Prion dimer interface
Application to the species barrier
Prion chameleon helices
It's not a bug, it's a feature
Quaternary structure and amyloid
X-ray Diffraction of Scrapie Prion
Interactive prion models
Pdb coordinates for mouse 121-231, the dimer, and amyloid H1
James Warwicker Institute of Food Research, Reading Uk, fax 44-1734-267917 BBRC 232 (2): 508-512 (1997) BBRC1996 Sep 24; 226(3): 777-82 [with PJ Gane]
In a couple of clever papers on prion quaternary structure, Warwicker sought to identify a long-suspected prion dimer interface through optimized molecular modelling. The 3D coordinates are basically those of fragment 121-231 of mouse and were infered from a 2D graphic using alpha-carbon spacing and characteristic helical parameters, with some element of uncertainty, of course. Warwicker has publicly released his coordinates.
The original NMR work was published in Nature 382 180-82 1986 but the coordinates, PDB accession code 1AG2, though submitted, may not be available until October 1, 1997. However, Dr. Martin Billeter has kindly provided on his Web page the refined alpha carbon positions for 124-226 as of 29 May 1997. These represent coordinates refined relative to the original Nature paper. Riek R, Wider G, Billeter M, Hornemann S, Glockshuber R, and Wuthrich K are the scientists responsible for the upcoming published refinement. Billeter has posted a wonderful superpositioning color graphic that compares the Warwicker model to the refined Zurich coordinates. The Warwicker model has a root-mean-square deviation (rmsd) of 2.7 angstroms from the unrefined Nature model and 3.3 angstroms from the refined coordinates for selected residues. The coordinates used in the July, 1996 Nature paper have a rsmd of 1.5 angstroms from the refined NMR structure released on 29 May 1997.
Warwicker begins by placing the invariant hydrophobic core domain 109-122 into the NMR structure onto an exposed hydrophobic surface component that is runs horizontally across figure 1b of the NMR paper. This is plausible from entropic considerations and also results in a compact globular protein.
Next a second prion monomer is added to the computer model and the steric hindrances associated with various angles of approach are considered. The best fit is shown in figure 2a of the 1997 BBRC paper: two 109-122 helices are anti-parallel in the core of the dimer interface. This is consistent with a standard closed homodimer Z(2) symmetry model predicted decades ago by Jacques Monod: the evolution of complementary surfaces is best attained with a 180 degree rotation.
The proposed structure, while speculative, could very well right. The issue is at the heart of the species orpolymorphism barrier, or rather that component of the species barrier concerned with recruitment of normal prion conformer by exogenous or endogenous mutant rogue conformer. (Other aspects of inter- and intra-species transfer have to do with routes and levels of exposure at the whole-organism level.) A number of known mutants and modulating polymorphisms are indeed situated where they could affect Warwicker's dimer interface.
However, Prof. Rudi Glockshuber has informed us that an in-press PNAS article wll discuss species differences mapped to the 3D structure.; human, sheep, and cow prion structures threaded onto mouse were shown by Wuthrich at a December meeting. A second paper on the atomic details of the refined structure has been submitted to Nature Structural Biology. Prusiner-Cohen also have NMR structure close to publication, possibly showing a larger fragment.
Warwicker's model is of a normal conformer dimer. However, the necessary transition from helix to sheet in the disease state has an interesting connection with recent observations on the chameleon nature of helix 1 and 2 made by MF Perutz in 27 Feb 97 Nature. pg 771. The idea, implemented by Minor DL and Kim PS, Nature 380 (6576): 730-734 (Apr 1996) in protein G insertions, is that some stretches of sequence can form either alpha-helix or beta-sheet depending on what secondary structure is adopted by adjacent residues. In other words, non-local influences are important: their 11 amino acid stretch formed an alpha-helix when inserted by robust alpha-helix and beta-sheet when next to robust beta-sheet.
