Adaptor protein Disabled-2 modulates low density lipoprotein receptor synthesis in fibroblasts from patients with autosomal recessive hypercholesterolaemia

Abstract

Autosomal recessive hypercholesterolaemia (ARH), characterized clinically by severe inherited hypercholesterolaemia, is caused by recessive null mutations in LDLRAP1 (formerly ARH). Immortalized lymphocytes and monocyte-macrophages, and presumably hepatocytes, from ARH patients fail to take up and degrade plasma low density lipoproteins (LDL) because they lack LDLRAP1, a cargo-specific adaptor required for clathrin-mediated endocytosis of the LDL receptor. Surprisingly, LDL-receptor function is normal in ARH patients' skin fibroblasts in culture. Disabled-2 (Dab2) has been implicated previously in clathrin-mediated internalization of LDL-receptor family members, and we show here that Dab2 is highly expressed in skin fibroblasts, but not in lymphocytes. SiRNA-depletion of Dab2 profoundly reduced LDL-receptor activity in ARH fibroblasts as a result of profound reduction in LDL-receptor protein, but not mRNA; heterologous expression of murine Dab2 reversed this effect. In contrast, LDL-receptor protein content was unchanged in Dab-2-depleted control cells. Incorporation of ³⁵S-labelled amino acids into LDL receptor protein revealed a corresponding apparent reduction in accumulation of newly synthesized LDL-receptor protein on depletion of Dab2 in ARH, but not in control, cells. This reduction in LDL-receptor protein in Dab2-depleted ARH cells could not be reversed by treatment of the cells with proteasomal or lysosomal inhibitors. Thus, we propose a novel role for Dab2 in ARH fibroblasts, where it is apparently required to allow normal translation of LDL receptor mRNA.

INTRODUCTION

Autosomal recessive hypercholesterolaemia (ARH, OMIM #603813) is a recessive disorder characterized by severe hypercholesterolaemia, xanthomas and premature atherosclerosis, suggestive of the common inherited disorder of familial hypercholesterolaemia (FH, OMIM #144010). The underlying cause is recessive mutations in a gene encoding the novel adaptor protein ARH, now called low density lipoprotein-receptor adaptor protein 1 (LDLRAP1, GenBank accession no. NM015627). As with FH, the hypercholesterolaemia in ARH results from impaired LDL receptor-mediated uptake of plasma LDL by the liver, the major site for cholesterol metabolism (1).

The LDL receptor is a recycling receptor that is internalized via clathrin-mediated endocytosis, a process by which virtually all eukaryotic cells internalize nutrients, antigens, growth factors and recycling receptors, including the transferrin (Tf) receptor (2). Epstein–Barr virus (EBV)-transformed lymphocytes isolated from ARH patients fail to internalize the LDL receptor, while Tf-receptor mediated uptake and recycling of ¹²⁵I-Tf occurs normally, suggesting that the internalization defect does not affect all clathrin-mediated endocytosis, but rather is specific for the LDL receptor (3). Retroviral expression of wild-type human LDLRAP1 in an EBV-lymphocyte cell line from an ARH patient restored LDL-receptor internalization, confirming that LDLRAP1 is required specifically for clathrin-mediated endocytosis of the LDL receptor in these cells (4). Moreover, LDLRAP1-deficient mice develop hyper-cholesterolaemia and exhibit defective internalization of LDL receptors in their hepatocytes in vivo (5).

Clathrin-mediated endocytosis is thought to be initiated by the recruitment of the adaptor complex AP2 to phosphatidylinositol 4,5-biphosphate [PtdIns (4,5)P2] on the plasma membrane, where it serves to anchor clathrin to the plasma membrane and to coordinate coat assembly with the selection of cargo proteins, as well as with the recruitment of accessory proteins required for vesicle formation (6). LDLRAP1 contains a phosphotyrosine-binding (PTB) domain that can bind directly to the NPVY internalization motif in the cytoplasmic tail of the LDL receptor and to phosphatidylinositol-4,5-bisphosphate [PtdIns (4,5)P₂] (7,8). LDLRAP1 also contains a potential clathrin-binding motif (LLDLE) which mediates high affinity binding to the clathrin heavy chain (7,8). In addition, a highly conserved AP2-binding domain has been identified in LDLRAP1 that is required for binding to the β-adaptin subunit of AP2 (7). Thus, LDLRAP1 demonstrates all the features of a cargo-specific adaptor protein and is now widely viewed as an LDL receptor-specific endocytic adaptor protein.

A puzzling feature of LDLRAP1 is that while LDL-receptor internalization is severely disrupted in EBV-lymphocytes (3,9), monocyte macrophages (4) and, presumably, liver of ARH patients, it appears to occur normally in patients' skin fibroblasts in culture (4,10,11). Disabled-2 (Dab2) is an alternative PTB domain-containing protein that, like LDLRAP1, binds the NPVY motif found in the LDL receptor, interacts directly with phosphoinositides (12) and binds both clathrin triskelia and the α-adaptin subunit of AP2 (13). Moreover, Dab2 has been implicated as an endocytic adaptor for other members of the LDL-receptor family such as the apolipoprotein E receptor 2 (apoER2) (14) and megalin (15,16). Recently, it has been shown that in HeLa cells, simultaneous siRNA-mediated reduction of both LDLRAP1 and Dab-2 results in impaired LDL receptor mediated internalization of LDL and accumulation of LDL receptor protein at the cell surface (17,18).

