Apolipoprotein A-I Activates Cellular cAMP Signaling through the ABCA1 Transporter*

It has been suggested that the signal transduction pathway initiated by apoA-I activates key proteins involved in cellular lipid efflux. We investigated apoA-I- mediated cAMP signaling in cultured human fibroblasts induced with (22 R )-hydroxycholesterol and 9- cis- reti- noic acid (stimulated cells). Treatment of stimulated fibroblasts with apoA-I for short periods of time ( < 45 min) increased ATP binding cassette A1 (ABCA1) phosphorylation in a concentration-dependent manner. Concomi- tantly, apoA-I increased the intracellular level of cAMP in a concentration- and time-dependent manner. The maximal cAMP level was reached within 10 min at 10 (cid:1) g/ml apoA-I representing a 1-fold increase. The ability of apoA-I to mediate cAMP production was only observed in stimulated fibroblasts. Furthermore, overex- pression of ABCA1 in Chinese hamster ovary cells resulted in a 1.5-fold increase in apoA-I-mediated cAMP accumulation as compared with untransfected cells. In contrast, forskolin increased cAMP production signifi- cantly in unstimulated fibroblasts as well as in untransfected Chinese hamster ovary cells. Pharmacological in- hibition of protein kinase A (H89) completely blocked apoA-I-mediated ABCA1 phosphorylation. Naturally oc- curring mutations of ABCA1 associated with Tangier disease (C1477R, 2203X, and 2145X) severely reduced apoA-I-mediated cAMP production, ABCA1

plays a crucial role in the biogenesis of HDL particles (1)(2)(3). The importance of ABCA1 in the reverse cholesterol transport process has been strikingly demonstrated by the identification of mutations in ABCA1 gene locus as the molecular defect of Tangier disease (TD) and familial HDL deficiency (4,5). These patients are characterized by extremely low HDL levels caused by inadequate transport of cellular cholesterol and phospholipids to the extracellular space, leading to hypercatabolism of lipid-poor nascent HDL particles (6). Thus, factors affecting the active lipidation of apoA-I are likely to affect the homeostasis of plasma HDL cholesterol and the reverse cholesterol transport process, one of the major mechanisms by which HDL may prevent atherosclerotic vascular disease (7).
We have documented previously (8) that the cAMP/protein kinase A (PKA)-dependent pathway plays a pivotal role in ABCA1 phosphorylation and modulates cellular lipid efflux mediated by apoA-I in fibroblasts. In a related observation, See et al. (9) reported that S2054 on the ABCA1 protein is essential for PKA phosphorylation and for regulation of ABCA1 transporter activity. In addition, it has been shown that PKA or protein kinase C (PKC) controls the function of ABC transporters, cystic fibrosis transmembrane regulator, and P-glycoprotein by their phosphorylation (10,11). A number of kinases have been implicated in cellular lipid efflux including mitogenactivated protein kinase and PKC (12,13). PKC inhibitors severely reduce apolipoprotein-mediated cholesterol efflux (14), and at the same time, protein kinase agonists such as 1,2dioctanoylglycerol and phorbol 12-myristate 13-acetate increase cholesterol efflux (15). On the other hand, it was reported that cAMP-mediated cholesterol efflux to apoA-I is associated with the binding, uptake, and resecretion of apoA-I in a calcium-dependent pathway in murine macrophages (16). Such cross-talk between cAMP signaling and Ca 2ϩ -and phospholipid-dependent signaling pathways might play a key functional role in vivo in the activation of ABCA1 by its phosphorylation, allowing apoA-I lipidation.
Although it is clear that the cellular signaling process plays an important role in cellular lipid homeostasis and lipid efflux, structural determinants of molecular interactions between cell signaling molecules and apoA-I have not been elucidated. We hypothesize that apoA-I activates cAMP signaling directly or via cross-talk mechanisms with other signal transduction pathways. The present study aims at providing evidence for links between apoA-I, intracellular cAMP production, and ABCA1 phosphorylation, and examining how these interactions could be affected by cAMP regulators, PKA inhibitors, or by naturally occurring mutants of ABCA1.

EXPERIMENTAL PROCEDURES
Patient Selection-Patients with TD were selected as described previously (2). All of these subjects carry a mutation at the ABCA1 gene locus and had clinical signs of TD. Cellular studies in all TD subjects revealed a marked impairment of apoA-I-mediated phospholipid and cholesterol efflux. Molecular analysis of the ABCA1 gene revealed a compound heterozygous state for subjects TD-1 and -3 and a homozygous nonsense point mutation in subject TD-2 (see Table I). Fibroblasts from three normolipidemic subjects were used as controls.