With prion protein, a shift in the region 104-122 from helix to sheet is widely taken as the structural correlate of the shift from normal to disease conformer. Perutz has examined the three prion NMR/xray helices for chameleon character, finding the long helix H3 to be normal [exterior polar/ interior apolar helical wheel], whereas the shorter two were not. (See color graphic of prion helical wheels.) Short helices are of course intrinsically less stable because of fewer applicable hydrogen bonds.
Helices H1 and H2 don't follow the rules for beta-sheets either [all non-polar if interior of protein, alternating polar-apolar if exterior -- see below]. This leaves us with two stretches of amino acid known to form helices [despite somewhat anomalous properties] and known adjacency to beta-sheet in the case of H1, in a protein known to transition from helix to sheet. Whence Perutz' suggestion for cameleon domains. This fits well with the pre-chameleon suggestion of Glockshuber and Wuthrich that the small original beta sheet nucleates prion structural conversion. (See color animation of this process.) The examples of cameleon peptides discuss the cis situation (same strand) but the concept might be extensible to trans (different peptide of a dimer influencing secondary structure of a chameleon peptide. Certainly, the quaternary structure has a modulating influence on each monomer.
If chameleon peptides are unstable and lead to lethal prion disease states, why weren't they eliminated during evolution? The answer is perhaps that they as unstable by design (two-state regulatory switches). That is, the two conformers are integral to normal prion function. As a concrete example, suppose normal prion role was gated transport and binding some regulator caused the gate to open (resp. close) through this conformation shift. So like they say onsoftware help lines, 'hey, that's not a bug, that's a feature.'
H1 DWEDR YYREN M; as alternates: (asp glu arg tyr glu met )(trp asp tyr arg asn) H2 CVNIT IKQHT VTTTT; as alternates: (cys asn thr lys his val thr thr)( val ile ile gln thr thr thr) H3 ETDVK MMERV VEQMC VTQ (human prion helix sequences, infered from mouse NMR)Perutz also reviewed the general case of amyloid-forming proteins, including prions, in the same note. The same issue of Nature contains an companion article, pp787, on a lysozyme disorder, hereditary amyloidosis, where amyloid consists entirely of mutant allele product. Lysozyme amyloid is composed exclusively of the mutant allele in heterozygotes [unlike some familial CJD] and that neither known mutant allele, ile56thr or asp67his, can recruit normal conformer, making this amyloidosis non-infectious, at least so far. Sporadic events or destablizing entropic or enthalpic mutations of the types discussed by Perutz, might prime the pump (provide the first rogue conformers) for prion protein as well.
It seems that the beta-sheets are quite consistently aligned normal to the fiber axis. Most likely there is a second interface, distinct from the closed dimer interface, responsible for higher order aggregation states (amyloid). This conformation could be of a generic nature applicable to diverse proteins and consist of nested-spoon beta-sheets.
Dr. Hideyo Inouye has addressed all this in a fiber xray diffraction paper in JMolBio 268,#2 , p375-389, May 2, 1997, fulltext online and free. This is a sophisticated paper makes clever choice of experimental material (syrian hamster 90-145; analogue of the human CJD mutant tyr145stop) and uses fiber diffraction to look directly at amyloid. Inouye proposes turns in the middle of both the H1 and H2 domains, in the final structure. Dr. Inouye kindly shared his pdb coordinates with this site on 27 March 1997.
* Dowload and install RasMol as appropriate to your computer platform
*Configure RasMol as a viewer for your browser
*Consider installing Cheme as a plug-in enhancer of RasMol
*Test your installation on the sample files that come with RasMol and/or Chime.
*Be sure you understand the mouse-down pop-up menu that controls the many viewing menus.
*Now sit back and play with these very cool prion images [if screen is black, use popup menu to change display option to spacefill or backbone).
A 'kinemage' [kinetic image] is a scientific illustration presented as an interactive computer display. Operations on the displayed kinemage respond within a fraction of a second: the entire image can be rotated in real time, parts of the display can be turned on or off, any point can be identified by picking it, and the change between different forms can be aninmated. Kinemages are created from Brookhaven Protein Data Bank files using the program PREKIN. The program MAGE is used to view them.
