The splice variant LOXIN inhibits LOX-1 receptor function through hetero-oligomerization

The splice variant LOXIN inhibits LOX-1 receptor function through hetero-oligomerization

Available online at www.sciencedirect.com Journal of Molecular and Cellular Cardiology 44 (2008) 561 – 570 www.elsevier.com/locate/yjmcc Original ar...

931KB Sizes 0 Downloads 35 Views

Available online at www.sciencedirect.com

Journal of Molecular and Cellular Cardiology 44 (2008) 561 – 570 www.elsevier.com/locate/yjmcc

Original article

The splice variant LOXIN inhibits LOX-1 receptor function through hetero-oligomerization Silvia Biocca a,⁎, Ilaria Filesi a , Ruggiero Mango b , Luana Maggiore a , Francesco Baldini a , Lucia Vecchione c , Antonella Viola c , Gennaro Citro d , Giorgio Federici b , Francesco Romeo b,e , Giuseppe Novelli c,e a

c

Department of Neuroscience and Laboratory of Clinical Biochemistry, University of Tor Vergata, Rome, Italy b Department of Internal Medicine, University of Tor Vergata, Rome, Italy Department of Biopathology and Diagnostic Imaging, Centre of Excellence for Genomic Risk Assessment in Multifactorial and Complex Diseases, School of Medicine, University of Tor Vergata, Rome, Italy d S.A.F.U. Department, Regina Elena Cancer Institute, Rome, Italy e Departments of Internal Medicine and Physiology and Biophysics, University of Arkansas for Medical Sciences, Little Rock, AR, USA Received 27 November 2007; accepted 28 November 2007 Available online 14 January 2008

Abstract Lectin-like oxidized low-density lipoprotein receptor-1 (LOX-1), encoded by the OLR1 gene, is a scavenger receptor that plays a central role in the pathogenesis of atherosclerosis. We have recently identified a truncated naturally occurring variant of the human receptor LOX-1, named LOXIN, which lacks part of the C-terminus lectin-like domain. In vivo and in vitro studies support that the new splicing isoform is protective against acute myocardial infarction. The mechanism by which LOXIN exerts its protective role is unknown. In this paper we report studies on the heterologous expression and functional characterization of LOXIN variant in mammalian fibroblasts and human endothelial cells. We found that LOXIN, when expressed in the absence of LOX-1, shows diminished plasma membrane localization and is deficient in ox-LDL ligand binding. When co-transfected with the full-length counterpart LOX-1, the two isoforms interact to form LOX-1 oligomers and their interaction leads to a decrease in the appearance of LOX-1 receptors in the plasma membrane and a marked impairment of ox-LDL binding and uptake. Coimmunoprecipitation studies confirmed the molecular LOX-1/LOXIN interaction and the formation of non-functional hetero-oligomers. Our studies suggest that hetero-oligomerization between naturally occurring isoforms of LOX-1 may represent a general paradigm for regulation of LOX-1 function by its variants. © 2007 Elsevier Inc. All rights reserved. Keywords: Lectin-like oxidized LDL receptor (LOX-1); LOXIN; oxidized LDL; oligomerization

1. Introduction The interaction of oxidatively modified low-density lipoprotein (ox-LDL) with macrophages, smooth muscular cells, lymphocytes and various components of artery walls, including endothelial cells, plays a central role in atherogenesis [1,2]. OxLDL elicits endothelial dysfunction, a key step in the initiation of atherosclerosis, favouring generation of reactive oxygen species, inhibition of nitric oxide synthesis, and enhancement of

⁎ Corresponding author. Tel.: +39 06 72596428; fax: +39 06 72596407. E-mail address: [email protected] (S. Biocca). 0022-2828/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.yjmcc.2007.11.017

monocyte adhesion to activated endothelial cells [3]. In addition, ox-LDL is involved in inducing smooth muscle cell migration and proliferation, and is avidly ingested by macrophages, resulting in foam cells formation [4]. Increased levels of ox-LDL relate to plaque instability in human coronary atherosclerotic lesions and to the severity of acute coronary syndromes such as myocardial infarction (MI) and unstable angina [5]. According to this scenario, ox-LDL induces apoptosis and necrosis of vascular endothelial, smooth muscle cells (SMCs), and macrophages [4]. Most of these effects are mediated by the interaction of ox-LDL with its major receptor lectin-like oxidized low density lipoprotein receptor-1 (LOX-1), characterized as the primary receptor for ox-LDL in endothelial

562

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

cells and up-regulated in atherosclerotic lesions [4,6]. LOX-1 belongs to the C-type lectin family of scavenger receptors and consists of four domains: a short N-terminal cytoplasmic, a single transmembrane, a connecting neck, and a lectin-like domain at the C-terminus. The C-terminal lectin-like domain is highly conserved among species, especially at the position of the six cysteine residues. This is consistent with its function as a ligand-binding domain and an initiator of internalization and phagocytosis [7,8]. By site-directed mutagenesis and crystal structure analysis, it was found that LOX-1 acts as a heartshaped disulfide-linked homodimer and that dimerization is indispensable for its ligand-binding activity [9–11]. LOX-1 is expressed in endothelial cells, macrophages, SMCs, and platelets. Furthermore, LOX-1 is present in atheroma-derived cells and is overexpressed in humans and animal atherosclerotic lesions in vivo. In addition, LOX-1 can activate different intracellular signaling pathways leading to the production of pro-inflammatory cytokines, matrix metalloproteinases, and transcription factors playing a fundamental role in the entire process of atherosclerosis [4,12]. LOX-1 is encoded by a single gene, OLR1, mapped on human chromosome 12p12.3–13.1. We and others have identified polymorphisms within the OLR1 gene which are associated with coronary atherosclerosis (CAD) and acute myocardial infarction (AMI) [13–15]. In particular, we characterized six single nucleotide polymorphisms (SNPs), embedded within a disequilibrium block (LD1) spanning from intron 4 up to the 3'UTR of the gene. All polymorphisms comprised within LD1 show a strong positive association with acute myocardial infarction. LD1 SNPs of OLR1 gene regulate the expression of a new functional splicing isoform of the OLR1 gene, named LOXIN, which lacks exon 5. In vivo and in vitro studies suggest a negative link between levels of LOXIN mRNA and protein expression and the incidence of myocardial infarction in humans [15]. The proapoptotic effect of LOX-1 receptor in cell culture is specifically rescued by the coexpression of LOXIN in a dose-dependent manner [15]. Programmed cell death is increasingly observed as atherosclerotic lesions progress, although the exact mechanism and consequences of apoptosis in the development of atherosclerosis are still unclear. LOXIN protein may balance against this very morbid outcome. Here we report the expression and functional characterization of LOXIN variant. This variant was deficient in ligand binding, but interacted with the full-length LOX-1 receptors and blocked their surface appearance, ox-LDL binding activity and its uptake. 2. Materials and methods