We also aimed to determine whether Dab2 could function as a specific adaptor for LDL-receptor internalization in fibroblasts from patients lacking functional LDLRAP1. We have found that siRNA-mediated depletion of Dab-2 results in almost complete loss of LDL-receptor protein from ARH cells, but not from control cells. We have also examined the effect of α-adaptin depletion in control and ARH skin fibroblasts to assess the importance of the AP2 adaptor complex to the function of LDLRAP1 and Dab2 as endocytic adaptor proteins for the LDL receptor.

RESULTS

Expression of Dab2 and analysis of LDL-receptor function in cells lacking Dab2

Immunoblotting of whole cell extracts revealed high levels of Dab2 protein in cultured normal skin fibroblasts, whereas it was barely detectable in normal EBV-lymphocytes (Fig. 1A). In contrast, LDLRAP1 was detectable in both cell types (Fig. 1B).

Figure 1.

Expression of Dab2 and LDLRAP1. Cell extracts were fractionated on non-reduced SDS–polyacrylamide gels (10%), transferred to nylon membranes and immunoblotted with antibodies to Dab2, LDLRAP1, AP2-α or AP2-µ2 as indicated on the side of the gel images, and with anti-γ-tubulin as a protein loading control. (A and B) Total cell extracts (25 µg of protein/lane) of skin fibroblasts (fibro) or EBV-transformed-lymphocytes (lym). The approximate MW (kDa) of protein markers run on the same gel is indicated. (C and D) Depletion of Dab2. Whole cell lysates of cultured skin fibroblasts from a control subject (C) or an ARH patient (D) were prepared 92 h after transfection with different siRNA, as indicated below: none, untransfected cells; con, non-targeting; Dab2-02 and Dab2-03, two different siRNA targeted to Dab2; AP2-α, siRNA to α-adaptin subunit of AP2.

Dab2 expression was specifically and efficiently knocked down in both control fibroblasts (Fig. 1C) and ARH fibroblasts (Fig. 1D) with two different siRNA oligonucleotides targeted to different regions of Dab2 (siDab2-02 and -03). Conditions for transfection were optimized with fluorescently labelled siRNA (data not shown). Both Dab2 siRNAs were effective in reducing Dab2 protein under conditions that resulted in 85–95% transfection efficiency, but a more complete knockdown was achieved using siDab2-02. A non-targeting siRNA had no effect on Dab2 expression in control or ARH cells (Fig. 1C and D) and depletion of Dab2 had no effect on the expression of AP2-α or -µ2.

Figure 2 shows the effect of Dab2-depletion on the uptake and degradation of labelled LDL by control fibroblasts and by fibroblasts from an ARH patient homozygous for Q136X in LDLRAP1. This mutation is known to result in undetectable levels of LDLRAP1 protein in cells (9). Saturable LDL receptor-dependent uptake (binding plus internalization) and degradation of ¹²⁵I-labelled LDL was determined after depletion of Dab2 and pre-incubation of the cells in sterol-depleted medium to upregulate LDL-receptor expression. In three separate experiments, depletion of Dab2 in control fibroblasts resulted in a slight reduction in uptake of LDL to 86 ± 2% of that in cells transfected with a non-targeting siRNA (Fig. 2A, filled triangles versus filled circles). In contrast, depletion of Dab2 in the ARH fibroblasts rendered the cells essentially unable to take up ¹²⁵I-LDL, reducing uptake to 28 ± 4% of that in cells transfected with a non-targeting siRNA (Fig. 2B). Degradation of ¹²⁵I-LDL reflected uptake in both control (Fig. 2C) and ARH (Fig. 2D) fibroblasts, showing that residual uptake by ARH cells was not due solely to binding to LDL receptors on the cell surface. These findings were confirmed in skin fibroblasts from two different control subjects and two unrelated ARH patients (Table 1).

Figure 2.

Uptake and degradation of ¹²⁵I-LDL by human skin fibroblasts transfected with siRNA. Skin fibroblasts from control subjects (A and C) or an ARH patient [homozygous Q136X (B and D)] were transfected with non-targeting siRNA (siCon), or siRNA targeted to either α-adaptin (siAP2) or Dab2 (siDab2). After 48 h, cells were transferred to medium containing LPDS for a further 40 h and then incubated for 4 h at 37°C with ¹²⁵I-labelled LDL at the indicated concentrations. Saturable uptake (A and B) and degradation (C and D) of LDL was determined in duplicate dishes as the difference in the amount of cell-associated or trichloroacetic acid-soluble, non-iodide radioactivity in the medium of cells incubated in the presence and absence of an excess of unlabelled LDL (1 mg/ml); the data shown are the mean of three separate experiments ± SEM. Non-saturable uptake of LDL by control cells was always <5% of total.

