ATP hydrolysis-dependent conformational changes in the extracellular domain of ABCA1 are associated with apoA-I binding

Abstract

ATP-binding cassette protein A1 (ABCA1) plays a major role in cholesterol homeostasis and high-density lipoprotein (HDL) metabolism. Although it is predicted that apolipoprotein A-I (apoA-I) directly binds to ABCA1, the physiological importance of this interaction is still controversial and the conformation required for apoA-I binding is unclear. In this study, the role of the two nucleotide-binding domains (NBD) of ABCA1 in apoA-I binding was determined by inserting a TEV protease recognition sequence in the linker region of ABCA1. Analyses of ATP binding and occlusion to wild-type ABCA1 and various NBD mutants revealed that ATP binds equally to both NBDs and is hydrolyzed at both NBDs. The interaction with apoA-I and the apoA-I-dependent cholesterol efflux required not only ATP binding but also hydrolysis in both NBDs. NBD mutations and cellular ATP depletion decreased the accessibility of antibodies to a hemagglutinin (HA) epitope that was inserted at position 443 in the extracellular domain (ECD), suggesting that the conformation of ECDs is altered by ATP hydrolysis at both NBDs. These results suggest that ATP hydrolysis at both NBDs induces conformational changes in the ECDs, which are associated with apoA-I binding and cholesterol efflux.

Fig. 1.

Secondary structure of ABCA1. The HA tag-insertion sites and Walker A lysine mutations are indicated by open circles and open squares, respectively. The TEV protease recognition sequence was introduced between R1272 and R1273, as indicated by the filled triangle. The two cysteine residues in ECD1 and three cysteine residues in ECD2 involved in S-S bond formation between the ECDs are indicated by filled circles.

Fig. 2.

Effect of ATP depletion on apoA-I binding. A: HEK or HEK/ABCA1 cells were incubated with 5 μg/ml Alexa 546-conjugated apoA-I for 10 min. B: HEK/ABCA1 cells were incubated with (middle panel) or without (upper panel) 10 mM NaN₃ and 10 mM 2-deoxy-D-glucose in glucose-free DMEM containing 0.02% BSA and 5 μg/ml Alexa 546-conjugated apoA-I for 10 min. After ATP depletion, the cells were incubated with DMEM containing glucose, 0.02% BSA, and 5 μg/ml Alexa 546-conjugated apoA-I for 10 min (lower panel). Differential interference contrast (DIC) images of cells are shown on the right. Bar, 10 μm.

Fig. 3.

Effects of Walker A lysine mutations on cholesterol efflux and apoA-I binding. A: Cell lysates (15 μg of protein) were separated by 7% polyacrylamide gel electrophoresis, and ABCA1-GFP was analyzed with an anti-GFP antibody. The amount of vinculin was examined as a loading control. B: Free cholesterol efflux was analyzed. HEK293 cells stably expressing wild-type (wt) ABCA1 or the Walker A lysine mutants ABCA1-K939M, ABCA1-K1952M, or ABCA1-K939M,K1952M were incubated for 24 h in DMEM containing 0.02% BSA with or without 5 μg/ml apoA-I. ApoA-I-dependent cholesterol efflux was calculated by subtracting the value obtained in the absence of apoA-I. C: Cells were incubated with DMEM containing 0.02% BSA and 5 μg/ml Alexa 546-conjugated apoA-I for 15 min. Bar, 10 μm. Experiments were performed in triplicate, and average values are shown with SD. ***P < 0.001, significantly different from the value for HEK293 cells.

Fig. 4.

Nucleotide binding and occlusion in wild-type and Walker A lysine mutants of ABCA1. A: ATP binding was analyzed by incubating membranes with 20 μM [α³²P]8N₃ATP for 10 min on ice. ABCA1 was immunoprecipitated after UV irradiation. The amounts of bound nucleotide in mutants were compared with that in wild-type (wt). B: ATP occlusion was analyzed by incubating the membranes with 5 μM [α³²P]8N₃ATP in the presence or absence of 0.4 mM vanadate (Vi) for 10 min at 37°C. ABCA1 in the membranes was photoaffinity-labeled after removing the unbound ATP. Proteins were transferred to PVDF membranes, and the amount of bound nucleotides was visualized by autoradiography with a BAS2500 image analyzer (Fujifilm) and compared with wild-type ABCA1 in the absence of vanadate. The amount of ABCA1 was analyzed by immunoblotting the same membrane (IB). C: ATP occlusion was analyzed with 5 μM [α³²P]8N₃ATP or [γ³²P]8N₃ATP at 37°C. The average values from three to five experiments are presented with the SD. *P < 0.05; **P < 0.01.

Fig. 5.