ABCA1-expressing CHO Cells-Green fluorescent protein-ABCA1expressing CHO cells were generously provided by Dr. Sean Davidson from the Department of Pathology and Laboratory Medicine, University of Cincinnati, and were characterized and cultured as described previously (17).
ABCA1 Mutational Analysis-The analysis of the mutations at the ABCA1 gene was performed as described previously (4,5). Exon-specific oligonucleotides were synthesized for each exon of the ABCA1 gene. Direct sequencing was performed on all patients, and single nucleotide changes were assessed by comparison between species to identify highly conserved residues. All detected nucleotide changes were assessed in large DNA samples that were from control individuals with normal HDL-C and from the same origin.
In Vivo Phosphorylation and Immunoprecipitation-Confluent cells were stimulated or not with 2.5 g/ml (22R)-hydroxycholesterol and 5 M 9-cis-retinoic acid for 20 h, and the medium was replaced by minimum Eagle's medium-free phosphate solution (Invitrogen). Cells were then labeled with 0.5 mCi/ml [ 32 P]orthophosphate (PerkinElmer Life Sciences) for 45 min at 37°C in the presence or absence of apoA-I (Biodesign International), 8-Br-cAMP, or okadaic acid. At the indicated time, the cells were washed twice with ice-cold phosphate-buffered saline and scraped into 0.5 ml of immunoprecipitation buffer containing 20 mM Tris (pH 7.5), 0.5 mM EDTA, 0.5 mM EGTA, 0.1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 25 g/ml okadaic acid, and 25 mM glycerophosphate, and the suspension was allowed to stand for 30 min at 4°C in the presence of a protease inhibitor mixture (Roche Diagnostics). [ 32 P]ABCA1 was immunoprecipitated by anti-ABCA1 antibody (Novus Biologicals), separated on SDS-gel, and detected by phosphorimaging as described previously (8).
Cellular ABCA1 Content and Cell Surface Biotinylation-This was performed as described previously (17) with slight modification. Confluent cells were stimulated with 2.5 g/ml (22R)-hydroxycholesterol and 5 M 9-cis-retinoic acid for 20 h, and then surface proteins were biotinylated with 500 g/ml sulfosuccinimidyl 2-(biotinamido)-ethyl-1,3-dithiopropionate (Pierce) for 30 min at 4°C on a platform rotator. Cells were then washed with ice-cold quench buffer (1 M Tris-HCl (pH 7.5)) and once with ice-cold phosphate-buffered saline. The cells were swollen on ice for 10 min and then homogenized with 20 strokes in a tight fitting Dounce homogenizer. After centrifugation at 2000 ϫ g, at 4°C, for 10 min to remove unbroken cells and nuclei, the supernatant was recentrifuged at 100,000 ϫ g, at 4°C, for 60 min. The resulting supernatant was discarded, and the final membrane pellet was resuspended in 250 l of immunoprecipitation buffer; 100 g of protein was added to 30 l of streptavidin-Sepharose (Amersham Biosciences) and incubated overnight on a platform mixer at 4°C. The gel was pelleted and washed five times with immunoprecipitation buffer. SDS-PAGE gel and Western blotting were performed as described above.
Intracellular cAMP Assay-Twenty-five thousand cells were seeded in 24-well plates (fibroblasts, ABCA1-expressing CHO cells, and untransfected CHO cells). The next day the cells were incubated overnight in a serum-free medium (Dulbecco's modified Eagle's medium/bovine serum albumin, 1 mg/ml) supplemented with 2 Ci/ml [ 3 H]adenine (PerkinElmer Life Sciences) in the presence or absence of (22R)-hydroxycholesterol and 9-cis-retinoic acid. Cells were then washed once with Dulbecco's modified Eagle's medium/HEPES (20 mM) and incubated at 37°C in the presence of 1 mM isobutylmethylxanthine with different concentrations of apoA-I for the indicated times. Adenylyl cyclase was activated with 40 M forskolin. After the removal of the medium, the reaction was stopped with 500 l of 5% trichloroacetic acid, and the cells were mechanically scraped and layered on the top of WA-1 acid Alumina columns (Sigma). The separation of [ 3 H]cAMP so formed was achieved according to the method already described (18) and slightly modified (19). In brief, the columns were washed sequentially with 4 ml of 0.5 mM HCl and 0.5 ml of 0.1 M ammonium acetate and then collected into scintillation vials with 4 ml of 0.1 M ammonium acetate. The intracellular cAMP level was expressed in fmol/mg of cell protein. All of the experiments were performed in triplicate.
