The uptake of OxLDLs (oxidized low density lipoproteins) by CD36-expressing macrophages in the arterial intima and the subsequent ‘foam cell’ formation represents a crucial step in the initiation and development of atherosclerotic plaques. The present study has addressed the function of the CD36 N-terminal cytoplasmic domain in the binding and internalization of OxLDL. A selection of CD36 N-terminal cytoplasmic domain mutants were generated and stably expressed in HEK-293 (human embryonic kidney) cells. The capacity of three mutants [CD36_C3/7-A (CD36-C3A/C7A), CD36_D4/R5-A (CD36-D4A/R5A) and CD36_nCPD− (CD36 lacking the N-terminal cytoplasmic domain)] to bind and endocytose OxLDL was then studied using immunofluorescence microscopy and quantitative fluorimetry. Each of the CD36 constructs was expressed at differing levels at the cell surface, as measured by flow cytometry and Western blotting. Following incubation with DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate)–OxLDL, cells bearing the CD36_wt (wild-type CD36), CD36_C3/7-A, CD36_D4/R5-A and CD36_nCPD− constructs all internalized DiI–OxLDL into endosomal structures, whereas empty-vector-transfected cells failed to do so, indicating that, unlike the C-terminal cytoplasmic domain, the N-terminal cytoplasmic domain is not essential for the endocytosis of OxLDL. In conclusion, the uptake of OxLDL by CD36 is not reliant on the presence of the CD36 N-terminal cytoplasmic domain. However, the N-terminal cytoplasmic domain may conceivably be implicated in the maturation of CD36.
The uptake of OxLDLs (oxidized low-density lipoproteins) by macrophages in the arterial intima and the subsequent ‘foam cell’ formation represents a crucial step in the initiation and development of atherosclerotic plaques. Uptake of OxLDL is mediated by certain members of the scavenger receptor family, a group of widely expressed and multifunctional membrane-embedded glycoproteins. A major receptor for OxLDL is CD36 , and CD36 has been implicated as a key player in foam cell formation and atherosclerotic lesion development. For example, macrophages obtained from CD36-deficient individuals bind and endocytose OxLDL far less effectively than those from normal subjects , and CD36-deficient macrophages are impaired in their capacity to secrete pro-inflammatory cytokines following OxLDL treatment , a hallmark event of atherogenesis. Furthermore, CD36/ApoE (apolipoprotein E) double-null mice develop markedly smaller atherosclerotic plaques than ApoE−/− mice .
On the basis of findings from studies such as Gruarin et al.  and Tao et al. , CD36 is predicted to adopt a ditopic configuration, meaning it comprises an extracellular domain (residues 29–438) that is flanked by two transmembrane domains (residues 7–28 and 439–460) and two cytoplasmic domains (residues 1–6 and 461–472). A 28-residue segment in the extracellular domain (155–183) has been found to mediate binding with OxLDL . Interestingly, despite the compartmental separation between the extracellular domain and the C-terminal cytoplasmic domain, the latter has been found to be essential for the binding and uptake of OxLDL, as well as in signal transduction processes. We have demonstrated previously  that HEK-293 (human embryonic kidney) cells expressing a construct without the last six residues (467–472) of the C-terminal cytoplasmic domain displayed an approx. 80% decrease in their capacity to bind and internalize OxLDL. The C-terminal cytoplasmic domain also possesses signal transduction properties; Lipsky et al.  reported a significant decrease in OxLDL-mediated NF-κB (nuclear factor κB) activation in CHO (Chinese-hamster ovary) cells expressing a CD36 construct without the C-terminal cytoplasmic domain compared with cells bearing the wild-type protein. Similarly, the anti-angiogenic effect of TSP-1 (thrombospondin-1) observed in CD36-transfected endothelial cells is abrogated if certain residues in the C-terminal cytoplasmic domain (Cys464, Arg467 and Lys469) are mutated .
To our knowledge, the CD36 N-terminal cytoplasmic domain has not been investigated with regards to its involvement in ligand binding and uptake. The oppositely charged aspartate–arginine sequence at positions 4 and 5 respectively may provide a means for CD36 to interact with co-receptors and/or facilitate the capping and endocytosis of CD36–ligand complexes. Similarly, the palmitoyl group shown to be appended to each of the cysteine residues at positions 3 and 7  may mediate an interaction with membrane or cytoplasmic proteins that is necessary for CD36 binding and endocytosis. Hence, to investigate the role of the N-terminal cytoplasmic domain in the binding and internalization of OxLDL, a selection of CD36 N-terminal cytoplasmic domain mutants were generated and stably expressed in HEK-293 cells. The capacity of these mutants to bind and endocytose OxLDL was studied using immunofluorescence microscopy and quantitative fluorimetry.