Table 1.

Binding and internalization of ¹²⁵I-LDL by human fibroblasts transfected with siRNA

siRNA	Binding + uptake of LDL by siRNA-transfected cellsa
	Control 1	Control 2	ARH 1 (Q136X)	ARH 2 (W22X)
siDab2-02	86 ± 2	82 ± 3	28 ± 4	30 ± 3
siDab2-03	81 ± 4	87 ± 3	37 ± 3	42 ± 3
siAP2-α	92 ± 2	93 ± 3	76 ± 4	79 ± 4

^aExpressed as % of LDL uptake by cells treated with a non-targeting siRNA, determined as described in the legend to Figure 2.

In order to assess the involvement of AP2 in endocytosis of the LDL-receptor in fibroblasts, the α-adaptin subunit (AP2-α) was depleted by siRNA targeting. As expected, this also resulted in slightly reduced levels of the µ2 subunit of AP2 (Fig. 1C and D). As shown by the open triangles in Figure 2A and B, LDL uptake was not affected by this partial depletion of AP2 in control fibroblasts, but in ARH fibroblasts, it was slightly reduced compared with that in cells transfected with a non-targeting siRNA (to 76 ± 4%). However, the observed effect was much less profound than that caused by Dab2 depletion (Fig. 2B). Similar observations were made in fibroblasts from two different control subjects and two ARH patients (Table 1).

Effect of AP2 and Dab2 depletion on LDL-receptor protein

To investigate further the observed reduction in LDL-receptor activity in Dab2-depleted ARH cells, LDL-receptor protein levels in whole cell extracts were assessed by immunoblotting, as shown in Figure 3A. Depletion of AP2-α had little or no effect on LDL-receptor protein in either control or ARH cells. Dab2 knockdown, however, resulted in significant reduction of LDL-receptor protein in ARH cells but not in control cells (Fig. 3A). Depletion of Dab2 with two different siRNA in cells from two control subjects and two ARH patients with different mutations confirmed a decrease in LDL-receptor protein of up to 75% in ARH cells, but little or no change in cells from control subjects (Fig. 3B). Thus the reduction in LDL receptor activity in Dab2-depleted cells resulted from the marked reduction in LDL-receptor protein content, but a similar reduction could not account for the 25% fall in activity in AP2-depleted cells.

Figure 3.

LDL-receptor protein in AP2- and Dab2-depleted fibroblasts. (A) Whole cell extracts of cells treated with siRNA (as indicated below the gel image) and pre-incubated with LPDS were fractioned on non-reduced SDS–polyacrylamide gels (10%) and immunoblotted with antibodies to the LDL receptor and γ-tubulin, as indicated on the right. (B) The ratios of LDL receptor to γ-tubulin in cells transfected with two different Dab2 siRNA (SiDab2-02 and SiDab2-03) were quantified by densitometry and expressed relative to cells transfected with non-targeting siRNA (control); Con-1 and -2 cells, skin fibroblasts from two control subjects; ARH-1 and -2, skin fibroblasts from ARH patients homozygous for the Q136X (ARH-1) and the W22X (ARH-2) mutations. The data shown are the mean of three separate experiments ± SEM. (C) ARH cells on cover slips were treated with Dab-02 siRNA, incubated with LPDS and then fixed and stained with antisera to Dab2 (red) and LDL receptor (green), with the overlay shown on the right; upper panels, cell surface; lower panels, section through centre of cell. Asterisks indicate cells in which Dab2 was depleted (D) Fibroblasts from an ARH patient (heterozygous for two defective alleles, a large deletion of approximately exons 6–9 and insC₆₂₀) were transfected on day 1 with Dab2 siRNA, then re-transfected on day 3 with a plasmid expressing either GFP (upper panels) or murine Dab2-GFP (mDab2-GFP, lower panels). On day 5, the cells were fixed and stained with anti-serum to the LDLR (red).

Confocal microscopy of cells stained with fluorescently labelled antibodies confirmed that while siRNA depletion of Dab2 had no effect on LDL receptor protein in control cells (data not shown), it had a profound effect in ARH cells. ARH cells depleted of Dab2 contained no detectable LDL-receptor protein, either inside the cell or on the cell surface (Fig. 3C). Three blue DAPI-stained nuclei of cells that were depleted of Dab2 are marked with an asterisk (*), and these contain no visible LDL-receptor protein; in contrast, a single cell on the same cover slip (upper left) in which Dab2 (red) was not depleted contains a normal amount of LDL-receptor protein (green).

To ensure that the effect of Dab2 siRNA on LDL-receptor protein levels was caused by the reduction in Dab2 expression and not by off-target effects, Dab2-depleted ARH cells were plated onto cover slips and then re-transfected 2 days later with a plasmid expressing green fluorescent protein (GFP) or GFP-tagged murine Dab2 p96 (13), which does not contain the siRNA target sequence. Two days later, the cells were immunostained for LDL-receptor protein and examined by confocal microscopy for expression of GFP (green) and LDL receptor (red), as shown in Figure 3D. The efficiency of Dab2 depletion in this experiment was at least 85%, as estimated from the number of cells with undetectable LDL-receptor protein. Expression of GFP did not restore LDL-receptor protein in any of 32 cells counted, but all GFP-Dab2-expressing cells also contained LDL receptor protein (28 cells counted).