ApoA-I Binding Assay-ApoA-I binding assay was performed as we have described previously (20) with minor modification. Briefly, purified human plasma apoA-I was iodinated with 125 I by IODO-GEN (Pierce) to a specific activity of 800 -1200 cpm/ng of apoA-I. Cells from three normal controls and three TD subjects (see Table I) were grown on 24-well plates and stimulated as described above. Cells were then incubated at 37°C with 10 g/ml 125 I-apoA-I (under saturating binding conditions) in Dulbecco's modified Eagle's medium/bovine serum albumin (1 mg/ml) for 2 h at 37°C and then washed twice with cold phosphate-buffered saline/ bovine serum albumin and twice with phosphate-buffered saline. Cell association radioactivity and cell protein were measured after digestion in 0.1 N NaOH. Nonspecific binding was determined in the presence of a 40-fold excess of unlabeled apoA-I. Results are expressed as nanograms of specific 125 I-apoA-I binding/mg of cell protein. For appropriate comparison, results were presented as percent of controls (see Table I).
Cellular Lipid Efflux-Phospholipid and cholesterol efflux were determined as described previously (2,21) with minor modifications. Briefly, 50,000 cells were seeded in 12-well plates. At midconfluence, the cells were labeled with 1-5 Ci/ml [ 3 H]choline or 0.2 Ci/ml [ 3 H]cholesterol (PerkinElmer Life Sciences) for 48 h. At confluence, cells were cholesterol-loaded (20 g/ml) for 24 h. After a 24-h equilibration period, cells were stimulated with 2.5 g/ml (22R)-hydroxycholesterol and 5 M 9-cis-retinoic acid for 20 h. Phospholipid or cholesterol efflux was determined at either 2 or 24 h with 10 g/ml apoA-I. Cellular lipid efflux was determined as follows: 3 H cpm in medium/( 3 H cpm in medium ϩ 3 H cpm in cells); the results were expressed as the percentage of the total radiolabeled phospholipids or cholesterol.
Statistical Analysis-Intracellular cAMP levels have been compared statistically by Student's t test. Two-tailed p Ͻ 0.05 values were considered significantly different.

RESULTS
In the present study, we labeled the cells with [ 32 P]orthophosphate for 45 min at 37°C to better assess a small variation of ABCA1 phosphorylation (8). As shown in Fig. 1A, ABCA1 is phosphorylated at a basal level in untreated normal intact cultured human fibroblasts in which ABCA1 was induced with (22R)-hydroxycholesterol and 9-cis-retinoic acid (stimulated cells). Inclusion of the Ser-Thr phosphatase inhibitor okadaic acid (10 M) during 32 PO 4 labeling resulted in a significant increase in the level of ABCA1 phosphorylation. In addition, the treatment of cells with 8-Br-cAMP (1 mM) in the presence of okadaic acid during the 32 PO 4 labeling further increased ABCA1 phosphorylation levels up to 6 -8-fold.
To establish the relationship between apoA-I and ABCA1 phosphorylation activity, normal intact stimulated cells were treated with increasing amounts of apoA-I (0, 10, and 50 g/ml) during 32 PO 4 labeling. Phosphorylated ABCA1 was immunoprecipitated and normalized to total and cell surface ABCA1 protein as described under "Experimental Procedures." As shown in Fig. 1B (lower panel), treatment of stimulated fibroblasts with apoA-I increased ABCA1 phosphorylation in a concentration-dependent manner, compared with the basal phosphorylation level of ABCA1 in untreated cells. One representative experiment is shown in the upper panel of Fig.  1B. On the other hand, apoA-I treatment (45 min) did not increase significantly the total ABCA1 level or cell surface ABCA1 protein. The effect of apoA-I treatment (10 g/ml, 45 min) on ABCA1 phosphorylation was also observed in unstimulated cells, and this effect was markedly enhanced in stimulated cells (100 Ϯ 7.8, 187 Ϯ 9.6, and 365 Ϯ 12.6% for base-line, unstimulated, and stimulated cells, respectively).