MATERIALS AND METHODS
HEK-293 cells were maintained in DMEM (Dulbecco's modified Eagle's medium), supplemented with 10% (v/v) fetal calf serum, 2 mM glutamine, 10 μg/ml streptomycin and 10 units/ml penicillin, in a 5% CO2/95% air atmosphere at 37°C.
The CD36 coding sequence was amplified by PCR from an M13mp19 cloning vector containing the CD36 coding sequence using the wild-type forward and reverse primers shown in Table 1 and subcloned into the pcDNA3.1(+) expression vector. Three N-terminal cytoplasmic domain mutant constructs (Figure 1) were generated by mutation of the CD36_wt (wild-type CD36) sequence by PCR using the primer sequences shown in Table 1. As with the wild-type construct, all mutated constructs were subcloned into the pcDNA3.1(+) expression vector and validated by DNA sequencing.
|Construct .||Forward (5′→3′) .||Reverse (5′→3′) .|
|Construct .||Forward (5′→3′) .||Reverse (5′→3′) .|
Sequence of generated CD36 N-terminal cytoplasmic domain mutants
Stable expression of CD36 constructs in HEK-293 cells
Transfection of plasmid DNA into HEK-293 cells was carried out using PolyFect reagent (Qiagen), according to the manufacturer's instructions. After transfection (48 h), growth medium was removed and replaced with complete medium containing geneticin (500 μg/ml) in order to select for stably transfected cells.
SDS/PAGE, Western blotting and immunodetection
Cells were harvested with trypsin/EDTA, washed twice with PBS and lysed into 300 ül of Triton X-100 (5%). Protein concentration was determined using the bicinchoninic acid assay. SDS/PAGE, Western blotting and immunodetection of CD36 were carried out essentially as described by Malaud et al. .
Cells (106) were incubated in PBS containing mouse anti-(human CD36) monoclonal antibody 10.5 (5 μg/ml) (Biocytex) for 1 h at 4°C with vertical rotation. After a single wash with PBS, cells were incubated in PBS containing a mouse anti-human FITC-conjugated antibody (80 μg/ml) for a further 1 h at 4°C with vertical rotation. Finally, cells were washed to remove any residual antibody, resuspended in 1% paraformaldehyde and analysed using a FC 500 flow cytometer (Becton Coulter) with RXP analysis software (Cytomics).
Visualization of DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate)–OxLDL uptake
HEK-293 cells expressing the wild-type CD36 construct, as well as the mutated CD36 constructs, were seeded on to six-well plates (400000 cells/well) and grown on coverslips for 24 h before being incubated with 50 μg/ml DiI–OxLDL (in complete medium) for 16 h. Cells were rinsed with PBS, fixed with 4% (w/v) paraformaldehyde, mounted on to microscope slides using Vectashield (Vector Laboratories) and viewed at the specified magnification using a fluorescence microscope.
Quantification of DiI–OxLDL uptake
HEK-293 cells stably transfected with the plasmids described above were seeded on to six-well plates (400000 cells/well) and cultured for 24 h before being incubated with 0, 25 and 50 μg/ml DiI–OxLDL (in complete medium) for 16 h at 37°C. Uptake of DiI–OxLDL was quantified according to the method of Luan and Griffiths . Cells were harvested with trypsin/EDTA, washed twice with PBS and lysed with 5% (w/v) Triton X-100. DiI–OxLDL fluorescence was measured at 590 nm following excitation at 520 nm using a SpectraMax microplate reader (Molecular Devices). Fluorescence values were converted into μg of DiI–OxLDL via a standard curve produced by measuring the fluorescence at 590 nm of lysis buffer (5% Triton X-100 in PBS) containing 0.1, 0.2, 0.5, 1.0 and 1.5 μg of DiI–OxLDL. The protein concentration of each sample was determined via the bicinchoninic acid assay. Finally, the quantity of DiI–OxLDL taken up by cells was calculated as the amount of OxLDL per mg of cellular protein.