The reduction in LDL-receptor protein was not secondary to a reduction in LDL-receptor mRNA content of the cells, as quantitative RT–PCR revealed no significant differences in LDL-receptor mRNA levels between control and ARH fibroblasts. The amount of LDL-receptor mRNA relative to β-actin mRNA in control fibroblasts transfected with the Dab2 siRNA and pre-incubated with lipoprotein-deficient serum (LPDS) for 40 h was 78 ± 26% of that in cells transfected with the non-targeting siRNA versus 91 ± 23.0% for ARH fibroblasts (mean of four different preparations of mRNA from two different control cells and five from two different ARH cells). Similar results were obtained when the amount of LDL-receptor mRNA was expressed relative to 18S rRNA (Dab2 siRNA versus control siRNA: 71 ± 18% for control cells; 83 ± 29% for ARH cells).

To investigate why LDL-receptor protein was reduced in Dab2-depleted ARH cells, LDL-receptor synthesis and turnover was determined in pulse-chase experiments with a mixture of [³⁵S]methionine and [³⁵S]cysteine. Cells treated with either the control siRNA or the Dab2 siRNA were incubated with ³⁵S-labelled amino acids for up to 2 h, and then with complete medium for up to 22 h. LDL-receptor protein was immunoprecipitated from cell lysates and analysed by SDS–PAGE to measure incorporation of radioactivity into the LDL receptor. Representative autoradiographs of gels are shown in Figure 4A for normal cells and in Figure 4B for ARH cells. As shown in Figure 4C and D, the amount of labelled LDL-receptor protein in the experiment shown was quantified by phosphorimaging and expressed relative to incorporation of ³⁵S-labelled amino acids into total trichloroacetic acid-insoluble protein at each time point. The overall rate of incorporation of radioactivity into total cell protein was not different between the control and ARH cells and was not affected by Dab2-depletion (data not shown). As seen from the gel shown in Figure 4A (compare upper and lower panels) and the quantified radioactivity (Fig. 4C, compare open and closed symbols), profound depletion of Dab2 in control cells had little discernable effect on either the rate of synthesis of mature LDL-receptor protein or its subsequent turnover during the chase period. Furthermore, both these parameters were indistinguishable from those in ARH cells treated with the control siRNA (Fig. 4B, compare upper panel with control cells in Fig. 4A, and in Fig. 4D, compare closed symbols with Fig. 4C). However, the rate of incorporation of ³⁵S-labelled amino acids into LDL-receptor protein was greatly reduced in Dab2-depleted ARH cells (Fig. 4D, compare open and closed symbols). It was not possible to determine whether the proportion of precursor converted to mature receptor was affected because the radioactivity in precursor bands was insufficient to permit accurate quantification in Dab2-depleted ARH cells and no accumulation of precursor protein was evident on autoradiographs (Fig. 4B, lower panel). Similar results were obtained with fibroblasts from two different ARH subjects and control subjects. Mean ( ± SD, n = 3) incorporation of ³⁵S-labelled amino acids into mature LDL receptor protein in Dab2-depleted ARH cells was 45.2 ± 18.9% of that in cells treated with control siRNA after 4 h chase and 33.5 ± 14.7% after 8 h chase; the same values for control cells were 91.3 ± 4.7 and 93.8 ± 2.8%. Thus, either LDL-receptor protein synthesis was reduced or newly synthesized protein was rapidly degraded.

Figure 4.

Incorporation of ³⁵S -labelled amino acids into LDL receptor protein. Control (A and C) or ARH cells (B and D) (heterozygous for two defective alleles, a large deletion of approximately exons 6–9 and insC₆₂₀) (4) treated with Dab2-02 si-RNA (Dab2) or control siRNA (Con) were pre-incubated with LPDS for 14 h and then incubated in methionine-free medium containing [³⁵S]methionine and cysteine for the indicated times (pulse); cells incubated for 2 h were then incubated with medium containing unlabelled amino acids (chase). (A and B), immunoprecipitated LDL receptor protein in cell lysates was fractionated by SDS–PAGE; horizontal lines indicate the precursor (P) and mature (M) forms. (C and D) Incorporated radioactivity in mature LDL receptor protein was measured by phosphorimaging and expressed relative to incorporation into total cell protein (closed symbols, cells treated with control siRNA; open symbols, cells treated with Dab2 siRNA). The insets show immunoblots of Dab2 and a loading control (γ-tubulin) in the same extracts (siRNA treatment indicated above the gel images).