Because cAMP has been suggested as the mediator of apoA-I-mediated ABCA1 phosphorylation (8,9), the question was raised whether the increase in ABCA1 phosphorylation induced by apoA-I was associated with an increase in cAMP production. As shown in Fig. 2A, treatment of stimulated fibroblasts with apoA-I (10 g/ml) resulted in a rapid increase of the intracellular levels of cAMP in a time-dependent manner, reaching maximal levels within 10 min, and then decreased slightly at 30 min. At the same time, treatment of stimulated normal cells with increasing amounts of apoA-I (0, 1, 5, 10, and 50 g/ml) increased intracellular cAMP levels in a concentrationdependent manner (Fig. 2B). The maximal cAMP level was reached at 10 g/ml apoA-I, representing a 1-fold increase, compared with the basal intracellular cAMP level in untreated cells (506 Ϯ 15 versus 1119.9 Ϯ 66.2 fmol/mg of cell protein, respectively). Above 10 g/ml apoA-I, the intracellular cAMP levels decreased gradually. In contrast, treatment of unstimulated normal fibroblasts with increasing amounts of apoA-I did not elevate cAMP concentrations above basal levels. A control experiment was performed to probe the specificity of apoA-I response. We demonstrate that apoA-I (10 g/ml) was able to mediate cAMP production in stimulated but not in unstimulated normal fibroblasts from three different normal cell lines (Fig. 3A). As expected, forskolin (FRK) increased intracellular cAMP similarly in either unstimulated or stimulated cells (Fig.  3A). We observed that stimulation of cells with (22R)-hydroxycholesterol and 9-cis-retinoic acid for 16 -24 h increased ABCA1 protein levels and specific 125 I-apoA-I binding by 4-fold, compared with unstimulated cells (data not shown).
Having determined that the ability of apoA-I to mediate cAMP production was observed only in stimulated fibroblasts, the question was posed whether ABCA1 expression is sufficient to increase intracellular cAMP. We next examined the ability of ABCA1 expression to induce apoA-I-mediated cAMP production in CHO cells. As shown in Fig. 3B, CHO cells stably transfected with a human ABCA1 showed a significant increase in cAMP production in the presence of 10 g/ml apoA-I, as compared with wild type CHO cells. On the other hand, FRK  32 P-Labeled ABCA1 was immunoprecipitated and separated by electrophoresis and then transferred to a polyvinylidene difluoride membrane as described under "Experimental Procedures." 32 P-Labeled ABCA1 separated in duplicate for each treatment was revealed by phosphorimaging. The ABCA1 protein was detected on the same membrane by anti-ABCA1 antibody. B, stimulated fibroblasts were [ 32 P]orthophosphate-labeled (as in A) in the presence of increasing amounts of apoA-I (0, 10, and 50 g/ml) for 45 min at 37°C. 32 P-Labeled ABCA1 was immunoprecipitated and quantified by phosphorimaging. Phosphorylated ABCA1 was normalized to total and cell surface ABCA1 protein as described under "Experimental Procedures" (lower panel). The results are representative of three independent experiments. The error bars represent the S.D. significantly increased cAMP production above basal levels in ABCA1-CHO cells as well as in wild type CHO cells.
To provide further evidence for a specific role of PKA in apoA-I-mediated ABCA1 phosphorylation activity, we examined the effect of H89 PKA inhibitor on ABCA1 phosphorylation. Treatment of normal stimulated cells with increasing amounts of H89 (0, 1, 5, and 10 M) in the presence of apoA-I (10 g/ml) almost completely inhibited apoA-I-mediated ABCA1 phosphorylation in a dose-dependent manner (Fig. 4).
To determine whether naturally occurring mutations of ABCA1 may affect the ability of apoA-I to mediate cAMP production and ABCA1 phosphorylation activity, stimulated fibroblasts from normal control and three ABCA1 mutants associated with Tangier disease (Table I) were incubated in the absence or presence of apoA-I (10 g/ml) or FRK (40 M) for 15 min, and the intracellular concentration of cAMP was determined. In parallel experiments, stimulated cells from normal control and TD cell lines were labeled with [ 32 P]orthophosphate in the absence or presence (10 g/ml) of apoA-I. The percent increase in ABCA1 phosphorylation above basal levels in untreated cells was then determined. As shown in Table I, intracellular cAMP levels and ABCA1 phosphorylation were increased significantly in response to apoA-I treatment in normal cells, but apoA-I failed to mediate either intracellular cAMP production, ABCA1 phosphorylation, or cellular lipid efflux in TD cell lines. As expected, FRK-mediated cAMP production was not significantly different in normal or TD cell lines (Table I).