TAGCGGCCGCCACTGTGCTCTTAGGTTTGC-3′) and LOXIN-NotI (5′-ATAGTTTAGCGGCCGCAAAATCAGATCAGCTGTGCT-3′). For the expression in mammalian cells, the same amplified EcoRI and NotI human LOX-1 and LOXIN cDNAs fragments were subcloned into pEF/myc-His and pEF/ V5-His vectors (Invitrogen). To generate the human non-tagged LOX-1 construct, LOX-1 cDNA was PCR amplified from pEF/ V5-LOX-1 and cloned into the BamH1/EcoR1 digested pCDNA 3.1(+) vector (Invitrogen). For PCR amplification, the following primers were used: LOX-1 BamH1 (5′-CGGATCCATGACTTTTGATGACCTAAAGA-3′) and LOX-1 EcoR1 (5′CGAATTCTCACTGTGCTCTTAGGTTTGC- 3′). To generate the non-tagged LOXIN construct, the coding sequence of LOXIN was PCR-amplified from cDNA derived from the human heart poly A+ RNA (Clontech), using the following oligos: LOXIN-FW (5′-TGTTGAAGTTCGTGACTGCTT-3′) and LOXIN-RW (5′-TTCTGCAGCCAGCTAAATGA-3′). The PCR products were cloned into pCR2.1 vector using TA cloning Kit (Invitrogen) and the BamH1/Xho1 fragment, containing LOXIN coding region, was subcloned into the pCDNA 3.1(+) vector. pEF/V5-R4 was generated by subcloning the PCR-amplified BamHI/NotI fragment derived from pscFvexpress-cytR4 [16] into the pEF/V5-His vector (Invitrogen). For PCR amplification, the following primers were used: (5'-CGGGATCCCACGTGGCCACCATGG-3') and (5'-GCGGCCGCCCGTTTGATCTCG3'). Generation of pscFvexpress-Sec-8H4 has been previously described [17]. The PCR products-derived constructs were confirmed by sequencing using ABI PRISM® 3130 XL Genetic Analyzer (Applied Biosystems). 2.2. Antibody preparation Purified recombinant LOX-1 was fused to glutathione-Stransferase (GST) to form the LOX-1–GST fusion protein and used to generate polyclonal antibodies in rats. Isolation of LOX1–GST from isopropyl-β-D-1-thiogalactopyranoside (IPTG) induced bacteria was performed as previously described [18]. The protein was purified to homogeneity using SDS–PAGE, the induced LOX-1-GST gel-band was cut and eluted by immersion in bidistillated water (H2O) overnight at 4 °C. Two Sprague– Dawley male rats, weighing 150–180 g, were subcutaneously injected with 25 μg of antigen (dissolved in 50 μl of sterile 100 mM phosphate buffer saline pH 7.2) adsorbed onto 20 μl of aluminum hydroxide (Pierce). The mixture was injected weekly for three consecutive weeks. The booster dose was administered 15 days after the last inoculation. Four days after the booster injection the antiserum was collected from the animals by tail bleed and tested to determine the strength of the specific antibody response.

2.1. DNA constructs 2.3. Antibodies and reagents For bacterial expression, LOX-1 cDNA was PCR amplified and the EcoRI–NotI fragment cloned into the pGEX-4T1 vector (Clontech). For PCR amplification of LOX-1, the following primers were used: LOX-1 EcoRI (5′-CCGGAATTCATGACTTTTGATGACCTAAAG-3′), LOX-1-NotI (5′-ATAGTT-

Mouse anti-myc IgG 9E10 and mouse anti-V5 antibodies were purchased from Invitrogen, rabbit anti-myc and goat anticalnexin from Santa Cruz, mouse anti-Golgi 58k protein and rabbit anti-V5 from Sigma. Secondary antibodies goat anti-

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

mouse IgG horseradish peroxidase (HRP) was from Jackson Immunoresearch and goat anti-rat HRP were from Pierce. Tunicamycin and chloroquine were purchased from Sigma, TNFα from Alexis and proteases inhibitor cocktail set III from Calbiochem. 2.4. Cell cultures and transfection COS cells were grown in DMEM medium (Euroclone) supplemented with 10% foetal bovine serum (Gibco) and 100U/ml penicillin/streptomycin (Euroclone). Primary endothelial human cell line HUVEC and immortalized HUVEC-ST [19] were grown in EGM2 medium (Cambrex). Human monocytes were isolated from peripheral blood mononuclear cells (PBMCs) from buffy coats of volunteers and promoted their transition to macrophages in vitro as described [20]. COS and HUVEC cells were transiently transfected with Superfect (Qiagen) following the manufacturer's instructions, with a DNA/transfectant reagent ratio (w/v) of 1:5 in both cases. 2.5. Western blot analysis Transfected COS, HUVEC and human primary macrophages were rinsed twice with ice-cold phosphate-buffered saline (PBS) and lysed for 20 min at 4 °C in ice-cold extraction buffer (EB) containing 10 mM Tris–HCl pH 7.6, 100 mM NaCl, 10 mM EDTA, 0.5% Nonidet P40, 0.5% sodium deoxycholate, proteases inhibitor cocktail set III (0.1 mM AEBSF hydrochloride, 0.5 μM aprotinin, 5 mM Bestatin, 1.5 μM E-64, 10 μM Leupeptin, 1 mM Pepstatin A and 1 mM phenylmethylsulfonyl fluoride). After centrifugation (16,000 × g for 15 min), the supernatants were precipitated with 5 volumes of MeOH, for 2 h at − 20 °C. Pellets were dissolved in sample buffer (125 mM Tris–HCl pH 6.8, 1% SDS, 5% glycerol, 10 mM dithiothreitol and 0.05% bromphenol blue) and heated at 95 °C for 5 min. Equal amount of proteins were separated by SDS–PAGE in 12% acrylamide gels and transferred to polyvinylidene difluoride (PVDF) membranes (Amersham Biosciences) for 16 h at 30 V. Immunoreactive bands were visualized by ECL (Sigma). For separation of oligomeric forms, proteins were separated by NuPAGE 3–8% Tris–acetate gels (Invitrogen) in nonreducing conditions. 2.6. Immunoprecipitation Cells were harvested 24 h after transfection, rinsed twice with ice-cold PBS and lysed for 30 min in ice-cold EB. Lysates were centrifuged for 15 min at 12,000 × g and supernatants were subjected to immunoprecipitation with mouse anti-V5 IgG or rabbit anti-myc IgG, linked to 30 μl of protein-A-Sepharose beads (Sigma). Bound proteins were washed six times with Tris-buffered saline (TBS, 25 mM Tris, 140 mM NaCl, pH 8) containing Nonidet P40 0.1%, once with 5 mM Tris–Cl pH 7.6, eluted with sample buffer and separated by SDS–PAGE.