Investigation of the effect of inhibitors of protein degradation

One conclusion from the observed results is that in fibroblasts lacking LDLRAP1, Dab2 is required to protect newly synthesized LDL-receptor protein, possibly as a chaperone. To investigate whether the reduction in apparent LDL-receptor synthesis was the result of rapid degradation of newly synthesized protein, ARH cells treated with control siRNA or Dab2-siRNA were incubated with inhibitors of protein degradation. As shown in Figure 5A, treatment of ARH cells with the proteasomal inhibitor MG132 (20 µm for 8 h) or the equivalent volume of solvent vehicle DMSO, as indicated above the gel image, had no influence on the LDL-receptor protein content of ARH cells treated with control siRNA, (left two lanes) or cells depleted of Dab2 (right two lanes). To confirm that treatment of cells with MG132 under these conditions could influence protein degradation, we examined its effect in Chinese hamster ovary (CHO) cells expressing two different mutant forms of the LDL receptor, I402T and C240F (19,20). Maturation and exit of these mutant proteins from the ER is known to be impaired, resulting in accelerated protease degradation in the ER (19,20) and thus their level in cells should be increased when cells are treated with the proteasomal inhibitor MG132. As shown in Fig. 5B, incubation of these cells with MG132 resulted in the expected increased accumulation of the mutant proteins. To determine whether other protease inhibitors had any effect on LDL receptor protein content in Dab2-depleted ARH cells, we also incubated siRNA-treated ARH cells with leupeptin (100µM for 20 h) and found no effect (Fig. 5C, compare lanes 7 and 8). Leupeptin was clearly having some effect on protein degradation in these cells because Dab2 protein levels increased slightly in both control and ARH cells treated with control siRNA (Fig. 5C, compare lanes 1 and 2; lanes 5 and 6). Changing the conditions (concentration of inhibitor or time of incubation) according to published reports (21–23) did not result in any increase in LDL-receptor protein in Dab2-depleted ARH cells, nor did incubation with other protease inhibitors, i.e. ammonium chloride, lactacystin or ALLN (data not shown).

Figure 5.

Effects of proteolysis inhibitors on LDL-receptor protein content of cells. ARH fibroblasts treated with control siRNA or Dab2 siRNA (A) or CHO cells transfected with cDNA for LDL receptor mutants I420T or C240F, as indicated (B), were incubated with MG132 (20 µm) or vehicle (DMSO) for 8 h. (C) Control (lanes 1–4) or ARH (lanes 5–8) fibroblasts treated with control siRNA (lanes 1, 2, 5 and 6) or Dab2 siRNA (lanes 3, 4, 7 and 8) were incubated with (+) or without (−) leupeptin (100 µm) for 20 h. Protein extracts from these cells were analysed by immunoblotting with antibodies to γ-tubulin (as a loading control) or to the LDL receptor as indicated. mLDLR, mature LDL-receptor protein; pLDLR, immature precursor form of LDL receptor.

Analysis of Tf receptor function

To determine whether the effect of Dab2 depletion was specific for the LDL receptor in ARH cells, or whether other endocytic receptors might be affected, the uptake of Tf via the Tf receptor was measured. Depletion of AP2 with siRNA directed to the µ2-subunit has been shown by others to reduce markedly the uptake of Tf by HeLa cells (24,25). Fig. 6 shows the effects of AP2-α and Dab2 knockdown on Tf uptake in fibroblasts. When cells treated with a non-targeting siRNA (closed circles in Fig. 6A or B) were incubated with ¹²⁵I-labelled Tf at 37°C, the amount of intracellular Tf peaked at about 10 min, after which it fell as it was recycled into the medium (26). As expected, depletion of AP2 with siRNA directed to the α-adaptin subunit strongly reduced the uptake of ¹²⁵I-Tf in both control and ARH fibroblasts to <26% of that in cells transfected with a non-targeting siRNA (Fig. 6, open triangles). Dab2 depletion had very little effect on Tf uptake in control cells (Fig. 6A, closed triangles), but in ARH cells, it was reduced to 64% of that in cells transfected with a non-targeting siRNA (Fig. 6B, closed triangles). These findings were confirmed in cells from two different control subjects and two ARH patients, using two different siRNA targeting Dab2 (Fig. 6C). Immunoblotting of Dab2-depleted ARH cell extracts showed that there was a slight reduction in cellular content of Tf receptor protein (Fig. 6D, compare left two lanes on gel image), but the reduction was much less profound than that of LDL-receptor protein under these conditions (Fig. 3B).

Figure 6.

Uptake of ¹²⁵I-Tf by human skin fibroblasts transfected with siRNA. Control (A) or ARH (B) skin fibroblasts were treated with siRNA as described in Figure 3, and 4 days later, the internalization and recycling of ¹²⁵I-labelled Tf was determined and expressed as a percentage of the total counts in the lysate. Data shown are the mean of duplicate values in three separate experiments. (C) Tf uptake after 10 min at 37°C was measured in fibroblasts from two control subjects (Con-1 and -2) and two ARH patients (ARH-1, homozygous for Q136X and ARH-2, homozygous for W22X) transfected with a control siRNA, siDab2-02, siDab2-03 or α-adaptin siRNA (siAP2-α); values were expressed as the percentage of uptake in the cells transfected with a non-targeting siRNA. (D) Extracts of control (Con) or ARH cells treated with control siRNA (Con) or siRNA targeted to Dab2 were immunoblotted with antisera to Tf receptor (TfR) or γ-tubulin, all as indicated.

