Try this speculation on for size: Maybe the reason more functional partners for amyloid precursor protein (APP)—whether ligand or receptor—have not yet been found is that one of its major partners turns out to be…APP. Speculative, indeed, but in an article published in today’s Molecular Cell, researchers provide evidence that the E2 domain of APP reversibly dimerizes, and suggest that this could occur between complete APP molecules, even those that are membrane-bound.

All aspects of the physiological role for APP are fertile ground for experimentation, since very little has been determined (although a recent report presents F-spondin as a putative ligand for APP; see ARF related news story). There is some evidence that the membrane-bound protein is a receptor, with possible roles in cell adhesion or signaling. The secreted ectodomain of APP may function as a growth factor. A better understanding of APP's tertiary structure will certainly help to elucidate its possible relationships with other molecules.

In their current study, Yongcheng Wang and Ya Ha of Yale University in New Haven, Connecticut, focused on E2, the largest of APP's three domains. This domain contains a high-affinity binding site for the extracellular matrix protein heparin sulfate proteoglycan (HSPG), as well as a sequence that might confer growth factor-like capabilities upon soluble APP. Wang and Ha found evidence that E2 is folded independently of the rest of APP, and their x-ray crystallography data indicate that the tertiary structure of E2 contains two coiled-coil substructures connected through a continuous helix. The authors found that this structure is quite similar to domains of spectrin and similar molecules, including α-catenin, and fits well with data recently obtained by NMR spectroscopy (see Dulubov et al., 2004).

The researchers further found that the E2 domain invariably dimerizes (as opposed to remaining a monomer or forming other oligomers) in solution. This dimer takes an antiparallel orientation with the N-terminal double-stranded coiled-coil of one monomer packed against the C-terminal triple-stranded coiled-coil of its partner. Interestingly, spectrin and α-catenin also make antiparallel dimers. One feature of the dimer specially noted by the authors is a site on the concave surface that is a likely binding site for heparin and HSPG.

Since E2 is the only domain of APP that has been experimentally shown to dimerize, this suggests to the authors that this interaction could account for the observed dimerization of secreted APP. There is also evidence, supported by work in the current study, that membrane-bound APP can dimerize, perhaps also thanks to the E2 region.

Beyond the interesting speculation that APP is its own ligand and receptor, the authors offer another possibility, also as yet untested. Such dimerization of membrane-bound APP could be a basis for cell-cell adhesion—one lonely, membrane-bound APP striking up a relationship with a soul-mate on another cell.—Hakon Heimer


  1. This is a very interesting paper which confirms the observations made in a number of previous studies. The structure of the E2 region gives us a much clearer picture of the overall structure of APP. While the structure of a number of regions of the ectodomain have not yet been determined, it is quite likely that these other regions will contain less structure than the N-terminal domain and the E2 domain, which is the subject of the present study.

    It was pleasing to see confirmation of our prediction that the E2 domain is structure rich and that it contains a heparin-binding domain (see Mok et al., 1997). The dimerization of APP (although previously demonstrated in vitro) is also important, although it is difficult to reconcile the models shown in Fig. 6A of the paper with the crystal structure of the dimer (Fig. 4A).

    The only major concern is whether the unglycosylated structure expressed in E coli will be similar to the native mammalian N-glycosylated structure.

    Finally, the idea that the ectodomain of APP may dimerize raises the question whether the cytoplasmic domain may also dimerize. Could this be a mechanism of signal transduction?


    . Expression and analysis of heparin-binding regions of the amyloid precursor protein of Alzheimer's disease. FEBS Lett. 1997 Oct 6;415(3):303-7. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. AβPP Ligand Found? Neurons Respond to F-Spondin

Paper Citations

  1. . Three-dimensional structure of an independently folded extracellular domain of human amyloid-beta precursor protein. Biochemistry. 2004 Aug 3;43(30):9583-8. PubMed.

Further Reading