563

2.7. Endoglycosidase digestion Clarified cellular lysates were MeOH-precipitated and protein pellets were resuspended in denaturating buffer (0.5% SDS, 1% β-mercaptoethanol) and boiled for 10 min. Denaturated proteins were digested with endoglycosidase H (Endo-H) and peptide-N-glycosidase F (PNGase-F) (1000 units, Cell Signaling Technology) for 2 h at 37 °C. The mobility shift of deglycosylated LOX-1 and LOXIN isoforms was visualized by immunoblotting using anti-V5 antibodies. All experiments were performed in triplicate. 2.8. Immunofluorescence staining Immunofluorescence was performed as described [17]. Samples were examined with a confocal microscope (Nikon Instruments Spa, C1 on Eclipse TE200; EZC1 software). 2.9. Surface labelling quantification To reliably measure only the receptor protein expressed on the cell surface, we have quantified the surface proteins using a cyto-ELISA assay. COS cells were plated (2 × 104 cells/well) in flat-bottom 96-well microtiter plates (Falcon) and transfected. Twenty-four hours after transfection, cells were washed twice with ice-cold PBS and incubated with mouse anti-myc 9E10 IgG for 1 h at 4 °C. After washing with ice-cold PBS, cells were fixed with 4% paraformaldehyde for 10 min at 4 °C. Detection was performed by using donkey anti-mouse-HRP followed by incubation with o-phenylenediamine (Sigma). For each point, mean absorbance at 492 nm of blank wells (containing nontransfected cells) was subtracted from mean values of wells with transfected COS cells, to account for non-specific binding. The data presented represent three independent experiments performed in triplicate. 2.10. Ox-LDL preparation, labelling and fluorometric assay Human LDL was prepared from fresh healthy normolipidemic plasma of volunteers by ultracentrifugation [21]. LDL was extensively dialysed against PBS and protein concentration was determined by Bradford. LDL was oxidized at a concentration of 0.4 mg/ml by exposure to 5 μM CuS04 for 20 h at 30 °C. The reaction was stopped with 100 μM EDTA and extensively dialysed against PBS/EDTA. Ox-LDL was labelled with 1,1′-dioctadecyl-3,3,3′,3′-tetramethyllindocarbocyanine perchlorate (DiI) as described [22]. A stock solution of 3 mg/ml DiI in dimethyl-sulfoxide (DMSO) was prepared and was added to the ox-LDL solution to reach a ratio of 300 μg DiI to 1 mg ox-LDL. The mixture was incubated 16 h at 30 °C, dialyzed against PBS and filtered in 0.45 μm filter. Twenty-four hours after transfection cells were incubated with (DiI)-labelled ox-LDL in serum-free medium on ice for 1 h and washed several times with PBS. DiI fluorescence was observed with a Leica fluorescence microscope. Quantitation of ox-LDL receptor activity in cells was assayed by DiI extraction in isopropanol and fluorescence determined in a Perkin Elmer

564

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

Fig. 1. Natural and ectopic expression of LOX-1 receptor and LOXIN variant. (A) Heterologous expression of human LOX-1 and LOXIN proteins in COS cells. Cos cells transfected with plasmids encoding LOX-1-myc (lane 1) and LOXIN-V5 (lane 2) were lysed and immunoblotted with, respectively, 9E10 (anti-myc) and Mab anti-V5. (B) Lanes 1 and 2 show the naturally occurring expression of LOX-1 and LOXIN in human macrophages incubated in the absence (lane 1) and in the presence (lane 2) of 50 μg/ml ox-LDL. Lanes 3–4 show the endogenous expression of LOX-1 and LOXIN isoforms in HUVEC-ST cells cultured in the absence (lane 3) or in the presence of 50 ng/ml TNFα (lane 4). TNFα-induced HUVEC-ST cells were also treated with 25 μM chloroquine for 16 h (lane 5). Lanes 6 and 7 show the expression of human recombinant non-tagged LOX-1 (lane 6) and LOXIN (lane 7) in HUVEC-ST cells. All blots were probed with rat anti-human LOX-1 antiserum. Protein level was monitored by probing the same blot with anti-β-actin IgG (as shown in lanes 3–5). (C) Western blot of lysates from COS cells transfected with LOX-V5 (lanes 1– 3) and LOXIN-V5 (lanes 4–6) incubated with PNGase or Endo-H as indicated and immunoblotted with Mab anti-V5. (D) Western blot of lysated derived from COS (lanes 1–4) and HUVEC-ST cells (lanes 5–8). Cells were transfected with LOX-1-V5 (lanes 1–2 and 5–6) and LOXIN-V5 (lanes 3–4 and 7–8) and cultivated in the presence of tunicamycin (4 μg/ml) as indicated. Molecular weight markers are given in kilodaltons (kDa). Asterisks point to a degradation product of LOX-1 protein. Blots shown are representative of at least three different experiments.

