UC Berkeley

Apolipoprotein (apo) E is an exchangeable apolipoprotein that is critical for the trafficking of lipid and cholesterol nutrients in the brain and peripheral circulation. ApoE is a 299 amino acid (37 kDa) protein comprised of two independently folded functional domains, the carboxy-terminal lipid-binding domain and the receptor-binding amino-terminal (NT) domain that only displays receptor competent activity upon association with lipid. In the absence of lipid, the isolated NT domain (residues 1-183) of apoE adopts an amphipathic four alpha-helix bundle architecture that is characteristic of several other related apolipoproteins.

Various models have been advanced that describe the predicted conformational change of the protein upon lipid binding. Experiments have shown that the alpha-helical secondary structure is preserved if not enhanced upon lipid binding and yet it is known that the protein undergoes a dramatic conformational change in the transition to the lipid-bound state. Low-resolution experiments have provided insight into the mechanism and possible path of this transition, but a high-resolution determination of the lipid-bound conformation of apoE has not been accomplished. Using a combination of unique protein engineering methods and nuclear magnetic resonance (NMR) spectroscopy, this thesis advances the understanding of the lipid-induced conformational change of the apoE N-terminus. Recombinant apoE NT (residues 1-183) is a representative model for apolipoprotein helix bundle conformational flexibility in the presence of lipid and on the surface of lipoprotein particles. The 22 kDa domain is predominantly alpha-helical, monomeric, and comparably stable relative to the native protein. This domain readily forms discoidal particles in the presence of phospholipids, which imparts low-density lipoprotein (LDL) receptor activity to the protein.

A protein engineering approach was used to further define the structural determinants of apoE NT that are necessary for lipid binding. A short helix connecting helix 1 and 2 in the four-helix bundle was replaced by a sequence predicted to adopt a beta-turn. The resulting stable recombinant protein was not compromised in its ability to function as a ligand for the LDL receptor, yet the protein displayed greatly enhanced binding affinity for lipid as assessed by phospholipid solubilization studies, a lipophilic fluorescent dye binding assay, and protection against phospholipase induced aggregation of human LDL fractions.

In order to define the detailed lipid-induced conformational change in apoE, a protein engineering approach termed segmental isotope labeling was deemed necessary to simplify the system for analysis by NMR. Using expressed protein ligation (EPL) methodology, a hybrid apolipoprotein was constructed from two independently generated fragments, apoE residues 1-111 and a 91 amino acid apolipophorin protein fragment. This protein ligation experiment tested the novel use of a pelB leader sequence for the generation of an N-terminal cysteine-containing protein fragment required for the joining of protein fragments by EPL.

Expressed protein ligation techniques were alternatively adapted to create an intact, semisynthetic apoE NT domain using apoE(1-111) and apoE(112-183) protein fragments. This semisynthetic protein displayed nearly identical structural and function properties as wild-type apoE NT by circular dichroism spectroscopy, guanidine denaturation studies, and functional lipid and LDL receptor binding studies. Stable isotope-labeled (15N) apoE(112-183) was produced and ligated to unlabeled apoE 1-111 protein to create a segmental isotope-labeled protein. NMR experiments of the segmental protein further confirmed a structural and functional correspondence between wild-type, fully 15N-labeled and segmental isotope-labeled apoE NT while affirming that the segmental system dramatically simplified the NMR system for examining the protein region containing the LDL receptor recognition sequence.