Lysosomal integral membrane protein-2 (LIMP-2/SCARB2) is involved in lysosomal cholesterol export

Abstract

The intracellular transport of cholesterol is subject to tight regulation. The structure of the lysosomal integral membrane protein type 2 (LIMP-2, also known as SCARB2) reveals a large cavity that traverses the molecule and resembles the cavity in SR-B1 that mediates lipid transfer. The detection of cholesterol within the LIMP-2 structure and the formation of cholesterol^-like inclusions in LIMP-2 knockout mice suggested the possibility that LIMP2 transports cholesterol in lysosomes. We present results of molecular modeling, crosslinking studies, microscale thermophoresis and cell-based assays that support a role of LIMP-2 in cholesterol transport. We show that the cavity in the luminal domain of LIMP-2 can bind and deliver exogenous cholesterol to the lysosomal membrane and later to lipid droplets. Depletion of LIMP-2 alters SREBP-2-mediated cholesterol regulation, as well as LDL-receptor levels. Our data indicate that LIMP-2 operates in parallel with Niemann Pick (NPC)-proteins, mediating a slower mode of lysosomal cholesterol export.

Fig. 1

Physical and functional interaction between LIMP-2 and cholesterol. a Electron microscopy revealed characteristic membrane-surrounded cholesterol-like inclusion bodies within the cytoplasm of Schwann cells of the Nervus phrenicus of a 6.5 month old (upper panel) and N. ichiadicus of a 6.5 month and 23 month old (lower panel) LIMP-2-deficient mouse. Scale bars: (upper panel left, lower panel right) 250 nm (upper panel right) 500 nm, (lower panel left) 1000 nm. b Spatial distribution function of cholesterol (orange) in the LIMP-2 extracellular domain (tan) computed from MDS, shown as a cross-section through an axial slice (lighter orange) of the structure. c Trans-sterol can be UV-crosslinked to the recombinant ectodomain (luminal part of LIMP-2 (aa35− aa430); +) whereas without the lipid (negative control; −) no crosslinking was observed. The lower panel shows the LIMP-2 protein after Coomassie staining. d Wildtype (WT) and LIMP-2-deficient (KO) mouse embryonic fibroblasts (MEFs) were grown for 16 h in lipoprotein-deficient medium and subsequently challenged for 6 h with LDL. Intracellular cholesterol was visualized with filipin. Lysosomes were stained with LIMP-2 in WT MEF cells. Scale bar: 10 μm. e Quantification of cellular filipin intensity from D (mean ± SD; LIMP-2-WT n = 25, LIMP-2-KO n = 22 cells from one experiment; ****P < 0.0001, t-test). f Wide-field fluorescence micrographs of filipin-stained NPC1-deficient CHO M12 cells overexpressing LIMP-2-WT-mCherry. Arrowheads indicate examples of filipin-positive LIMP-2-containing LEs. Note the decrease of filipin fluorescence in LIMP-2-WT-mCherry-containing LEs. Scale bar, 10 μm. g Quantification of mean filipin intensity in wild type (WT.mCherry), tunnel-blocking mutant form of LIMP-2 (G379W/V415W.mCherrry) and non-expressing (LIMP-2 neg.) LEs (n = 75, 77, 152 organelles, respectively, from in 15 cells). Data (mean ± SEM) from two independent experiments, unpaired two-tailed Student’s t-test (****P < 0.0001). Source data are provided as a Source Data file

Fig. 2

Lipoprotein-derived cholesterol can translocate through the LIMP-2 ectodomain. a Confocal microscopy of HeLa cells transiently expressing C-terminally GFP-tagged wild-type (WT.GFP) or chimeric LIMP-2 (cmr.GFP). Cells were labeled with dextran to visualize late endosomal organelles. Cell boundaries are indicated by dotted lines. b Plasmalemmal expression of chimeric C-terminally GFP-tagged LIMP-2 (LIMP-2.cmr.GFP) in CHO cells was detected using Western blotting of biotinylated cell surface proteins after immunoprecipitation using streptavidin-beads. LIMP-2.GFP was used as a negative control. β-actin was used as a loading control for the input samples. c Confocal microscopy and quantification (d) of DiI-LDL binding to LIMP-2.cmr.GFP in CHO cells. Cells transiently expressing LIMP-2.cmr.GFP were incubated with DiI-LDL at pH 5 (upper panel) and pH 7 (lower panel). pH 5: n = 78, pH 7: n = 9 cells from two experiments. t-test (****P < 0.0001). e Confocal microscopy of CHO cells transiently expressing LIMP-2.cmr.mCherry and incubated with doubly labeled LDL, AF647-LDL(BC), after binding (upper panel) and washing with HBSS (lower panel). Note that AF647 was removed upon washing but BC was not. f Representation of the tunnel-blocking mutations of the human LIMP-2 luminal tunnel, shown as a slice through the body of the protein (sliced solid shown in orange). Residues A379 and V415, which point towards the cavity of the tunnel, are shown in blue. g, h Confocal microscopy (g) and quantification (h) of AF647-LDL(BC)-derived BC transport in CHO cells expressing LIMP-2.cmr.mCherry or LIMP-2.cmr.A379W/V415W.mCherrry proteins. Cells were pre-incubated with doubly labeled LDL (AF647-LDL(BC)) and washed with HBSS. The ratio of green (BC) over far-red (AF647-labeled LDL) fluorescence, measured in ZEN Lite (Zeiss), was used to estimate uptake (h); LIMP-2.cmr.mCherry n = 59, LIMP-2.cmr.A379W/V415W.mCherrry n = 50 cells from 3 experiments). i Quantification of BC uptake from SapA(BC) by CHO cells transiently expressing C-terminally mCherry-tagged-wild type (LIMP-2.cmr.mCherry) or tunnel-blocking mutant LIMP-2 chimera (LIMP-2.cmr.A379W/V415W.mCherrry). Also presented is the total fluorescence intensity of non-transfected cells. The total cellular fluorescence intensity of BC was measured using Volocity (LIMP-2.cmr.mCherry n = 67, LIMP-2.cmr.A379W/V415W.mCherrry n = 90, non-transfected cells n = 152 cells from 3 experiments). BC data in h, and i are the mean ± SEM of triplicate samples (****P < 0.0001, unpaired two-tailed Student’s t-test). a.u., arbitrary units. Scale bars, 10 μm. Source data are provided as a Source Data file