spectrofluorometer with excitation and emission wavelengths set at 520 and 578 nm, respectively. 3. Results 3.1. Cloning and expression of LOX-1 and LOXIN tagged isoforms In order to verify the hypothesis of a direct protein–protein interaction between LOX-1 receptor and its splice variant LOXIN, we have added myc and V5 tag sequences to the Cterminal of both proteins and introduced the new constructs into different cell lines by transfection. The protein expression in transfected COS cells was studied by Western blot. As shown in Fig. 1A, specific double bands at 48 and 46 kDa for full-length

LOX-1 receptor and 38 and 36 kDa for LOXIN were visualized with anti-myc for LOX-1-myc (lane 1) and with anti-V5 for LOXIN-V5 (lane 2). Similar patterns of expression were also observed in kidney embryonic HEK-293, adrenal PC12 (not shown) and transformed human embryonic endothelial HUVEC-ST cell lines (Fig. 1D, lanes 5 and 7). The presence of endogenous LOX-1 and LOXIN isoforms was studied in human monocyte/macrophages by probing the blots with antiLOX-1/LOXIN polyclonal rat antiserum (Fig. 1B, lanes 1 and 2), and in transformed HUVEC-ST cells (Fig. 1B, lanes 3–5). As it can be seen, more than two immunospecific bands are specifically induced. To better understand the pattern of expression, we have transfected HUVEC-ST cells with nontagged LOX-1 (lane 6) and non-tagged LOXIN recombinant proteins (lane 7) and compared the exogenously transfected

Fig. 2. Co-immunolocalization of LOX-1 and LOXIN isoforms. (A) COS cells were transiently co-transfected with human recombinant LOX-1-myc and LOX-1-V5 (panels a–c) and with LOX-1-myc and LOXIN-V5 (panels d–i). The surface (panels a and d) and intracellular (panel g) localization of LOX-1-myc was determined using anti-myc antibodies. The surface distribution of LOX-1-V5 (panel b) and LOXIN-V5 (panel e) was analyzed with Mab anti-V5. For LOXIN-V5 the intracellular localization is shown in panel h. Panels c, f and i represent the merged images. Scale bar, 10 μm. (B) COS cells were transfected with LOXIN-V5 (panels a–c and g–i) or with LOX-myc and LOXIN-V5 (panels d–f and j–l). Double immunofluorescence shown in panels a–f was performed with Mab anti-V5 (panels a and d) and goat anti-calnexin IgG (panels b and e). In panels g–l staining was performed with rabbit anti-V5 IgG (panels g and j) and Mab anti-Golgi 58k protein (panels h and k). Panels c, f, i and l represent the merged images.

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

isoforms with the endogenously induced bands. A doublet at 46 and 42 kDa and a faint 48-kDa band which are up-regulated in the presence of 50 μg/ml of ox-LDL (lane 2), represent most

565

likely the full-length glycosylated LOX-1 receptor. A similar pattern is seen in HUVEC-induced cells. In these cells the higher molecular weight 48-kDa band is often the most intense

566

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

band. A 28-kDa band (Fig. 1B, lane 2 and lanes 4–5) is likely the LOXIN variant because the size of the heterologous expressed non-tagged LOXIN in HUVEC cells is also similar. A faint, but often visible, immunospecific lower molecular weight band at 24 kDa is also seen in induced HUVEC-ST (Fig. 1B, lanes 4–5) and in human macrophages (not detected in this gel). Degradation bands are seen, sometimes, when LOX-1 is ectopically expressed (see asterisks in Fig. 1C, lanes 1–3 and Fig. 1D, lanes 1 and 2). We used chloroquine to inhibit the activity of acidic hydrolases in the lysosomal degradation pathway. As shown in Fig. 1B (lane 5), 16 h incubation with chloroquine induced a significant accumulation of LOX-1/ LOXIN bands in TNFα-treated HUVEC-ST cells, indicating that these proteins are substrates of lysosomal enzymes and the detected bands are not proteolytic products. We used in vitro deglycosylation with PNGase-F to determine the contribution of N-linked glycans to the complexity of LOX-1 and LOXIN isoforms observed in SDS–PAGE, when heterologously expressed. Removal of the N-linked glycans resulted in the appearance of a single band at 37 kDa for LOX-1, which is the backbone of the full-length receptor (Fig. 1C, lane 2) and a single 28-kDa band for LOXIN (Fig. 1C, lane 5). Similar results were also obtained by the inhibition of N-linked glycosylation in vivo with tunicamycin in COS (Fig. 1D, lanes 2 and 4) and in HUVEC-ST cells (lanes 6 and 8). These results suggest that the LOX-1 and LOXIN doublets seen in gels resulted from a different extent of glycosylation in the post-translational processing. Since LOXIN has a diffuse ER localization [15] and proteins with complex N-linked oligosaccharides present in the midGolgi are resistant to Endo-H, we then assessed whether LOX-1 and LOXIN differ in their susceptibility to be cleaved by EndoH. The cell lysates derived from COS cells expressing LOX-1myc or LOXIN-V5 proteins were first treated with Endo-H and then immunoblotted with Mabs anti-Myc and anti-V5 (Fig. 1C, lanes 3 and 6). Endo-H treatment produced a complete removal of glycans in LOX-1 and LOXIN, indicating that both isoforms undertake the same glycosylation/maturation step.

(merged image in panel c). Interestingly, when we cotransfected LOX-1-myc and LOXIN-V5 at 1:1 ratio, we found a dramatic decrease of wild type LOX-1-myc fluorescent signal at the cell surface in all co-transfected cells (panel d). The merged image (panel f) shows a variable amount of areas of colocalization, with some cells in which the red colour (LOXIN) is more abundant. We then explored the effect of LOX-1 receptors on intracellular LOXIN localization in fixed and permeabilized cells (Fig. 2A, panels g–i). At 24 h after transfection, we found a complete co-localization of the two isoforms in the secretory compartment of all transfected cells (Fig. 2A, merged image in panel i). In order to better analyze the immunolocalization of LOXIN isoform, either when expressed alone or in combination, we performed a double immunofluorescence with the ER marker calnexin (Fig. 2B, panels a–f) and the Golgi marker 58K protein (Fig 2B, panels g–l). When LOXIN is transfected in the absence of LOX-1, it co-localizes with calnexin (Fig. 2B, panels a–c) but many cells have only a partial co-localization with the Golgi marker (Fig. 2B, panels g–i). On the contrary, when the two isoforms are co-transfected, LOXIN staining is superimposable with the Golgi marker in all transfected cells (Fig. 2B, panels j–l). These results suggest that LOX-1 and LOXIN interact in the secretory compartment and both molecules reaches the cell surface. 3.3. Down-regulation of surface LOX-1 receptors and inhibition of LOX-1 activity by LOXIN co-expression In order to better evaluate the effect of the co-expression, we have quantified the surface appearance of LOX-1 either when expressed by itself or in combination with increasing amounts of LOXIN variant by using a surface labelling quantification