We next investigated whether the absence of cAMP response was due to defective apoA-I binding to ABCA1 mutants. Specific 125 I-apoA-I cell association was determined in three normal and three ABCA1 mutant cell lines, as described under "Experimental Procedures." As shown in Table I, specific 125 I-apoA-I was found severely reduced in TD cell lines as compared with normal cells.
To determine whether the absence of ABCA1 phosphorylation in response to apoA-I stimulation in TD cells was because of abnormal PKA phosphorylation sites, in vitro experiments were carried out in which immunoprecipitated ABCA1 from normal and TD-stimulated cells were phosphorylated by the catalytic subunit of PKA. As shown in Fig. 5A, the PKA catalytic subunit was able to similarly phosphorylate ABCA1 from normal and TD cells.
TD mutations resulted in a distinct influence on the function and subcellular localization of ABCA1. To examine whether defective apoA-I binding, cAMP production, or ABCA1 phosphorylation observed in ABCA1 mutants was because of abnormal localization of ABCA1 at the plasma membrane, cell surface proteins from stimulated normal and TD cells were biotinylated as described under "Experimental Procedures," and the isolated cell surface proteins and the total cellular ABCA1 protein were then analyzed by Western blot probed with a polyclonal antibody to ABCA1. Fig. 5 shows that all three ABCA1 mutants were similarly and efficiently trans- lated, but they differed markedly in their ability to reach the plasma membrane. As shown in Fig. 5B, the 2203X (TD-2) mutant was localized to the plasma membrane as much as the normal control. The C1477R (TD-1) mutant showed decreased presence at the plasma membrane, relative to the normal transporter (50% or greater reduction); however, the 2154X (TD-3) mutant failed to reach the membrane. DISCUSSION It is well accepted that the removal of cholesterol from peripheral cells by HDL involves signal transduction systems in which HDL receptors at the cell surface are believed to transmit the signal to intracellular effectors (22,23). In this report, we demonstrate that the treatment of stimulated normal fibroblasts with apoA-I for short periods of time (Յ45 min) induces ABCA1 phosphorylation activity in a concentration-dependent manner (Fig. 1B). However, a recent study by Tall and coworkers (24) documented that apoA-I treatment for longer periods of time (Ն1 h) results in the dephosphorylation of the ABCA1 PEST sequence and thereby inhibits calpain degradation leading to an increase of both ABCA1 cell surface expression and activity. It is possible that apoA-I is increasing phosphorylation because it is stabilizing ABCA1 protein at plasma membrane sites accessible to PKA. In the present study, however, we did not observe evidence in support of such a mechanism under our experimental phosphorylation system. First, treatment of fibroblasts with apoA-I (Յ45 min) affected neither the total ABCA1 level nor the cell surface ABCA1 (Fig. 1B). Second, ABCA1 phosphorylation was almost completely inhib-ited with the H89 PKA inhibitor, which did not affect the level of total ABCA1 protein (Fig. 4). In addition, we have been able to demonstrate in our phosphorylation system that phosphatase 1 and/or phosphatase 2A play a role in dephosphorylation of ABCA1 (Fig. 1A). It is likely that apoA-I signaling activity is an upstream step(s) of the calpain-mediated proteolysis. Concomitantly, apoA-I increased intracellular cAMP levels in a concentration-and time-dependent manner in stimulated normal fibroblasts (Fig. 2), consistent with the concept that the cAMP/PKA-dependent pathway may regulate ABCA1 activity and consequently cellular lipid efflux (8,9). This is supported by a previous study (25) demonstrating that anion flux mediated by ABCA1 expression in Xenopus oocytes can be stimulated by cAMP or inhibited by the PKA inhibitor, H89. Furthermore, it was documented that PKA phosphorylation of ABCA1 S2054 is important for maintaining normal phospholipid efflux function (9).
