The low density lipoprotein receptor (LDLR) maintains cholesterol homeostasis by binding and removing low density lipoproteins (LDL), and intermediate density lipoproteins (IDL). Absence of functional LDLRs leads to the disease Familial Hypercholesterolemia (FH), which is associated with an increase in endogeneous cholesterol levels. A single span of seven tandemly repeated modules makes up the extracellular ligand-binding (LB) domain of the LDLR. These LB modules share several common features; each is approximately 40 amino acids long, and contains 6 conserved cysteine residues, which form 3 canonical intra-disulfide bonds. The LB modules of the LB domain are arranged in a head to tail manner. This combination of modules enables the LDLR to bind to its ligands. LDLR and ligand interactions are dependent on the presence of Ca2+ ions.
When this study was begun, little was known about the role of Ca2+ in determining the structure and function of LB modules of the LDLR. The work described in this thesis focuses on the interaction between LB modules and Ca2+ and on the role of Ca2+ in (i) the thermodynamic stability of LB modules, (ii) the formation of a canonical set of disulfide bonds, and (iii) the dissociation of the LDLR-ligand complex in the acidic endosomal environment. Our data support a model in which the LB domain of the LDLR exists as a string of independent Ca2+ -complexed LB modules with the flexibility to bind its small ligand, apoE, and its much larger ligand, apoB 100. We hypothesise that the loss of Ca2+ in the acidic environment of the endosome contributes to the dissociation of the receptor-ligand complex.
Each LB module binds a single Ca2+ with high affinity. Despite the highly conserved nature of the LB modules, the Kds of Ca2+ -binding for different LB modules range from 200 nM up to 14 ɥM. The differences in affinities of Ca2+ -binding arise from differences in the environment of the Ca2+ -binding sites.
Although differences in the Ca2+ -binding sites lead to different Kds for different modules, the metal-binding site in LB modules is optimised for binding by Ca2+. A selection of divalent cations (Cd2+ and Mn2+ with similar sizes to Ca2+ bind rLB1 with Kds of 12 and 34 ɥM respectively, but are unable to induce the native conformation. Similarly, the other Group 2 elements are not able to induce the native conformation.
Ca2+ is also required for the formation of the native disulfide bonded isomers of rLB1 and rLB2 from a mixture of non-native disulfide bonded isomers. Of all the Group 2 divalent cations (Mg2+, Sr2+, Ca2+, and Ba2+) tested, Ca2+ is the only metal ion capable of inducing the efficient formation of native disulfide bonded LB modules.
LB modules with a set of canonical disulfide bonds bind Ca2+ to form a thermodynamically stable isomer. In the absence of Ca2+, the native isomer is stable as long as free thiols are not present. In the presence of a thiol disulfide exchange buffer, the Ca2+ -free native isomer is converted to an equilibrium mixture of native and non-native isomers. The observations made for individual modules were also made for functional bovine LDLRs. In conditions conducive to disulfide bond rearrangements, bovine LDLR retains binding to LDL only in the presence of Ca2+ but not in the absence of Ca2+ We hypothesize that the formation of the Ca2+-complex 'traps' and removes native LB isomers from equilibrium with a pool of non-native isomers.
A two-equilibrium model of folding was used to assess this hypothesis. The equilibria that govern these reactions are (i) the equilibrium that exists between the native isomer of rLB1 and fourteen non-native isomers, and (ii) the equilibrium between the Ca2+-free native isomer of rLB1, and the Ca2+ -complex of this isomer. Support for the model comes from the observation that (i) more native isomer was formed when Ca2+ concentrations were increased, and (ii) the experimental Kds of Ca2+ -binding for rLB1 and rLB2 are 5 ɥM and 6 ɥM respectively, consistent with experimentally determined Kds from separate studies.
In the model, the equilibrium constant (KA1) governs the equilibrium between the Ca2+ free native isomer and non-native isomers. Kinetic analyses show that the KA1s range from 0.007 at 2.5 mM Ca2+ to 0.09 at 50 ɥM Ca2+, or 0.07 in the absence of Ca2+ The model is simplified, and does not take into account the possibility that at high concentrations of Ca2+ additional Ca2+ -dependent pathways may contribute to the formation of the native isomer and hence decrease the KA1. As a first approximation, the model confirms that the 'trapping' of native isomers by formation of the Ca2+-complex results in accumulation of the native isomer. Further refinements to the model will need to accommodate the contribution to Ca2+ -dependent folding by additional Ca2+ -dependent pathways.
The release of ligands from the LDLR occurs in the acidic (pH 5.5) and low Ca2+ concentration (3 ɥM) environment of the endosome. A corresponding increase in the Kds of Ca2+-binding for LB modules, from pH 7.5 to 5.5, is related to the protonation of carboxylate side-chains of Ca2+ -coordinating residues. The experimental Kds of Ca2+-binding suggest that within the endosome, the majority of LB modules are partially saturated with Ca2+ or are free of Ca2+ Conformational changes in LB modules, as a result of Ca2+ release, may act in concert with acid-dependent conformational changes in the EGF-precursor homology domain to allow the efficient release of ligands within the time frame of LDLR recycling.
Sedimentation equilibrium analyses showed that the LB modules exist as monomers in the presence or absence of Ca2+, indicating a lack of strong interactions between modules. The lack of inter-modular interactions may result in a flexible 'beads on a string' conformation capable of binding small ligands such as apoE or larger ligands such as apoB-100.