Expression of CD36 constructs in HEK-293 cells
HEK-293 cells were stably transfected with a plasmid encoding the CD36_wt sequence, as well as plasmids encoding the N-terminal cytoplasmic domain mutant constructs. All constructs were expressed, although at different levels, as determined by Western blotting (Figure 2). Interestingly, the anti-CD36 antibody 10.5 recognized two distinct bands in proteins derived from cells expressing the CD36_nCPD− (CD36 lacking the N-terminal cytoplasmic domain) construct. These may represent differentially glycosylated forms of CD36 and therefore would suggest an involvement of the N-terminal cytoplasmic domain region in the maturation process. Each of the four CD36 constructs was expressed at different levels at the cell surface, as measured by flow cytometry (Figure 3a). The CD36_C3/7-A (CD36-C3A/C7A) construct was marginally more abundant than the wild-type construct, whereas the CD36_D4/R5-A (CD36-D4A/R5A) and the CD36_nCPD− constructs were approx. 2- and 3-fold more abundant respectively than the wild-type construct (Figure 3b).
Detection of CD36 constructs in stably transfected HEK-293 cells by Western blotting
Analysis of CD36 construct expression in stably transfected HEK-293 cells by flow cytometry
Visualization of DiI–OxLDL uptake by HEK-293 cells bearing CD36 mutants
To determine whether the CD36 N-terminal cytoplasmic domain was required for binding and internalization of OxLDL, the passage of fluorescently labelled DiI–OxLDL was tracked in HEK-293 cells stably expressing each of the CD36 constructs using immunofluorescence microscopy. Following incubation with 10 μg/ml DiI–OxLDL for 16 h at 37°C, cells with the CD36_wt, CD36_C3/7-A, CD36_D4/R5-A and CD36_nCPD− constructs all internalized DiI-OxLDL into endosomal structures, whereas ev (empty vector)-transfected cells failed to do so (Figure 4), indicating that, unlike the C-terminal cytoplasmic domain, the N-terminal cytoplasmic domain is not essential for the endocytosis of OxLDL.
Visualization of OxLDL uptake in HEK-293 cells bearing different N-terminal cytoplasmic domain CD36 constructs
Quantification of DiI–OxLDL uptake by HEK-293 cells with the CD36 mutants
Despite the apparent expendability of the N-terminal cytoplasmic domain for OxLDL uptake, it was considered that different constructs may still display variations with respect to their binding and internalization efficacy. In order to compare the relative effectiveness of each construct to bind and internalize OxLDL, cells stably expressing each construct were incubated with DiI–OxLDL, as described in the Materials and methods section. Each of the cell types accumulated DiI–OxLDL in a dose-dependent manner, again indicating that the N-terminal cytoplasmic domain is not essential for OxLDL binding and uptake. HEK-293-CD36_wt cells accumulated approximately twice as much DiI–OxLDL as ev-transfected control cells, but an equal amount to HEK-293-CD36_C3/7-A cells, whereas HEK-293-CD36_D4/R5-A and HEK-293_nCPD− cells accumulated approx. 3- and 4-fold more DiI-OxLDL respectively than ev-transfected control cells (Figure 5a). This variation was thought to be a result of the different levels of expression of each construct. Indeed, when the results were normalized to compensate for the differential expression of each construct, the efficacy of each cell type to accumulate DiI–OxLDL is approximately equivalent (Figure 5b).
Accumulation of DiI–OxLDL in HEK-293 cells bearing wild-type and mutant CD36 constructs
Class B scavenger receptors, such as CD36, are composed of an extracellular domain, two transmembrane domains and two cytoplasmic domains. The C-terminal cytoplasmic domain of CD36 has been shown previously to be crucial in the binding and internalization of OxLDL  and the Gram-positive bacterium Staphylococcos aureus , and is also required for TSP-1-induced anti-angiogenic signalling , as well as NF-κB transactivation . To the best of our knowledge, the role of the CD36 N-terminal cytoplasmic domain in such events has yet to be addressed.