DISCUSSION

In this study, we have taken advantage of gene silencing by siRNA to seek an explanation for the surprising ability of skin fibroblasts from ARH patients to internalize the LDL receptor; these cells lack the specific LDL-receptor adaptor LDLRAP1 (formerly known as ARH) that is essential for internalization of the LDL receptor in other cells types (4,9). Efficient knockdown of Dab2 (Fig. 1) severely disrupted LDL uptake in ARH fibroblasts that lack LDLRAP1 (Fig. 2), but since this was accompanied by a corresponding reduction in LDL receptor protein (Fig. 3), the role of Dab2 in clathrin-mediated endocytosis of the LDL receptor in skin fibroblasts remains unresolved. Instead we propose a novel function for Dab2 in modulating LDL receptor protein synthesis in cells lacking LDLRAP1.

Two siRNA targeted to different regions of Dab2 were used in this study; both were effective and specific (Fig. 1C and D). In two different ARH fibroblast cell lines, we found that LDL-receptor protein was markedly reduced after depletion of Dab2 with either of the siRNA, yet while transfection of the same siRNA into control fibroblast cell lines dramatically reduced Dab2 protein, LDL-receptor expression was unaffected (Fig. 3B). Hence, it seemed unlikely that the reduction of LDL receptor protein in Dab2-depleted ARH cells was due to off-target effects of the siRNA. We confirmed this by our observation that over-expression of a murine Dab2 construct not targeted by the siRNA reversed the loss of LDL receptor protein in these cells. In addition, we found by quantitative real-time RT–PCR that there was no associated change in LDLR mRNA in Dab2-depleted ARH cells; the reduction in LDL receptor protein is therefore likely to be the result of increased degradation or reduced translation rather than decreased transcription. Interestingly, we also observed a slight reduction in Tf receptor expression in ARH cells depleted of Dab2 (Fig. 6); it is worth noting, however, that other proteins involved in trafficking were unaffected, for example, expression of two subunits of AP2 was unchanged by depletion of Dab2 (Fig. 1C and D). During their studies to investigate a potential role for Dab2 as an endocytic adaptor in HeLa cells, Maurer and Cooper (17) and Keyel et al. (18) also depleted Dab2 from ARH fibroblasts and concluded that LDL-receptor internalization was impaired as a result. However, they did not examine LDL-receptor protein levels in any detail and, from the results shown, it is difficult to determine whether LDL-receptor protein levels were decreased in their Dab2-depleted cells.

The LDL receptor is synthesized in the endoplasmic reticulum (ER) as a precursor with an apparent molecular mass of 120 KDa. Within 30 min, it is transported to the Golgi complex where post-translational modifications result in the production of mature protein with increased apparent molecular mass (160 KDa) (27). In our study, no accumulation of precursor protein was evident in the Dab2-depleted ARH cells (Fig. 4), which might have been explained by its rapid degradation in the ER due to lack of protection by Dab2. Such a protective role has been proposed by others for Dab1, the neuronal counterpart and structural homologue of Dab2, which has been reported to chaperone the very low density lipoprotein (VLDL) receptor and apolipoprotein E receptor 2 (ApoER2) to the cell surface (28). It is possible that in addition to their roles as endocytic adaptors, Dab2 and LDLRAP1 may also be required for forward trafficking of LDL receptor. However, we found that treatment of the Dab2-depleted ARH cells with several different inhibitors of protein degradation had no effect on their LDL-receptor protein content (Fig. 5A and C), suggesting that rapid degradation of newly synthesized protein was also unlikely. In marked contrast, we showed that these inhibitors were able to inhibit the degradation of mutant forms of the LDL receptor that become trapped in the ER (Fig. 5B). Combining these results with the very low amounts of newly synthesized LDL-receptor protein detected within 30 min (Fig. 4) strongly suggests that reduced translation is the likely cause for the diminished LDL-receptor we observed in Dab2-depleted ARH cells.

Clearly, our original intention to determine whether or not Dab2 facilitates LDL-receptor internalization in ARH fibroblasts was not possible from our data. However, recently published observations by others on the effect of depletion of LDLRAP1 and Dab2 in HeLa cells certainly support this proposal. In these cells, simultaneous depletion of Dab2 and LDLRAP1 with siRNA resulted in decreased uptake of fluorescently labelled LDL and accumulation of LDL-receptor protein on the cell surface (17,18). Our data from confocal microscopy of immunolabelled cells showed no evidence for accumulation of LDL receptor protein on the surface of fibroblasts lacking Dab2 and LDLRAP1 (Fig. 3C), but the HeLa cell data are reminiscent of our earlier finding that LDL-receptor protein accumulates on the surface of EBV-transformed lymphocytes from ARH patients (4), which, as we have shown here, do not express detectable Dab2 (Fig. 1). Quite why LDL-receptor protein is stable but trapped on the cell surface in the absence of ARH and Dab2 in lymphocytes and HeLa cells, but is unstable in ARH skin fibroblasts depleted of Dab2 is puzzling. Clearly, LDL-receptor protein trafficking differs in different cell types, and some component that is absent from skin fibroblasts must maintain LDL receptor protein integrity in HeLa cells and lymphocytes depleted of Dab2 and LDLRAP1. Presumably, this component is also present in the liver, since Dab2 expression in whole liver has been reported to be very low (12) and has recently been shown to be absent from hepatocytes (18), and in arh−/− mice lacking LDLRAP1, LDL-receptor protein also accumulates on the surface of hepatocytes in vivo (5).