3.2. Cellular localization of co-expressed LOX-1/LOXIN isoforms As indicated by genetic and biochemical studies, both LOX1 and LOXIN isoforms are co-expressed in vivo [15] (Fig. 1B). In a previous study we have found that LOXIN has a more diffuse ER staining compared to LOX-1 distribution and only 10% of LOXIN transfected cells show surface staining [15]. To address whether the two isoforms interact, we co-transfected myc-tagged LOX-1 and V5-tagged LOXIN in COS cells. Surface localization of LOX-1/LOXIN receptors was analyzed by double immunofluorescence analysis. We used polyclonal rabbit anti-myc antibodies to detect myc-tagged LOX-1 and Mab anti-V5 to detect LOX-1-V5 and LOXIN-V5. This analysis revealed that both human recombinant LOX-1-myc (Fig. 2A, panel a) and LOX-1-V5 fusion proteins (panel b) are efficiently expressed and co-localize in the cell surface, indicating that the two tags do not influence LOX-1 localization

Fig. 3. LOXIN effect on down regulation of LOX-1 surface receptors. COS cells were transfected with LOX-1-myc alone or LOX-1-myc and increasing concentration of LOXIN-V5 or the irrelevant protein R4-V5 as indicated. Surface receptors were measured by cytoELISA assay by using anti-myc antibodies.

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

567

Down-regulation of full-length LOX-1 receptors could give rise to a consequent impairment of ox-LDL binding and internalization. To study the receptor function, we have transfected COS cells with LOX-1 or LOXIN and, 24 h after transfection, incubated them with DiI-ox-LDL at 4 °C for 1 h. As shown in Fig. 4A, DiI-ox-LDL efficiently binds to fulllength receptors (panel b). A 100-fold excess of cold ox-LDL completely abolishes DiI-ox-LDL binding to LOX-1 (panel d), indicating the specificity of the binding. No fluorescence can be visualized in panel f, demonstrating that DiI-ox-LDL does not bind to LOXIN. Quantification of bound DiI-ox-LDL was obtained by its extraction with isopropanol and spectrofluorometric analysis. As can be seen in Fig. 4B, inhibition of ox-LDL binding is very severe and almost complete when LOXIN was co-expressed with the full-length receptor at a ratio 1:1. We also measured the intracellular uptake of ox-LDL in co-transfected cells by incubating the cells with DiI-ox-LDL at 37 °C for 4 h. Again, an almost complete inhibition of LOX-1 receptors was observed in the presence of LOXIN (data not shown). 3.4. LOX-1/LOXIN in vivo interaction The in vivo interaction between LOX-1 and the splice variant LOXIN in transfected cells was further confirmed by coimmunoprecipitation analysis. Constructs expressing myctagged LOX-1, the non-relevant Sec8H4-myc and V5-tagged LOXIN proteins were transiently co-transfected in COS cells, and cells were processed for analysis 24 h after transfection. Cellular lysates were first immunoprecipitated with anti-V5 antibodies, separated by SDS–PAGE and then immunoblotted

Fig. 4. Effect of LOXIN expression on ox-LDL binding. (A) Cells were incubated with 10 μg/ml of DiI-ox-LDL on ice for 1 h, washed three times and surface expressed LOX-1-myc or LOXIN-V5 receptors were visualized by antimyc (panels a and c) or anti-V5 IgG (panel e). Fluorescence of DiI-ox-LDL was detectable in cells transfected with LOX-1-myc (panel b) and not with LOXINV5 (panel f). Specific binding of ox-LDL was measured as displaceable binding with 100 μg/ml of unlabelled ox-LDL (panel d). Scale bar, 5 μm. (B) Binding of DiI-ox-LDL to COS cells transfected with LOX-1-myc alone, LOXIN-V5 alone or co-transfected with LOX-1-myc and increasing amounts of LOXIN-V5 was measured by extraction of DiI fluorescence with isopropanol as described in Materials and methods.

assay. LOX-1-myc, LOXIN-V5 and the non-relevant recombinant protein R4-V5 were co-transfected in COS cells at different concentrations, as indicated in Fig. 3, and the amount of exposed LOX-1 receptors was analyzed by a cytoELISA assay, as described in Materials and methods. As can be seen in the figure, when we co-transfected LOX-1/LOXIN at a ratio of 4:1 (100 ng of LOX-1 and 25 ng of LOXIN), it results in a 40–45% decrease of exposed LOX-1 receptors. The decrease is about 80% when the ratio between the two isoforms is 1:1. Note that co-expression of increasing amounts of the non-relevant R4 protein does not give any effect on LOX-1 receptors trafficking to the plasma membrane.

Fig. 5. In vivo interaction between LOX-1 full-length receptors and LOXIN truncated variant. COS cells were co-transfected with LOX-1-myc/LOX-V5 (lanes 1 and 6), LOX-1-myc and increasing concentrations of LOXIN-V5 as indicated (lanes 2–4 and 7–9) and Sec-8H4 (lanes 5 and 10). Half lysates were immunoblotted with anti-myc antibodies (lanes 1–5) and the remaining extracted proteins were first immunopurified with anti-V5 antibodies and then probed with anti-myc 9E10 (lanes 6–10). Note the 48 and 46 kDa LOX-1 immunopositive bands in lanes 7, 8 and 9 which indicate a direct interaction between LOX-1/LOXIN isoforms. The migration of molecular weight markers (kDa) is indicated on the right and the heavy (γ) and light (k) chains are indicated on the left.

