The "beta-clasp" model of apolipoprotein A-I--a lipid-free solution structure determined by electron paramagnetic resonance spectroscopy

Abstract

Apolipoprotein A-I (apoA-I) is the major protein component of high density lipoproteins (HDL) and plays a central role in cholesterol metabolism. The lipid-free/lipid-poor form of apoA-I is the preferred substrate for the ATP-binding cassette transporter A1 (ABCA1). The interaction of apoA-I with ABCA1 leads to the formation of cholesterol laden high density lipoprotein (HDL) particles, a key step in reverse cholesterol transport and the maintenance of cholesterol homeostasis. Knowledge of the structure of lipid-free apoA-I is essential to understanding its critical interaction with ABCA1 and the molecular mechanisms underlying HDL biogenesis. We therefore examined the structure of lipid-free apoA-I by electron paramagnetic resonance spectroscopy (EPR). Through site directed spin label EPR, we mapped the secondary structure of apoA-I and identified sites of spin coupling as residues 26, 44, 64, 167, 217 and 226. We capitalize on the fact that lipid-free apoA-I self-associates in an anti-parallel manner in solution. We employed these sites of spin coupling to define the central plane in the dimeric apoA-I complex. Applying both the constraints of dipolar coupling with the EPR-derived pattern of solvent accessibility, we assembled the secondary structure into a tertiary context, providing a solution structure for lipid-free apoA-I. This article is part of a Special Issue entitled Advances in High Density Lipoprotein Formation and Metabolism: A Tribute to John F. Oram (1945-2010).

FIGURE 1. EPR spectroscopic resolution of protein structure

Electron paramagnetic resonance spectroscopy can provide a quantitative measure of the local environment (steric and chemical) of the spin-labeled (nitroxide) amino acid side-chain, as well as the distances between two spin-labeled side-chains (intra- or intermolecular). MTSL (methane thiosulfonate spin label) is used to covalently modify a thiol-group within the target protein (A), thereby introducing a stable unpaired electron. The shape of the EPR spectra (B), the derivative of microwave absorption as a function of the applied magnetic field strength, reports on the relative mobility of the spin labeled side chain. Examples of spectra that indicate low (K94, left) and high (A95, right) mobility are shown. Residues that exhibit high solvent mobility (A95, right) are likely to reside within a random coil and exhibits the sharpest spectral line shapes due to its higher degree of motional freedom, whereas more broadened spectra (appearing more attenuated and rounded) are exhibited by residues in a more ordered environment such as alpha helix (K94, left). The inverse values of the peak-to-peak distances (arrows) are used to provide a semi-quantitative measure of the mobility (inverse spectral linewidth or δ⁻¹), which gives insight into conformational state of the protein at that residue and the degree of steric hinderance. EPR power saturation experiments (C) with hydrophobic (O₂) or hydrophilic (CrOx) relaxation agents gives a measure of the chemical milieu proximal to the labeled side-chain (N₂ is used for background subtraction). Dotted lines indicate the applied power (mW) at the maximal central line amplitude. The analysis provide the accessibility parameters (Π) for the two relaxers, which are used to calculate a polarity index, or contrast function (Φ). Dipolar interaction of two (or more) spin-labels at a distance shorter >20Å results in characteristically broadened spectra (D), and can be used to determine intra- or intermolecular distances.

FIGURE 2. EPR analysis of apoA-I’s lipid-free secondary structure

The inverse central line width (δ⁻¹; Gauss⁻¹), accessibility parameter data for the hydrophilic relaxer (Π_CrOx), hydrophobic relaxer (Π_O2), and contrast value (Φ) are plotted as a function of residue number (A). These were jointly used to determine secondary structure. A periodicity of 3.6 residues per turn was used to identify the regions displaying α-helical character, indicated by yellow sinusoids. Dashed lines indicate regions where the helical structure pattern is less obvious in the mobility score analysis. A periodicity of 2 was used to identify β-strand structure (shown in blue). The linear representation of the resulting secondary structure model is compared to the secondary structure distribution defined by hydrogen-exchange/mass-spectroscopy experiments [13], and by X-ray crystallization [12] (B). α-helical structure is shown in yellow, random coil in green and β-strands in blue. Sites displaying intermolecular interaction are indicated with arrows.

FIGURE 3. Arrangement of apoA-I’s secondary structural elements

ApoA-I assumes higher order associations at concentrations greater than 0.1 mg/ml. Because of this, specific nitroxide labeled locations within apoA-I elicit spin coupling due to their proximity to their counterpart on a paired molecule. Six locations with in apoA-I were identified as exhibiting spin-coupling characteristics (residues 26, 44, 64, 167, 217, 226; red circles) (A). These sites of spin coupling define a central plane for the dimer and provide insight into how these secondary structures align into a tertiary context (B). The α-helices (tan) and β-strands (blue) are drawn to relative scale so that the position of turns (green) can be defined accurately.

FIGURE 4. The “Beta Clasp” model of lipid-free apoA-I

The combination of solvent accessibility, mobility (geometric constraint) and spin coupling was used to conceive the model. Whereas apoA-I typically is presented as a predominantly α-helical protein, this model (A), notably bears four anti-parallel β-strands, potentially arrayed in a single β-sheet (blue). This is surrounded by amphipathic α-helical bundle (tan) from the N-terminal, central, and C-terminal region of apoA-I. We hypothesize the β-strands serve as a hydrophobic core, which stabilizes the amphipathic α-helices in the absence of lipid. At the center of the protein lies the central loop, which is a sustained conformational feature present also on discoidal HDL. On the left is the cartoon structure of lipid-free apoA-I, in the unit cell conformation it would adopt within a dimer. On the right is a schematic representation of the apoA-I unit cell illustrating the relative position of functional elements. The residues of apoA-I that exhibit spin coupling are colored in red. (B) The unit cell monomers of apoA-I (as depicted in panel A) colored in ocher and blue, respectively are aligned based on the residues that exhibit dipolar coupling (red). In the presence of lipid, the β-strands transition to α-helix and the protein adopts an extended helical structure. ApoA-I is a highly dynamic molecule so the lipid-free model represents the average state of the protein in solution and not necessarily the precise conformation of apoA-I.