The N-terminal cytoplasmic domain of CD36 comprises the following residues: Met-Gly-Cys-Asp-Arg-Asn-Cys. Two regions of this sequence were felt to be of interest. First, the cysteine residues (at positions 3 and 7) are known to carry palmitate groups. Such moieties have been demonstrated previously to mediate protein–protein and protein–lipid interactions. Secondly, the oppositely charged aspartate (negative)–arginine (positive) sequence (at positions 3 and 4 respectively) may also permit interactions with other receptor proteins and/or cytosolic proteins. To investigate the role played by the N-terminal cytoplasmic domain (with specific emphasis on the above-mentioned regions) in the expression of CD36 and the internalization of OxLDL, three different mutants were generated by standard PCR mutagenesis techniques. In the first mutant (CD36_C3/7-A), the two cysteine residues were mutated into alanine; in the second mutant (CD36_D4/R5-A), the aspartate and arginine residues were mutated into alanine; in the third mutant (CD36_nCPD−), the entire N-terminal cytoplasmic domain was removed. The three CD36 mutants, as well as the wild-type construct, were then stably transfected in HEK-293 cells. All constructs (including the wild-type) were synthesized and expressed at the cell surface, as determined by flow cytometry, albeit at differing levels; the wild-type and the CD36_C3/7-A construct were expressed at the cell membrane at a similar level, whereas the CD36_D4/R5-A construct and the CD36_nCPD− construct were expressed approx. 2- and 3-fold higher respectively than the wild-type construct. The fact that the N-terminal cytoplasmic domain constructs were expressed more highly than the wild-type protein at the cell surface would suggest that this region may suppress the migration of the CD36 protein from the cytoplasm to the plasma membrane, perhaps in order to lengthen its exposure to the various post-translational modification enzymes needed for tertiary structure acquisition and attachment of extra groups, such as sugar moieties. Indeed, this may explain why the CD36_nCPD− construct seemed to comprise two species when analysed by Western blotting; instead of a single band being present at approx. 85 kDa, two bands were present: one at approx. 85 kDa and another at approx. 80–82 kDa. The smaller band may represent an incompletely glycosylated species and would may indicate that part of or the entire N-terminal cytoplasmic domain is required for glycosylation of CD36. An assessment of the effect of removing glycosylation sites in CD36 on its membrane trafficking would be advantageous. It should be noted that, in our hands, PNGase F (peptide N-glycosidase F)-mediated deglycosylation of CD36_wt eliminated recognition by the anti-CD36 monoclonal antibody 10.5 (results not shown), suggesting that the epitope recognized by the 10.5 antibody is composed of sugar groups, as well as protein. Indeed, the region of CD36 thought to be bound by the 10.5 antibody holds one N-glycan group.
Immuofluorescence microscopy was next used to follow the movement of fluorescently labelled DiI–OxLDL in the transfected HEK-293 cells expressing the various CD36 mutants. DiI–OxLDL was consistently observed to accumulate in cells expressing the CD36_wt construct, as well as the three mutated forms, but not in ev-transfected (control) cells, following an overnight incubation, indicating that the N-terminal cytoplasmic domain is not essential for OxLDL uptake. A fluorescence-based assay adapted from Luan and Griffiths  was utilized to quantify the uptake of DiI–OxLDL by cells expressing each of the CD36 constructs. Consistent with the immunofluorescence study, cells bearing the different mutants were capable of binding and endocytosing OxLDL. In such analyses, there was a notable difference in the quantity of DiI–OxLDL taken up by each cell type. Cells expressing the wild-type and CD36_C3/7-A constructs both internalized approx. 2-fold more DiI–OxLDL than the ev-transfected cells, whereas cells expressing the CD36_D4/R5-A and CD36_nCPD− constructs internalized approx. 4- and 5-fold more DiI–OxLDL respectively than the ev-transfected cells. These differences were felt to be due at least partially to the fact that the different constructs were expressed at different levels at the cell surface of transfected cells. As a means of compensating for the disparity in construct expression, all DiI–OxLDL uptake values were normalized against the amount of CD36 expressed by each cell line. After doing this, there was a negligible difference in the efficacy of each cell type to take up DiI–OxLDL, again suggesting that the CD36 N-terminal cytoplasmic domain is not involved in the endocytotic mechanism. To conclude, the uptake of OxLDL by CD36 is not reliant on the presence of the CD36 N-terminal cytoplasmic domain. However, this region may somehow be implicated in the maturation of CD36. Specifically, it could be involved in retarding the movement of the CD36 polypeptide, following translation, from the cytoplasm to the plasma membrane in order for it to attain its final native state structure and receive additional moieties, such as sugar chains and palmitoyl groups.
CD36 lacking the N-terminal cytoplasmic domain
human embryonic kidney
nuclear factor κB
oxidized low-density lipoprotein
C.M.-R., J.M. and S.C.-T. were supported at the Thrombosis Research Institute by a Garfield Weston Foundation Grant.