568

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

with anti-myc Mab 9E10 to reveal LOX-1-myc and Sec8H4myc. The efficiency and specificity of the experiment were confirmed by analyzing the input of LOX-1-myc and Sec8H4myc in extracts before immunoprecipitation. As seen in Fig. 5, a similar amount of myc-tagged proteins were present in all samples (lanes 1–5). After immunoprecipitation with anti-V5 antibodies, we found the 48 and 46 kDa bands, corresponding to LOX-1-myc, either in extracts derived from cells expressing LOX-1-V5 and in cells expressing LOXIN-V5 (Fig. 5, lanes 6– 9). On the contrary, the non-relevant protein Sec8H4-myc, used as a control, did not co-immunoprecipitate with LOXIN-V5 (lane 10). Interestingly, while, as expected, the relative amount of the two 46–48 kDa immunoprecipitated bands was similar in extracts derived from cells expressing LOX-1-myc/LOX-1-V5 (lane 6), the lower 46-kDa band was selectively co-immunoprecipitated in a dose-dependent manner in cells expressing different amount of LOXIN-V5. 3.5. LOX-1/LOXIN homo- and hetero-oligomerization Previous studies suggested that full-length LOX-1 receptors function as a disulfide-linked homodimer [10–12]. This speculation is supported by the experiment shown in Fig. 6 by using 3–8% acrylamide gels in non-reducing conditions to better detect and separate high molecular weight forms. In COS cells expressing LOX-1 or co-expressing LOXIN plus LOX-1, we found that the wild-type LOX-1 (usually referred to 46– 48 kDa proteins) is expressed as a major 90-kDa band

Fig. 6. Homo- and hetero-oligomerization of LOX-1 and LOXIN isoforms. Lysates of COS cells expressing human recombinant LOX-1-V5 and LOXINV5 alone or in combination (as indicated) were separated by a NuPAGE 3–8% polyacrylamide gradient gel in non-reducing conditions and immunoblotted with anti-V5 antibodies (lanes 1–4). Lysates derived from COS cells transfected with LOX-1-myc and LOXIN-V5 or LOX-1-myc and LOX-1-V5 were first immunoprecipitated with anti-myc antibodies and then immunoblotted with anti-V5 antibodies (lanes 5 and 6) to identify the bound V5-tagged proteins. Separation of proteins was performed in non-reducing conditions. LOX-1 and LOXIN dimers, tetramers and hexamers are indicated, respectively, with D, T and HE and LOX/LOXIN hetero-tetramers are indicated with HT. The migration of molecular weight markers are shown on the right. Experiments were repeated three times with similar results.

corresponding to the homodimer and a minor band corresponding to the homotetramer that runs at 180 kDa (lane 4). The expression of LOXIN variant alone resulted in the appearance of two major bands, corresponding to LOXIN homodimer at 65 kDa and the homotetramer at 130 kDa. Other homooligomer forms at higher molecular weight are also visible, as the homohexamers (lane 1). Interestingly, in cells co-expressing LOX-1 and increasing amount of LOXIN (lanes 2 and 3), the 90-kDa band (LOX-1 homodimer) is still present, although at a decreasing level, and bands corresponding to LOX-1/LOXIN hetero-tetramers and hetero-hexamers (see asterisks in lanes 2 and 3) are now evident. In order to confirm the formation of stable hetero-oligomers, lysates derived from COS cells co-expressing myc- and V5tagged LOX-1 and LOXIN were first co-immunoprecipitated with anti-myc antibodies and then analyzed by Western blot with anti-V5 antibodies. As shown in Fig. 6, one major band at 90 kDa corresponding to LOX-1 homodimers is present when LOX-1-myc and LOX-1-V5 were co-transfected (lane 6), as expected. On the contrary, immunoprecipitation of lysates derived from COS cells co-expressing LOX-1-myc and LOXIN-V5 at a ratio 1:1 resulted in the co-immunoprecipitation of a major band at 160 kDa, corresponding to the LOX-1/ LOXIN hetero-tetramers (lane 5). 4. Discussion LOX-1 targeting could be a useful clinical approach to confer protection from atherosclerosis. In vivo infusion of antiLOX-1 monoclonal antibodies in animal myocardial ischemia– reperfusion models inhibits cellular apoptosis, decreases myocardial infarct size, and improves left ventricular functions [23]. In a previous study, we identified a new LOX-1 isoform, LOXIN, able to protect macrophages against ox-LDL-mediated apoptosis [15]. Here we describe the effects of heterologous expression of LOXIN in mammalian fibroblasts and human endothelial cells. When expressed in the absence of the wild-type LOX-1 receptor, LOXIN localizes mostly intracellularly, widespread in the endoplasmic reticulum. When, instead, both variants are co-expressed in host cells, LOXIN translocates to the plasma membrane and co-localizes with the full-length LOX-1 receptor. Biochemical studies confirmed the molecular LOX-1/LOXIN interaction and the formation of protein multimeric complexes. Hetero-oligomerization of LOX-1/LOXIN isoforms promotes cell-surface expression of intracellularly localized LOXIN and leads to a significant down-regulation of surface LOX-1 receptors with a consequent impairment of ox-LDL binding and uptake. Our studies suggest for the first time that heterooligomerization between naturally occurring isoforms of LOX1 receptor disrupts its functional properties. Therefore, in cells expressing both isoforms, abundance of expression of LOXIN may cause a resistance to ox-LDL induced apoptosis and cytotoxicity. The predicted size of LOXIN protein is 21.4 kDa. The polypeptide is composed of 188 aa N-terminal of wild-type

