Wuthrich et al. made an important observation (Nature 382: 180 1996): mouse prion fragment 121-231 is markedly bipolar: one side of the disks is negative, the other positive. They interpret this latter feature as electrostatic binding to a negatively-charged outer phospholipid plasma membrane surface which then orients the negative face of prion protein outwards.
The dimensional scale is missing from the article but fortunately the axis of helix H1 is parallel to the plane of the diagram, so an axial translation of 1.5 Angstrom per residue is appropriate. Overall the fragment of prion protein studied would fit into a 35 x 35 x10 angstrom box. It thus has a high surface area to volume ratio compared to a globular protein of the same size.
Prion protein has a GPI anchor already so it doesn't need electrostatic binding to anchor itself to the membrane. Major plasma membrane components like sphingomyelin, cerebrosides, phosphatidyl ethanolamine, phosphatidyl choline, sterols, and integral membrane proteins can provide neutral membrane amphipathicity; neuronal membranes are highly specialized mosaics; and prion protein is reported localized within calveolar patches. Bipolarity is not consistently reported from other GPI caveolar proteins. The undetermined amino terminus of mature prion could mask the fragment's positive surface. In short, the proferred explanation for bipolarity is problematic.
The authors do not calculate the electric dipole moment and dipole axis corresponding to this surface, nor did they measure it by dielectric constant relaxation frequency. An accurate calculation is easy given refined coordinates [Biophys J 94:1550 1993]; as a point of departure, the Glockshuber-Wuthrich structure can be modelled as a oblate ellisoid and the dipole moment estimated (with an unscreened uniform dielectric coulombic model) at some 225 debyes orthogonal to the plane of the disc with only the 3 histidines presenting a computational issue at the isoelectric point. (Coordinates can be reverse-engineered from published rendered projections using "thin-sectioning" techniques in IHS color space.) Charge separation is thus a force holding the protein together and could contribute to the exceptional stability of prion protein to heat denaturation or to kinetics of renaturation.
Now the three alpha helices have significant dipole moments of their own (from aligned hydrogen bonds and peptide bonds) not related to that from surface charge, unlike the beta sheet. These align constructively along the helical axis to form macrodipoles (positive axis N-terminal) of 39, 43, and 59 debyes. The pattern is out-of-sequence anti-parallel, ie, there are constructive dipole-dipole interactions between pairs of helices and also with the bipolar dipole, as the dipoles are oriented nose to tail and a few angstroms apart in the 3D picture. Normally,1H NMR directly correlates helix dipole strength and amide proton chemical shifts (J Mol Biol 222: 311 1991).
The accompaning graphic shows the relationship between these features -- it is consistent with the idea of an important dipole contribution to structural stability of prion protein. A previously predicted helix H0 at 109-122 (PNAS 91: 7139 1996) is shown in faint colors in a speculative alignment that matches its dipole with the unpaired end of helix H2. This stretch is the core invariant region of prion protein, hardly changing from chicken to marsupial to mammal since the Devonion, and a likely site for ligand binding. The figure also shows (with a red X) a proposed binding site for a small ligand or catalytic binding site based on dipole distribution. (Helix HO should folded over this but is not, for clarity.) These roles for the observed surface charge distribution could either supplement the membrane binding proposal, or displace it.