Fig. 3

LIMP-2-WT but not LIMP-2-G379W/V415W overexpression rescues cholesterol efflux from LEs in LIMP-2 deficient cells. a Principle of late endosomal BODIPY-cholesterol (BC) efflux assay in A431 cells. Cells are loaded for 24 h with oleic acid-conjugated BSA to generate lipid droplets (LD). In parallel, cells are labeled with dextran to visualize late endosomal organelles (LE). Cells are then labeled for 2 h with BODIPY cholesteryl linoleate-labeled LDL (BC LN-LDL) that enters LE, and chased in serum-free medium to monitor the transfer of BC from LE to LD (labeled with LipidTox during the last 30 min of BC-LN-LDL labeling). b Overlaid confocal images of BC (green), dextran/LIMP-2 (blue), and LipidTox (red) in live A431 cells transfected with the indicated siRNAs and/or cDNA constructs (LIMP-2-WT or LIMP-2-G379W/V415W overexpressing cells are outlined) at 0 h and 4 h of chase. Scale bar, 10 μm. c Quantification of the fraction of BC fluorescence residing in dextran-positive or LIMP-2-positive late endosomes (LE) vs. LipidTox-positive lipid droplets (LD) after 4 h of chase (n = 21 (siLIMP-2, siCtrl), 22 (siLIMP2 + WT.mCherry) or 23 (siNPC1, siLIMP-2 + G379W/V415W.mCherry) cells). Data (mean ± SEM) from 3 experiments, t-test (**P ≤ 0.01, ****P < 0.0001). d Wide-field fluorescence micrographs of filipin stained NPC1-deficient CHO M12 cells overexpressing LIMP-2-G379W/V415W-mCherry. Arrowheads indicate examples of filipin-positive LIMP-2-containing LEs. Scale bar, 10 μm. For quantification see 1 G. e Analysis of [³H]cholesterol incorporation into cholesteryl esters in LIMP-2-WT and LIMP-2-KO MEFs in the presence and absence of U18666A or PKF (n = 7 (0 h/5 h), n = 3 (9 h), n = 4 (PKF/U18666A) samples). Data (mean ± SEM) from two (9 h chase) to 3 experiments (5 h chase), unpaired two-tailed Student’s t-test (*P ≤ 0.05). Source data are provided as a Source Data file

Fig. 4

Regulation of cholesterol homeostasis is altered in LIMP-2 deficient cells. Analysis of LIMP-2 influence on transcriptional regulation of cholesterol homeostatic genes. a Termination of SREBP2 processing in response to abundant cholesterol in the ER (induced by LDL loading) is absent or incomplete in NPC1-silenced and LIMP-2-silenced HeLa cells, respectively. HeLa WT cells were treated with scrambled (ctrl), NPC1 or LIMP-2 siRNA, respectively, and incubated in either full medium, cholesterol-depleted medium or LDL-loaded medium. The processing of SREBP2 in response to high or low cholesterol levels was analyzed by immunoblot. b The quantification depicts the amount of nSREBP2 after LDL loading in cells treated with scrambled (ctrl), NPC1 or LIMP-2 siRNA, respectively (n = 3; data (mean ± SEM) from two experiments; *P ≤ 0.05, t-test). c Analysis of the effect of cholesterol depletion and re-addition of LDL cholesterol for 24 h in NPC1-deficient HeLa cells on nSREBP2 formation. Cells were either treated with control siRNA or with LIMP-2 specific siRNA. The processing of the SREBP2, as well as the expression of the LDL receptor (LDLR) and LIMP-2 was followed by immunoblot. Note that additional depletion of LIMP-2 led to incomplete termination of nSREBP2 fragment and LDLR formation after cholesterol loading for 24 h. d Quantification of nSREBP2 levels. Nuclear SREBP2 fragment as percentage of total SREBP2 is depicted. (n = 2; data (mean ± SD) from one experiment). e Quantification of LDLR levels after cholesterol loading, normalized to the cholesterol-depleted samples (n = 5; data (mean ± SEM) from two experiments; *P ≤ 0.05, **P ≤ 0.01, unpaired two-tailed Student’s t-test). f Model of the role of LIMP-2 in lysosomal cholesterol efflux. Source data are provided as a Source Data file