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

LOX-1 and an altered amino acid at its C-terminal. It lacks the distal 84 amino acids of LOX-1 including two extracellular lectin binding domains. Here we show that LOXIN protein is translated in vivo and induced by ox-LDL in human macrophages and by TNFα in human umbilical endothelial HUVEC cells. The size of the endogenous protein is slightly higher than that predicted, probably due to different extent of glycosylation. C-terminal sequences were found to be essential for the intracellular traffic and function of LOX-1. Deletion of the last nine residues of LOX-1 resulted in failure of cell-surface localization [24]. C-terminal truncated LOXIN splice isoform, as expected, had a typical ER distribution and only 10% of cells transfected with LOXIN-GFP showed surface staining [15]. Interestingly, LOXIN intracellular distribution markedly changed in the presence of full-length LOX-1 receptors. Most of cotransfected cells exhibited LOXIN variant also at the level of the plasma membrane. Their co-localization suggests that the two isoforms interact intracellularly during their biosynthesis. This conclusion is supported also by co-immunoprecipitation experiments in which LOXIN is shown to preferentially interact with the under-glycosylated 46-kDa band of LOX-1 (Fig. 5). We cannot exclude, however, that LOXIN may interfere with the correct glycosylation and maturation of wild-type LOX-1. Furthermore, there is an emerging role for binding and oligomerization of individual subunits in the biosynthesis and maturation of membrane receptors, as for example for Gprotein-coupled receptors (GPCRs) [25], and it is possible that expression of LOX-1 with LOXIN relieves a block on the intracellular or post-translational processing of the latter that allow it to be expressed on the cell surface. Although we do not have evidence for such processes yet, further studies are likely to clarify whether this is important in this interaction. One of the most surprising facts of this study was the decrease of LOX-1 receptors fluorescence signal at the cell surface when cells are co-transfected with increasing amount of LOXIN. Membrane-associated LOX-1 receptors decrease despite the fact that the expression level of LOX-1 does not change, as detected by immunoprecipitation and Western blot techniques. Co-expression of both isoforms appears to facilitate the appearance of LOXIN at the plasma membrane and a concomitant down-regulation of LOX-1 receptors. Whether this event relates to the receptor maturation, glycosylation and trafficking to the plasma membrane, as previously discussed, or to receptor recycling, internalization and degradation is still unknown. Using a biochemical approach and site-directed mutagenesis, it has been shown that although human LOX-1 is expressed at the cell surface as a disulfide-linked homodimer in vivo, it can form larger functional oligomers [9,26]. Plasma membrane localization and oligomerization of LOX-1 receptors determine their functionality [10,11,26] and single amino acid substitution in the dimer interface may cause a severe reduction in LOX-1 binding activity, as indicated by crystal studies [10]. In our study, we show that also LOXIN forms stable disulfide-linked homodimers. In particular, LOXIN forms a higher order molecular weight multimeric complexes compared to those formed by LOX-1 alone. Interestingly, by co-immunoprecipita-

569

tion assay in non-reducing conditions, LOX-1/LOXIN dimers interact and give rise, preferentially, to heterotetramers. This result is particularly important in view of the fact that, in vivo, increasing amount of LOXIN leads to loss of ox-LDL binding to LOX-1 expressing cells. It appears that the affinity of oxLDL to LOX-1/LOXIN receptors decreases dramatically. Since LOXIN protein, by itself, does not bind to ox-LDL (Fig. 4, panel f), LOX-1/LOXIN heterotetramers are, likely, nonfunctional and could explain the reason for the observed dramatic loss of function of LOX-1 receptors in cells expressing both isoforms. The increased presence of LOXIN in the plasma membrane would mean competition with the wild-type LOX-1 monomers/ dimers for oligomerization and their functional inhibition. This is particularly interesting given that LOX-1 is highly induced by ox-LDL ligand(s) and a small increase in LOXIN expression would result in a dramatic loss of receptor activity. In support to this hypothesis, it is worth mentioning that a truncated splice variant of the human receptor P2X7, P2X7-j, interacts with the full-length wild-type receptor, to form inactive hetero-oligomers, which showed diminished binding and channel function capacity [27]. Moreover, also expression of the mutant form of LOX-1 protein LOX-1-K267A, which acts as a dominant negative receptor of wild-type LOX-1, blocks ERK1/2 activation [28]. Heterodimerization between mutant and wild-type LOX-1 receptors may attenuate the expression of functional receptors and therefore the intake of ox-LDL inside the cells. Our results demonstrate a functional model of the LOX-1/ LOXIN system based on genetic variation in the human OLR1 gene [15]. The OLR1 LD1 polymorphisms regulate the expression of LOXIN mRNA as the mRNA levels were increased among the “no risk” homozygotes compared to other genotypes. Previous genetic studies reported positive data about the association of OLR1 gene polymorphisms with both coronary artery disease (CAD) susceptibility and CAD severity. These results, supported by functional analysis, strongly suggested that the protein products coded by OLR1 gene are an important factor in the pathogenesis of CAD and myocardial infarction [15]. There are increasing evidences for the role of allele specific non-coding SNPs in the regulation of human gene expression and its concomitant association with complex diseases [29]. Here, we provide the first direct functional characterization of a genetically produced OLR1 splicing isoform, LOXIN, that leads to a “protective” effect versus atherosclerosis and myocardial infarction. Observations in genetically modified mice (LOX-1 KO) provide evidence that deletion of LOX-1 ameliorates ox-LDL-mediated endothelial dysfunction and inhibits atherogenesis in the LDL receptordeficient mice model of atherosclerosis [30]. These findings, together with evidences that LOXIN exerts dominant-negative effect on LOX-1 function, make LOXIN an attractive new target for prevention and treatment of atherosclerosis. Acknowledgments We are very grateful to A. Desideri and M. Falconi for helpful discussions, T. Parasassi and G. Greco for discussions

570

S. Biocca et al. / Journal of Molecular and Cellular Cardiology 44 (2008) 561–570

and help in LDL preparation and G. Graziani for immortalized HUVEC-ST cells. This work was supported by the Italian Ministry of Health and Italian Ministry of University and by R.E.D.D. s.r.l., a spinoff of the Tor Vergata University of Rome.