Effects of inserting the TEV sequence on the function of ABCA1. A: Cells were incubated with DMEM containing 0.02% BSA and 5 μg/ml Alexa 546-conjugated apoA-I for 15 min. Bar, 10 μm. B: Cholesterol efflux was analyzed. Cells transiently transfected with ABCA1 were incubated for 24 h in DMEM containing 0.02% BSA with or without 10 μg/ml apoA-I. C: Membranes were incubated with the indicated units of TEV protease for 30 min at 30°C. D: ATP occlusion was analyzed with 5 μM [α³²P]8N₃ATP in the absence of vanadate at 37°C. *P < 0.05; **P < 0.01.

Fig. 6.

Nucleotide binding and occlusion in each NBD of ABCA1. A: Membranes were incubated with 20 μM [α³²P]8N₃ATP for 10 min on ice and UV irradiated. After immunoprecipitating with KM3110, ABCA1 was treated with or without 0.2 units/μl TEV protease for 30 min at 30°C. B: Membranes were incubated with 5 μM [α³²P]8N₃ATP or [γ³²P]8N₃ATP for 10 min at 37°C. After immunoprecipitating with KM3110, ABCA1 was treated with 0.2 units/μl TEV protease for 30 min at 30°C. The proteins were transferred to PVDF membranes and analyzed by autoradiography and immunoblotting (IB). Experiments were performed in duplicate and yielded similar results.

Fig. 7.

ApoA-I binding and anti-HA antibody staining to ABCA1 with an inserted HA tag. Cells were incubated with DMEM containing 0.02% BSA and 5 μg/ml Alexa 546-conjugated apoA-I (A) or 1 μg/ml anti-HA antibody (B) for 15 min. Bar, 10 μm.

Fig. 8.

Effects of ATP depletion on anti-HA antibody accessibility. A: Cells expressing ABCA1 with an HA tag inserted in the ECDs were incubated with or without 10 mM NaN₃ and 10 mM 2-deoxy-D-glucose in glucose-free DMEM containing 0.02% BSA for 10 min. The cells were then incubated with glucose-free DMEM containing 0.02% BSA and 1 μg/ml anti-HA antibody in the presence or absence of NaN₃ and 2-deoxy-D-glucose for 15 min. Bar, 10 μm. B: The intensity of anti-HA staining was calculated and normalized to the intensity of GFP fluorescence using images from 10 cells (136HA, 349HA, 1421HA) or 40 cells (207HA, 443HA) and ImageJ 1.40 software. The relative anti-HA staining was compared in the presence or absence of ATP depletion (upper panel) or between 136HA and 1490HA (lower panel). ***P < 0.001.

Fig. 9.

Effects of Walker A lysine mutations on anti-HA and anti-Flag antibody accessibility. Cells expressing ABCA1 with or without an HA tag inserted at position 207 (A) or 443 (C) were incubated with DMEM containing 0.02% BSA and 1 μg/ml anti-HA antibody for 15 min. Bar, 10 μm. B, D: The intensity of anti-HA staining was calculated and normalized to the intensity of GFP fluorescence using images from 40 cells (207HA) or 44 cells (443HA) and ImageJ 1.40 software. The relative anti-HA staining of the Walker A mutants and the wild-type protein was compared. E: Cells expressing ABCA1 with or without a Flag tag inserted at position 443 were incubated with DMEM containing 0.02% BSA and anti-Flag antibody (1:1000) for 15 min. Bar, 10 μm. F: The intensity of anti-Flag staining was calculated and normalized to the intensity of GFP fluorescence using images from 50 cells and ImageJ 1.40 software. ***P < 0.001.

Fig. 10.

Effects of an HA or a Flag tag inserted at position 443 on the function and subcellular localization of ABCA1. A: Cells were incubated with DMEM containing 0.02% BSA and 5 μg/ml Alexa 546-conjugated apoA-I for 15 min. Bar, 10 μm. B: Cholesterol efflux was analyzed. Cells transiently transfected with ABCA1 were incubated for 24 h in DMEM containing 0.02% BSA with or without 10 μg/ml apoA-I. C: Cells were treated with sulfo-NHS-biotin and cell lysates were prepared. Biotinylated surface proteins were precipitated with avidin agarose from 150 μg of cell lysates. The precipitated surface proteins (upper panel) and cell lysates (15 μg; lower panel) were separated and detected with an anti-GFP antibody. Western blots were analyzed using a Fujifilm LAS-3000 imaging system. The amount of cell surface ABCA1 was normalized with total ABCA1. D: Cell lysates were separated and detected with the indicated antibodies. The band intensity of anti-HA or anti-Flag antibody was normalized with that of anti-GFP antibody. **P < 0.01; ***P < 0.001.