacts; stress response; thapsigargin.

Figure 1

APOE accumulates in the MAMs of cultured hepatocytes, equally evident during APOE3 and APOE4 overexpression: (a) subcellular fractions were isolated from Huh7 cells by Percoll density gradient ultracentrifugation and analyzed by Western blotting. Representative images show the accumulation of the APOE protein in pure MAMs. Respective marker proteins were detected to visualize the purity of mitochondria (COX), MAMs (CANX) and cytosol (TUB); (b) APOE protein levels were quantified in the whole cell sample and crude and pure MAM fractions and normalized to the APOE level in the whole cell sample. In the pure MAM, the APOE protein level was significantly higher compared to the whole cell (p < 0.05, unpaired t-test) as indicated by an asterisk (*). The data are means ± SEM (n = 3); (c) no APOE isoform-dependent difference was observed in the accumulation of APOE in the pure MAM fraction in APOE3- and APOE4-transfected Huh7 cells. Data are means ± SEM (n = 2) and the accumulation of APOE in MAMs is related to the APOE protein level in the whole cell samples; subcellular fractions: c. mito., crude mitochondria; p. mito., pure mitochondria; c. MAM, crude mitochondria ER-membranes (MAMs), p. MAM, pure MAM.

Figure 2

The number of visualized MERCs in cultured hepatocytes and MAM-assembling protein levels in the mouse liver are similar in presence of APOE3 and APOE4: (a) representative images of MERC-PLA experiments in APOE-transfected Huh7 cells that were incubated with 1 or 5 g/L glucose for six hours. VDAC1-IP3R1 PLA signals are shown in green, APOE was additionally stained in red to identify successfully transfected cells, and cell nuclei appeared blue by staining with DAPI. 400× magnification; scale bar 5 µm; (b) PLA signals from at least 100 cells per sample from three independent experiments were quantified showing a significant reduction of PLA signals per cell in all samples in response to the glucose challenge. No difference was observed comparing APOE3-, APOE4- and mock-transfected cells. Significance was accepted at p < 0.05, indicated with an asterisk (*); a two-way ANOVA was performed followed by the Šídák’s multiple comparisons test; (c) MAM-assembling as well as ER and mitochondrial marker proteins were analyzed in the livers of APOE-targeted replacement mice by Western blotting and representative images are shown; (d) densitometric analysis revealed no significant differences between APOE3 and APOE4 mice (unpaired t-test). Target band intensity was normalized by total protein load and related to the mean of APOE3 mice. Data are means ± SEM (n = 5–6).

Figure 3

Mitochondrial APOE-interacting proteins are remotely connected to MAMs, with no apparent difference comparing APOE3- and APOE4-transfected cells. (a) Visualization of the eleven higher abundant protein–protein interactions (PPIs) shared by APOE3- and APOE4-transfected cells compared to unmodified Huh7 cells. PPIs were categorized by their main cellular localization identified according to The Human Gene Database GeneCards. Hitherto unknown PPIs (7/11) are highlighted by the greater thickness and blue color of the borderline. (b) Representative Western blot images of the presence of the APOE-interacting proteins LONP1, TOMM40 and VDAC1 in subcellular fractions isolated from Huh7 cells. Respective marker proteins for mitochondrial and ER-related MAM proteins were detected (GRP75, mitochondria, MAM; and PEMT, ER and MAM). (c) Representative Western blot images showing APOE, LONP1, TOMM40 and VDAC1 as well as the ER/MAM marker CANX in subcellular fractions of APOE3- and APOE4-transfected cells. (d) Target band intensities were quantified and normalized by total protein load per lane. The target protein level in the pure MAM fraction was related to the whole cell sample showing similar results in APOE3- and APOE4-transfected cells. Data are means ± SEM (n = 2). Subcellular fractions: c. mito., crude mitochondria; p. mito., pure mitochondria; c. MAM, crude mitochondria ER membranes (MAMs), p. MAM, pure MAM.

Figure 4

Presence in MAMs and extent of the PPI of selected candidates and APOE in ER-stressed cultured hepatocytes. (a) Western blot images of LONP1, TOMM40, VDAC1 and APOE detection in subcellular fractions of unmodified Huh7 cells. Thapsigargin treatment (50 µM, 24 h) provoked the accumulation of LONP1 and APOE in MAMs relative to the whole cell sample and compared to untreated control cells. GRP75 served as a marker for MAMs and the positive control for thapsigargin induced MAM protein translocation. c. MAM, crude mitochondria ER membranes (MAMs), p. MAM, pure MAM; (b) representative Western blot images of LONP1, TOMM40 and VADC1 in APOE co-IP samples from unmodified Huh7 cells treated with thapsigargin (50 µM, 24 h). No signals were visible in the IP negative control (no antibody used, no AB). S, IP supernatant, inp., input control; (c) target band intensity was normalized by the corresponding APOE band intensity. Relative target protein levels in thapsigargin-stressed cells were related to the mean of untreated cells (expressed as thapsigargin-induced change). Data are means from three individual cell culture experiments and co-IP (n = 3).