Although the molecular mechanism of apoA-I-mediated cAMP signaling has not yet been elucidated, the present study shows that at a constant level of ABCA1 transporter expression in unstimulated cells, increasing apoA-I concentration did not elevate cAMP concentrations above basal levels (Fig. 2B). At the same time, increasing ABCA1 expression by (22R)-hydroxycholesterol and 9-cis-retinoic acid (4-fold) caused a significant increase in the intracellular cAMP concentration (80 -100%) in response to apoA-I treatment (Figs. 2B and 3A). We postulate that the amount of ABCA1 transporter in unstimulated fibroblasts is below a threshold level needed to trigger cAMP production or, alternatively, (22R)-hydroxycholesterol/ 9-cis-retinoic acid stimulation induces other proteins or pathways that potentiate the ability of apoA-I to promote cAMP production and ABCA1 phosphorylation. Our current results support the first hypothesis. We demonstrate that overexpression of ABCA1 in CHO cells significantly increased apoA-Imediated cAMP accumulation as compared with untransfected cells (Fig. 3B). At the same time, FRK significantly increased cAMP production in both transfected and untransfected CHO cells. Thus, ABCA1 expression is sufficient to increase apoA-Imediated intracellular cAMP production.
It is well documented that a wide variety of neurotransmitters, peptide and protein hormones, chemokines, growth factors, and other ligands elicit specific cellular responses by binding to plasma membrane receptors that are coupled to one or more heterotrimeric guanidine nucleotide binding regulatory proteins (G protein). Once activated, G protein-coupled receptors couple to and activate specific G s protein isoforms that promote the production of intracellular second messengers such as cAMP (26,27), supporting the concept that the amount of apoA-I-dependent cAMP accumulation in normal stimulated fibroblasts may reflect the level of adenylyl cyclase (AC) activation and thereby indicate the efficiency of ABCA1 transporter to couple to G s protein ␣ subunits (G s ), known as a physiological activator of AC and consequently intracellular cAMP production. Our results suggest that apoA-I activates cAMP signaling through an ABCA1 transporter coupled to the G s protein. This concept of ABCA1-mediated signaling was supported by the identification of a functional association between ABCA1 and the fas-associated death domain protein, an adaptor molecule mainly described in death receptor signal transduction (28). Furthermore, Cdc42 was reported to interact directly with ABCA1, and this interaction was suggested to influence intracellular lipid trafficking and apoA-I-mediated cholesterol efflux (29,30).
The proposed mechanism of ABCA1-mediated cAMP signaling was further strengthened by our results demonstrating that either apoA-I-mediated cAMP production or ABCA1 phos- phorylation was severely impaired in stimulated fibroblasts from TD subjects (Table I). Thus, we provide further evidence for a functional linkage between the ABCA1 transporter and the cAMP pathway. We postulate that functional ABCA1 is required for selective coupling to G␣ s leading to cAMP production or, alternatively, the structural characteristics of apoA-I binding to the ABCA1 transporter can directly affect cAMP signaling. In the present study, evidence was in fact obtained to support both of these possibilities. First, specific binding of 125 I-apoA-I to the ABCA1 transporter was severely reduced in TD cells, suggesting that ABCA1 mutants that failed to bind apoA-I also failed to mediate an increase in cellular cAMP levels, PKA-mediated ABCA1 phosphorylation, or cellular lipid efflux (Table I). This conclusion is consistent with the finding that the capacity of FRK to stimulate AC was unimpaired in TD cells (Table I). Although the three ABCA1 mutants were similarly and efficiently translated (Fig. 5A, lower panel), they differed markedly in their ability to reach the plasma membrane. An ABCA1 mutation at the carboxyl-terminal of ABCA1 associated with TD-3 (2145X) failed to reach the plasma mem-brane and showed near absence of binding to apoA-I (Table I). This suggests that the main defect resides in the inability to bind apoA-I in line with the current concept. However, the 2203X (TD-2) mutant was localized to the plasma membrane as much as the normal control but elicited near absence of binding to apoA-I. These results indicate that the cAMP-signaling defect observed in the 2203X ABCA1 mutant is not caused by impaired localization of ABCA1 to the plasma membrane. This suggests that a specific domain within ABCA1 may be necessary for triggering apoA-I-mediated cAMP signaling, most likely through coupling to G s protein. On the other hand, we have documented previously (8) that 8-Br-cAMP and FRK did not induce ABCA1 phosphorylation in intact fibroblasts from TD-1 (C1477R) as compared with normal cells. In contrast, we show here that the PKA catalytic subunit was able to phosphorylate similarly ABCA1 from the three mutants and control cell lines in vitro (Fig. 5A, upper panel). It is possible that ABCA1 mutations affect protein folding in intact cells that may destabilize ABCA1 at plasma membrane sites accessible to PKA. We are currently investigating the structural requirements for the  5. In vitro phosphorylation of ABCA1 and localization of ABCA1 at plasma membranes in ABCA1 mutants. A, fibroblasts from a normal control (CTR) and TD subjects (TD-1 (C1477R), TD-2 (2203X), and TD-3 (2145X)) ( Table I) were stimulated, and ABCA1 was immunoprecipitated and incubated with [␥-32 P]ATP in the absence or presence of the PKA catalytic subunit (PKA-c) as described under "Experimental Procedures." 32 P-Labeled ABCA1 was separated by electrophoresis and then transferred to polyvinylidene difluoride membrane. 32 P-Labeled ABCA1 was revealed by phosphorimaging. The ABCA1 protein was detected on the same membrane by anti-ABCA1 antibody and used as a control for protein loading. B, cell surface proteins from stimulated normal and TD cells were biotinylated as described under "Experimental Procedures," and the isolated cell surface proteins were then analyzed by Western blot probed with a polyclonal antibody to ABCA1. ABCA1 transporter to couple to the G s protein, activate AC, and regulate receptor activity at the cell surface.
