The interplay between size, morphology, stability, and functionality of high-density lipoprotein subclasses

Abstract

High-density lipoprotein (HDL) mediates reverse cholesterol transport (RCT), wherein excess cholesterol is conveyed from peripheral tissues to the liver and steroidogenic organs. During this process HDL continually transitions between subclass sizes, each with unique biological activities. For instance, RCT is initiated by the interaction of lipid-free/lipid-poor apolipoprotein A-I (apoA-I) with ABCA1, a membrane-associated lipid transporter, to form nascent HDL. Because nearly all circulating apoA-I is lipid-bound, the source of lipid-free/lipid-poor apoA-I is unclear. Lecithin:cholesterol acyltransferase (LCAT) then drives the conversion of nascent HDL to spherical HDL by catalyzing cholesterol esterification, an essential step in RCT. To investigate the relationship between HDL particle size and events critical to RCT such as LCAT activation and lipid-free apoA-I production for ABCA1 interaction, we reconstituted five subclasses of HDL particles (rHDL of 7.8, 8.4, 9.6, 12.2, and 17.0 nm in diameter, respectively) using various molar ratios of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine, free cholesterol, and apoA-I. Kinetic analyses of this comprehensive array of rHDL particles suggest that apoA-I stoichiometry in rHDL is a critical factor governing LCAT activation. Electron microscopy revealed specific morphological differences in the HDL subclasses that may affect functionality. Furthermore, stability measurements demonstrated that the previously uncharacterized 8.4 nm rHDL particles rapidly convert to 7.8 nm particles, concomitant with the dissociation of lipid-free/lipid-poor apoA-I. Thus, lipid-free/lipid-poor apoA-I generated by the remodeling of HDL may be an essential intermediate in RCT and HDL's in vivo maturation.

FIGURE 1

Nondenaturing gel electrophoresis analysis of rHDL particles. Lanes 1, 5, 11, and 17 in gels A, B, C, and D: molecular weight markers (High Molecular Weight Calibration Kit from GE Healthcare) (24, 32). Gel A: Unpurified lipidation mixtures. Lanes 2, 3, and 4, respectively: 30:2:1, 80:4:1, and 160:8:1 POPC:FC:apoA-I ratios (mol/mol). Gel B: Isolated single size particles. Lanes 6 and 7: 7.8 and 8.4 nm rHDL particle purified from 30:2:1 (POPC:FC:apoA-I) lipidation mixture. Lane 8: 9.6 nm rHDL particle purified from 80:4:1 (POPC:FC:apoA-I) lipidation mixture. Lanes 9 and 10: 12.2 and 17.0 nm rHDL particle purified from 160:8:1 (POPC:FC:apoA-I) lipidation mixture. Gel C: Particle samples after 4 months storage at 4 °C. Lanes 12–16, respectively: 7.8, 8.4, 9.6, 12.2, and 17.0 nm. Gel D: Low mobility band from the 9.6 nm particle stored at 4 °C for 4 months (lane 14) repurified by size exclusion chromatography (lane 18).

FIGURE 2

Stoichiometric analysis of apoA-I on rHDL by DMS cross-linking (see Experimental Procedures). Lane 1: Lipid-free WT apoA-I:DMS, 1:935 (mol/mol). Lane 2: Lipid-free WT apoA-I: DMS, 1:10. Lane 3: E136C mutant apoA-I, no DMS. Lane 4: E146C mutant apoA-I, no DMS. Lanes 5–9: apoA-I:DMS, 1:935 (mol/mol). Respectively, 7.8, 8.4, 9.6, 12.2, and 17.0 nm WT rHDL. Oligomeric species yield and mobility are also dependent on the ratio of protein to DMS (e.g., lanes 1 and 2). As the ratio of DMS to lipid-free WT apoA-I increases (lane 1 vs lane 2), the yield of dimeric products is very low compared to higher molecular weight oligomers. Because dimeric apoA-I can migrate in a broad range of apparent molecular weights, we reproduced this range using cysteine bearing apoA-I, which forms disulfide dimers with variable migration properties (lanes 3 and 4).

FIGURE 3

Panel A: Elution profile of size exclusion chromatography purification of a 30:2:1 POPC:FC:apoA-I lipidation mixture. Panel B: NDGGE analysis of the mixture prior to column separation (lane 2) and elution fractions (lanes 4–7). Lanes 5 and 7 show pure 8.4 and 7.8 nm rHDL particles, respectively. Lanes 1 and 3: Molecular weight markers.

FIGURE 4

Electron microscopy of homogeneous samples of 7.8 nm (panels A and B), 8.4 nm (panels C and D), 9.6 nm (panels E and F), and spherical (panels G and H) rHDL particles. Spherical rHDL were provided by Dr. Kerry-Anne Rye and were generated with standard procedures, as described in ref . Bars in the left column equal to 20 nm. See Table 2 for EM sizing. Individual particles are shown enlarged in the right column with boxes of 20 × 20 nm. In the first row of panel F are side views, in the second row are top views, and in the third row are tilted views of individual 9.6 discoidal particles. In general, particle views were assigned by the convention put forth by Catte et al. (13). Briefly, top views were visually identified as the objects showing maximum surface area. Particles with minimal surface area and showing a bidimensional geometry were selected as side views. All particles showing an intermediate shape between the two mentioned above were assumed to represent tilted rHDL views.

FIGURE 5

HDL concentration-dependent activation of LCAT by rHDL particles of different size. LCAT activation by rHDL was monitored by following the cholesterol to cholesteryl ester conversion. After rHDL particles were incubated with radiolabeled cholesterol, BSA, and β-mercaptoethanol, the indicated amount (apoA-I concentration) of rHDL particles of different size was incubated with 50 ng of LCAT at 37 °C for 1 h. The reaction was terminated with ethanol, and then CE was separated from cholesterol and quantified as described in Experimental Procedures. Panels B and C: Lineweaver–Burk linear regression analysis. Plot of the reciprocals of initial velocities vs reciprocals of apoA-I concentrations for 7.8, 8.4, and 9.6 nm rHDL (panel B) and 12.2 and 17.0 nm rHDL (panel C), respectively.

FIGURE 6

Nondenaturing gel electrophoresis analysis of 8.4 nm rHDL particle remodeling over a period of 2 months at 4 °C (A) or at 37 °C (B). Markers are High Molecular Weight Calibration Kit from GE Healthcare (24, 32).