It has been well documented that AP2 plays a major role in the assembly of clathrin-coated vesicles and is necessary for the uptake of cargo proteins via interaction of the YXXΦ internalization signal in the cytoplasmic tail of the cargo with the µ2 subunit of AP2. Indeed siRNA targeting of µ2 has been shown by others to prevent uptake of ¹²⁵I-Tf in HeLa cells (24,25). However, there is a body of evidence to suggest that although uptake of both Tf and LDL receptors is dependent on clathrin-mediated endocytosis, the internalization of these cargos involves distinct mechanisms (3,24,30,31). We transfected control fibroblasts with siRNA targeted to the α-subunit of AP2 and found that LDL uptake was not impaired (Fig. 2). Although the knockdown of AP2 achieved by the α-adaptin siRNA was incomplete (Fig. 1), it was sufficient to reduce Tf uptake in the same cells to a negligible level (Fig. 6). Others have observed that mutations in the AP2 binding site of LDLRAP1 did not prevent LDLRAP1-mediated clustering of LDL receptor in clathrin-coated pits unless the binding sites for clathrin were also disrupted (32). Our experiments are consistent with this, and indicate that LDLRAP1 is sufficient to promote endocytosis of the LDL receptor in fibroblasts in the absence of AP2. However, when we depleted the α-subunit of AP2 in ARH fibroblasts which lack LDLRAP1, LDL uptake was slightly impaired while LDL-receptor protein levels were unchanged (Figs 2 and 3A). This suggests that in the absence of LDLRAP1, AP2 may play some part in LDL-receptor internalization. This highlights another difference between LDL-receptor function in skin fibroblasts and HeLa cells, since others have found that depletion of LDLRAP1 and AP2 apparently had no measurable effect on LDL-receptor internalization in these cells as long as Dab2 was present (17,18). Surprisingly, we found that although Tf uptake in control cells was not affected by depletion of Dab2, it was slightly less efficient in Dab2-depleted, LDLRAP1-deficient cells, probably due to a small reduction in Tf receptor protein in the cells (Fig. 6).

In summary, we have demonstrated that Dab2 depletion results in the reduction of LDLR protein in ARH fibroblasts but not in normal fibroblasts. We propose that functional redundancy may exist between LDLRAP1 and Dab2 in modulating LDL receptor synthesis. Since the reduction in LDL-receptor protein is not caused by reduced transcription and cannot be rescued by proteasomal or lysosomal inhibitors under the conditions we have tried, it is likely that it is caused by reduced translation of LDL receptor mRNA. Of course, we cannot exclude the possibility that proteolytic pathways were induced that are unaffected by any of the classes of protease inhibitors we employed. There are a number of possible mechanisms, including the possibility that the absence of both LDLRAP1 and Dab2 somehow results in reduced binding of LDL receptor mRNA to polysomes or in decreased initiation or elongation of the polypeptide chain or even in blockage of translocation due to the presence of some repressors or absence of some necessary factors. Extensive studies will be required to address the complex nature of this process.

MATERIALS AND METHODS

Cell culture

All cell culture reagents were purchased from Invitrogen. Cells were cultured at 37°C with 5% CO₂. Skin fibroblasts were grown in DMEM supplemented with GlutaMAX, antibiotics and 10% (v/v) fetal calf serum (growth medium). EBV-lymphocytes were maintained in RPMI supplemented with 10% (v/v) heat-inactivated fetal calf serum and antibiotics [1% (v/v) penicillin/streptomycin solution (Invitrogen)]. LPDS was obtained from Sigma. CHO cells lacking endogenous LDL receptor activity (CHO ldlA7 cells), kindly provided by Dr M Krieger (33), were maintained in Ham's F12 medium with 10% (v/v) fetal calf serum and antibiotics [1% (v/v) penicillin/streptomycin solution (Invitrogen)].

Measurement of cell protein and mRNA content

Immunoblotting of cell extracts was performed as described previously (34), with the following antibodies: monoclonal anti-Dab2 (p96) (Transduction Laboratories, diluted 1 in 2000); anti-γ-tubulin (Sigma, diluted 1 in 10 000); anti-a-adaptin (Santa Cruz Biotechnology, diluted 1 in 2000); anti-µ2-adaptin (AP50, BD Transduction Laboratories, diluted 1 in 250); rabbit polyclonal anti-LDL receptor (Progen Biotechnik, diluted 1 in 3000); mouse monoclonal anti-Tf receptor (Zymed Laboratories Inc., diluted 1 in 1000). Affinity-purified rabbit polyclonal anti-ARH was generously provided by Linton Traub (7). Peroxidase-conjugated secondary antibodies were obtained from DAKO, Denmark.