[15]

[16]

References [17] [1] Binder CJ, Chang MK, Shaw PX, Miller YI, Hartvigsen K, Dewan A, et al. Innate and acquired immunity in atherogenesis. Nat Med 2002;8:1218–26. [2] Steinberg D. Atherogenesis in perspective: hypercholesterolemia and inflammation as partners in crime. Nat Med 2002;8:1211–7. [3] Chen M, Masaki T, Sawamura T. LOX-1, the receptor for oxidized low-density lipoprotein identified from endothelial cells: implications in endothelial dysfunction and atherosclerosis. Pharmacol Ther 2002;95:89–100. [4] Metha JL, Chen J, Hermonat P, Romeo F, Novelli G. Lectin-like, oxidized low-density lipoprotein receptor-1 (LOX-1): a critical player in the development of atherosclerosis and related disorders. Cardiovasc Res 2006;69:36–45. [5] Ehara S, Ueda M, Naruko T, Haze K, Itoh A, Otsuka M, et al. Elevated levels of oxidized low density lipoprotein show a positive relationship with the severity of acute coronary syndromes. Circulation 2001;103: 1955–60. [6] Sawamura T, Kume N, Aoyama T, Moriwaki H, Hoshikawa H, Aiba Y, et al. An endothelial receptor for oxidized low-density lipoprotein. Nature 1997;386:73–7. [7] Chen M, Narumiya S, Masaki T, Sawamura T. Conserved C-terminal residues within the lectin-like domain of LOX-1 are essential for the oxidized low-density-lipoprotein binding. Biochem J 2001;355: 289–96. [8] Xie Q, Matsunaga S, Shi X, Ogawa S, Niimi S, Wen Z, et al. Refolding and characterization of the functional ligand-binding domain of human lectinlike oxidized LDL receptor. Protein Expr Purif 2003;32:68–74. [9] Xie Q, Matsunaga S, Niimi S, Ogawa S, Tokuyasu K, Sakakibara Y, et al. Human lectin-like oxidized low-density lipoprotein receptor-1 functions as a dimer in living cells. DNA Cell Biol 2004;23:111–7. [10] Ohki I, Ishigaki T, Oyama T, Matsunaga S, Xie Q, Ohnishi-Kameyama M, et al. Crystal structure of human lectin-like, oxidized low-density lipoprotein receptor-1 ligand binding domain and its ligand recognition mode to ox-LDL. Structure 2005;13:905–17. [11] Park H, Adsit FG, Boyington JC. The 1.4 angstrom crystal structure of the human oxidized low density lipoprotein receptor lox-1. J Biol Chem 2005;280:13593–9. [12] Vohra RS, Murphy JE, Walker JH, Ponnambalam S, Vanniasinkam SH. Atherosclerosis and the lectin-like oxidized low-density lipoprotein scavenger receptor. Trends Cardiovasc Med 2006;16:60–4. [13] Mango R, Clementi F, Borgiani P, Forleo GB, Federici M, Contino G, et al. Association of single nucleotide polymorphisms in the oxidised LDL receptor 1 (OLR1) gene in patients with acute myocardial infarction. J Med Genet 2003;40:933–6. [14] Chen Q, Reis SE, Kammerer C, Craig WY, LaPierre SE, Zimmer EL, et al. Genetic variation in lectin-like oxidized low-density lipoprotein receptor 1

[18]

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28]

[29] [30]

(LOX1) gene and the risk of coronary artery disease. Circulation 2003;107:3146–51. Mango R, Biocca S, Del Vecchio F, Clementi F, Sangiuolo F, Amati F, et al. In vivo and in vitro studies supports that a new splicing isoform of OLR1 gene is protective against acute myocardial infarction. Circ Res 2005;97:152–8. Lener M, Horn IR, Cardinale A, Messina S, Nielsen UB, Rybak SM, et al. Diverting a protein from its cellular location by intracellular antibodies. The case of p21Ras. Eur J Biochem 2000;267:1196–205. Cardinale A, Filesi I, Vetrugno V, Pocchiari M, Sy MS, Biocca S. Trapping prion protein in the endoplasmic reticulum impairs PrPc maturation and prevents PrPSc accumulation. J Biol Chem 2005;280:685–94. Smith DB, Johnson KS. Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 1988;67:31–40. Tentori L, Vergati M, Muzi A, Levati L, Ruffini F, Forini O, et al. Generation of an immortalized human endothelial cell line as a model of neovascular proliferating endothelial cells to assess chemosensitivity to anticancer drugs. Int J Oncol 2005;27:525–35. Yuan XM, Brunk UT, Olsson AG. Effects of iron- and haemoglobinloaded human monocyte derived macrophages on oxidation and uptake of LDL. Arterioscler Thromb Vasc Biol 1995;15:1345–51. Sattler W, Bone P, Stocker R. Isolation of human VLDL, LDL, HDL and two HDL subclasses in the TL-100 tabletop centrifuge using the TLA100.4 rotor. Technical information Beckman; 1992. Stephan ZF, Yurachek EC. Rapid fluorometric assay of LDL receptor activity by DiI-labeled LDL. J Lipid Res 1993;34:325–30. Li D, Williams V, Liu L, Chen H, Sawamura T, Romeo F, et al. Expression of lectin-like oxidized low-density lipoprotein receptors during ischemia– reperfusion and its role in determination of apoptosis and left ventricular dysfunction. J Am Coll Cardiol 2003;41:1048–55. Shi X, Niimi S, Ohtani T, Machida S. Characterization of residues and sequences of the carbohydrate recognition domain required for cell surface localization and ligand binding of human lectin-like oxidized LDL receptor. J Cell Sci 2001;114:1273–82. Bulenger S, Marullo S, Bouvier M. Emerging role of homo- and heterodimerization in G-protein-coupled receptor biosynthesis and maturation. Trends Pharmacol Sci 2005;26:131–7. Matsunaga S, Xie Q, Kumanu M, Niimi S, Sekizawa K, Sakakibara Y, et al. Lectin-like oxidized low-density lipoprotein receptor (LOX-1) functions as an oligomer and oligomerization is dependent on receptor density. Exp Cell Res 2007;313:1203–14. Feng YH, Li X, Wang L, Zhou L, Gorodeski GI. A truncated P2X7 receptor variant (P2X 7-J) endogenously expressed in cervical cancer cells antagonizes the full-length P2X7 receptor through hetero-oligomerization. J Biol Chem 2005;281:17228–37. Tanigawa H, Miura S, Matsuo Y, Fujino M, Sawamura T, Saku K. Dominant-negative Lox-1 blocks homodimerization of wild-type Lox-1induced cell proliferation through extracellular signal regulated kinase 1/2 activation. Hypertension 2006;48:294–300. Knight JC. Regulatory polymorphisms underlying complex disease traits. J Mol Med 2005;83:97–109. Mehta JL, Sanada N, Hu CP, Chen J, Dandapat A, Sugawara F, et al. Deletion of LOX-1 reduces atherogenesis in LDLR knockout mice fed high cholesterol diet. Circ Res 2007;100:1634–42.