Nature 380 (6576): 730-734 (Apr 1996) Minor DL Jr, Kim PSProtein secondary structures have been viewed as fundamental building blocks for protein folding, structure and design. Previous studies indicate that. To examine the extent to which non-local factors influence the formation of secondary structural elements, we have designed an 11-amino-acid sequence (dubbed the 'chameleon' sequence) that folds as an alpha-helix when in one position but as a beta-sheet when in another position of the primary sequence of the IgG-binding domain of protein G (GB1). Both proteins, chameleon-alpha and chameleon-beta, are folded into structures similar to native GB1, as judged by several biophysical criteria. Our results demonstrate that non-local interactions can determine the secondary structure of peptide sequences of substantial length. They also support views of protein folding that favour tertiary interactions as dominant determinants of structure.
The spatial coordinates for mouse prion have been submitted to the protein structure database at Brookhaven. However, the data will apparently be under review for another 5 months by staff there:
Data under processing at PDB Idcode: 1AG2 Tracking Number: T11143 Entry Description: PRION PROTEIN DOMAIN PRP(121-231), DOMAIN 121-231 Authors: M. BILLETER,R. RIEK,G. WIDER,K. WUTHRICH,S. HORNEMANN,R. GLOCKSHUBER Status for 1AG2: -incomplete->-processing->-depositor->**REVIEW**>**HLD On hold until: Sep 30 1997 All materials arrived as of: Mar 31 1997 Accession Date: Mar 31 1997
JMB May 1997 Hideyo Inouye, Daniel A. KirschnerLimited proteolysis of infectious scrapie prions PrPSc yields an N-truncated polypeptide termed PrP 27-30, which encompasses residues 90 to 231 of PrPSc and which assembles into 100 to 200[emsp14]‰ wide amyloid rods. It has been hypothesized that the infectious prion is converted from its non-infectious cellular form (PrPC) by means of an [Alpha]-helical to [Beta]-sheet conformational change. Secondary structure analysis, computer modeling, and structural biophysics methods support this hypothesis.
Residues 90 to 145 of PrP, which contain two putative [Alpha]-helical domains H1 and H2, may be of particular relevance to the disease pathogenesis, as C-terminal truncation at residue 145 was found in a patient with an inherited prion disease. Moreover, our recent X-ray diffraction analysis suggests that the peptide consisting of these residues (designated SHa 90-145) closely models the amyloidogenic [Beta]-sheet core of PrP.
In the current study, we have analyzed in detail the X-ray diffraction patterns of SHa 90-145. Two samples were examined: one that was dehydrated under ambient conditions whilst in an external magnetic field (to induce fibril orientation), and another that was sealed after partial drying. The dried, magnetically oriented sample showed a cross-[Beta] diffraction pattern in which the fiber axis (rotation axis) was parallel to the H-bonding direction of the [Beta]-sheets. The major wide-angle peaks indicate the presence of [sim]40[emsp14]‰ wide [Beta]-crystallites, which constitute the protofilament. Each crystallite is composed of several orthogonal unit cells, normal to the fiber (a-axis) direction, having lattice constants a=9.69[emsp14]‰, b=6.54[emsp14]‰, and c=18.06[emsp14]‰. Electron density maps were calculated by iterative Fourier synthesis using [Beta]-silk as an initial phase model.
The distribution of density indicated that there were two types of [Beta]-sheet, suggesting that larger and smaller side-chains localized to different sheets. This would arise from folding of the polypeptide in which there are turns in the middle of both the H1 and H2 domains. A monoclinic macrolattice, with a=9.61[emsp14]‰, b=c=52.99[emsp14]‰ and [Alpha]=114.6ƒ, was found to index all the reflections, including those in the low-angle region. This suggests that the [Beta]-crystallites are nearly hexagonally packed. To account for the [sim]100[emsp14]‰ wide fibers visualized by negative staining in the electron microscope, the [Beta]-crystallites would be arranged in 4-mers.
The structure, which occurred during dehydration, could be a transient in the [Alpha]-helical to [Beta]-sheet conversion, suggesting that formation of hydrogen bonding precedes the inter-sheet interaction and assembly into the amyloid of scrapie prion.