LDL-receptor mRNA was assayed by real-time RT–PCR as described previously (4). In brief, total RNA was isolated using RNAzol kit (Biogenesis Poole, Dorset, UK). The probe and primers for assay of LDL receptor mRNA were: 5′FAM-CCCAACCTGAGGAACCTGGTCGCT-TAMRA-3′ (probe); 5′-CTGGACCGGAGCGAGTACAC-3′ and 5′-TGGGTGCTGCAGATCATTCTC-3′ (primers); probes and primers for β-actin and 18S rRNA were obtained from PE Applied Biosystems. LDLR mRNA was normalized to either β-actin or 18S rRNA assayed on the same plate. All samples and standards were assayed in triplicate.

Transfection of cells

All siRNA oligonucleotides were obtained from Dharmacon. The target sequences for SMARTselection™ siDab2-02 was AAGAACCAGCCUUCACCCUUU and for siDab2-03 was AACAAAGGAUCUGGGUCAACA, and the non-targeting control siRNA was siCONTROL Non-Targeting siRNA #1. Transfection efficiency was optimized using siGLO (RISC-Free siRNA).

Fibroblasts at 60–70% confluence were transfected on day 1 with OligofectAMINE™ (Invitrogen) and 50 nm siRNA, using the protocol recommended by the supplier (Dharmacon) and re-fed with fresh medium the following day. On day 3, cells were trypsinized and seeded at ∼70% confluence into dishes or on to cover slips for microscopy. For measurement of LDL-receptor protein or mRNA, or uptake or degradation of ¹²⁵I-labelled LDL, cells were then incubated for 16–18 h in medium supplemented with 10% (v/v) LPDS. For each experiment, cells from one well of a six-well plate were assayed for knockdown efficiency by immunoblotting of whole cell extracts with appropriate antibodies. For plasmids, cells were transfected with lipofectamine 2000™ as recommended by the supplier (Invitrogen).

CHO ldlA7 cells that lack endogenous LDL-receptor activity were transfected with LDL-receptor cDNA by electroporation as described previously (19,20).

Internalization assays

Uptake and degradation of ¹²⁵I-labelled LDL was determined as described (34). For the assay of Tf receptor function, Tf was labelled with ¹²⁵I (3) and uptake assays were performed, all essentially as described previously (24). Briefly, siRNA-transfected cells (day 5) in 12-well plates were incubated on a rocker for 30 min at 4°C in 0.5 ml pre-chilled serum-free medium (DMEM supplemented with 1% w/v bovine serum albumin) containing ¹²⁵I-labelled Tf (2 µg of protein/ml). After five washes in cold serum-free medium, cells were incubated at 37°C in 1 ml growth medium. At various time points, the medium was harvested, surface-bound ligand was removed with an acid wash and cells were lysed in 0.1 m NaOH. Radioactivity in the medium (recycled ¹²⁵I-labelled Tf), acid wash (surface-bound) and the cell lysate (intracellular) was quantified and the intracellular component was expressed as a percentage of total counts recovered at each time point.

Incorporation of ³⁵S-labelled amino acids

Skin fibroblasts at 60% confluence in 6 cm diameter dishes were treated on day 1 with 50 nM siRNA (either siDab2-02 or siCONTROL) as described above. On day 3, the cells were incubated for 14 h with medium containing 10% (v/v) LPDS, then washed and incubated at 37°C with methionine-free medium containing 375 µCi of [³⁵S]methionine and [³⁵S]cysteine (Redivue Pro-mix L-[³⁵S] in vitro cell labelling mix; GE Healthcare, Bucks, UK) for up to 2 h; for chase experiments, the medium was replaced after 2 h with medium containing unlabelled amino acids. LDL-receptor protein was immunoprecipitated from cell lysates with anti-LDL receptor IgGC7 as described previously (29,35). Immunoprecipitates were fractionated by SDS-PAGE on reduced gels; radioactivity in LDL receptor was quantified by Phosphorimaging and expressed relative to that in total cellular protein (arbitrary units), determined as incorporation into total trichloroacetic acid-insoluble protein in cell lysates by standard techniques (36).

Confocal microscopy

Cells treated with siRNA and/or transfected with mDab2 as described above were plated on to cover slips, incubated for 16 h in medium containing 10% LPDS and then fixed in 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and stained with appropriate antibodies for confocal microscopy as described previously (4). The following antisera were used: mouse monoclonal anti-Dab2 (p96) (Transduction Laboratories, diluted 1 in 40); rabbit polyclonal anti-LDL receptor (Progen Biotechnik, diluted 1 in 40); Alexa Fluor 568- or 488-conjugated goat anti-mouse or anti-rabbit IgG (Invitrogen, diluted 1 in 100).

Conflict of Interest statement. The authors have no conflicts of interest.

FUNDING

This work was supported by the UK Medical Research Council and by a project grant from the British Heart Foundation (PG/03/020//15126). Funding for open access charge: UK Medical Research Council.

Author notes

‡

The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.

†

Present address: Cell Biology, Institute of Ophthalmology, 11-43 Bath Street, London EC1V 9EL, UK.