HDL triggers a variety of intracellular signaling events including activation of either PI-PLC, PC-PLC, or PC-PLD, PKC, mitogen-activated protein kinase, tyrosine kinase (31)(32)(33), nitric oxide, intracellular Ca 2ϩ release, and ceramide production (34). The diversity of HDL-mediated cellular responses can in part be explained by the heterogeneity and composition of HDL particles (apolipoproteins and lipids) (35)(36)(37) as well as by the different HDL receptors possibly involved (SR-BI, ABCA1, G protein coupled to phospholipases). Defective regulation of either PC-PLC or PC-PLD in response to apoA-I was observed previously in TD cells (38). The same group has also reported that proteins of the Rho family (RhoA, RhoB, RhoG, Rac-1) are enriched in fibroblasts from TD patients (39), suggesting that G protein-or Rho family protein-dependent coupling of phospholipases was affected by ABCA1 mutations. This concept is supported by a recent study by Yokoyama and co-workers (40) demonstrating that apoA-I activates PKC by PC-PLC-mediated generation of diacylglycerol initiated by removal of cellular sphingomyelin and subsequently phosphorylates and stabilizes ABCA1. This is in agreement with our observations showing that treatment of normal fibroblasts with apoA-I increased membrane-associated PKC activity and treatment of intact normal fibroblasts with phorbol 12-myristate 13-acetate, an activator of PKC, increased ABCA1 phosphorylation (41). This is also consistent with our finding that apoA-I mediated ABCA1 phosphorylation in unstimulated normal fibroblasts even though it does not increase cAMP levels. It is likely that apoA-I-mediated ABCA1 phosphorylation in unstimulated cells is because of a PKC pathway. These observations raise the possibility that ABCA1 may be phosphorylated via various signal-relating molecules, allowing apoA-I to be lipidated through different pathways. Receptor coupling to multiple G proteins, i.e. dual coupling or cross-talk, has been observed in several reports. Thus, receptors for luteinizing hormone/chorionic gonadotropin, cholecystokinin A, and thyrotropin-releasing hormone can couple to both phospholipase C and AC pathways (42,43). In addition, ␤ 2 -adrenergic receptor was found to couple to both G␣ s and G␣ i , and protein kinase A-mediated phosphorylation of the ␤ 2 -adrenergic receptor was shown to serve as a switching mechanism to regulate the dual G protein coupling specificity of the receptor (44). Undoubtedly, further work is necessary to establish more definitively a possible functional link between the ABCA1 transporter, phospholipases, and G protein.
In conclusion, we have found that intracellular cAMP release induced by apoA-I leads to ABCA1 phosphorylation via a PKAdependent mechanism, most likely involving a G protein coupled to ABCA1 transporter. This model requires initial apoA-I binding to the ABCA1 transporter, which couples to G␣ s , leading to activation of AC, cAMP production, and subsequent PKA-mediated ABCA1 phosphorylation allowing lipidation of apoA-I. Further elucidation of the molecular interactions between the apoA-I-ABCA1-G protein complex should clarify the mechanism by which ABCA1 is involved in the biogenesis of HDL particles and suggest a novel therapeutic potential of signaling molecules in preventing or treating atherosclerotic vascular disease.