We have identified three structural and biophysical differences among the three apoE isoforms: (1) apoE4 domain interaction; (2) protein stability and folding; and (3) cysteine content (Figure 2).web.jpg)
Our working hypothesis is that one or more of these differences are responsible for the isoform-specific effects of apoE4 on plasma lipoprotein metabolism, atherosclerosis, and neurodegeneration. To test this hypothesis, we will engineer each of these isoform differences individually into the mouse Apoe gene by gene targeting. In this way we can determine the individual contribution of each difference to isoform-specific effects that are known to occur in plasma and the brain (Figure 2). Sorting out the relative contributions of these differences has important implications in developing effective, apoE4-based therapeutic strategies.
ApoE4 Domain Interaction. ApoE4 binds preferentially to very low density lipoproteins (VLDL), whereas apoE3 binds preferentially to high density lipoproteins (HDL) (Figure 1). We determined that the two domains in apoE4 interact and that this interaction is restricted to apoE4. The interaction is modulated by Arg-61 (amino-terminal domain) and Glu-255 (carboxyl-terminal domain) and is responsible for the VLDL binding preference of apoE4 (Figure 3).
web.jpg)
We are collaborating with Drs. Fred Cohen and Irwin Kuntz (UCSF) and using their DOCK program to identify small molecules that will bind to apoE4 in the vicinity of Arg-61 but not to apoE3 and thereby interfere with domain interaction. We expect that such molecules will represent a therapeutic approach by converting apoE4 into an ‘apoE3-like’ molecule (Figure 3).
In several species, including the mouse, apoE contains arginine and glutamic acid at positions equivalent to positions 112 and 255, respectively, in human apoE. However, these species lack the critical human Arg-61 required for domain interaction. Their apoE contains threonine and, like human apoE3, displays a preference for HDL. Based on these results, we used a ‘knock-in’ gene targeting approach to introduce an arginine codon into the mouse gene to ‘humanize’ mouse apoE at position 61 and introduce domain interaction.
In heterozygous targeted mice, as in human apoE4 heterozygotes, the plasma level of the Arg-61 form is 20–40% lower than the level of the wildtype apoE. This characteristic pattern reflects the more rapid clearance of apoE4 from plasma. The ‘Arg-61’ mouse apoE also displays the expected preference for VLDL. These results demonstrate that domain interaction was successfully introduced in vivo. With this model, we can determine the contribution of domain interaction to the increased risk of cardiovascular disease, neurodegeneration, and cognitive impairment associated with apoE4.
