A 3-D clue to mad cows
3D Structure of Normal Prion Protein
NMR structure of the mouse prion protein domain PrP(121-231)

NMR structure of the mouse prion protein domain PrP(121-231)

Nature 18 July 1996 ... Received 13 May; accepted 5 June 1996.

Roland Riek, Simone Hornemann, Gerhard Wider, Martin Billeter, Rudi Glockshuber & Kurt Wuthrich
Institut fur MoleRularbiologie und Biophysik,
Eidgenossische Technische Hochschule-HonAgerberg,
CH-8093 Zurich, Switzerland

THE 'protein only' hypothesis' states that a modified form of normal prion protein triggers infectious neurodegenerative diseases, such as bovine spongiform encephalopathy (BSE), or Creutzfeldt-Jakob disease (CJD) in humans. Prion proteins are thought to exist in two different conformations: the 'benign' prpc form, and the infectious 'scrapie form', PrPSc. Knowledge of the three-dimensional structure of prpc is essential for understanding the transition to PrPSc.

The nuclear magnetic resonance (NMR) structure of the autonomously folding PrP domain comprising residues 121-231 (ref. 6) contains a two-stranded antiparallel beta-sheet and three a-helices. This domain contains most of the point-mutation sites that have been linked, in human PrP, to the occurrence of familial prion diseases. The NMR structure shows that these mutations occur within, or directly adjacent to, regular secondary structures. The presence of a beta-sheet in PrP(121-231) is in contrast with model predictions of an all helical structure of prpc (ref. 8), and may be important for the initiation of the transition from prpc to PrPSc.

The NMR structure of PrP(121-231) contains three a-helices and a two-stranded antiparallel beta-sheet. The approximate lengths of the helices are from residues 144 to 154, 179 to 193, and 200 to 217, and the lengths of the p-strands are from residues 128 to 131, and 161 to 164. The first turn of the second helix and the last turn of the third helix are linked by the single disulphide bond in the protein. The twisted V-shaped arrangement of these two longest helices forms the scaffold onto which the short beta-sheet and the first helix are anchored. At the present stage of refinement, all regular secondary-structure elements and the connecting loops are well defined, with the sole exception of residues 167 to 176.

The polypeptide fold is stabilized by hydrophobic interactions in a core that contains side chains of the second helix (residues 179,180 and 184), the third helix (residues 203, 206,209, 210, 213 and 214), the beta-sheet (Val 161), the first, mostly hydrophilic, helix (Tyr 150), and three loop regions (residues 134, 137, 139, 141, 157, 158 and 198). With the exceptions of Ile 139, Ile 184 and Val 203, the residues of the hydrophobic core are invariant in the known mammalian prion protein sequences9 (Fig. 3a). Hydrophobic surface patches in PrP(121-231) are located near the ,B-sheet and the loop preceding the first helix. The surface of PrP(121231) is otherwise characterized by a markedly uneven distribution of positively and negatively charged residues.

Mature mouse prpc is a glycosylated 208-residue protein (codons 23-231, with deletion of codon 55 (ref. 9)) that is attached to the cell surface by means of a glycosyl phosphatidyl inositol anchor at its carboxy-terminal Ser 231 (ref.10). It seems to be necessary for normal synaptic functionit (but see ref. 12), longterm survival of Purkinje neuronsi3, and the regulation of circadian activity rhythms and sleep.

The segment of residues 109 - 218 was predicted to form a four-helix-bundle with helices at residues 109-122, 129-141, 178-191 and 202-218 (ref. 8), whereas prediction algorithms did not yield conclusive results for the segment 23-108, which contains the prior-characteristic octapeptide repeats. Attempts to express PrP(108-231) in the periplasm of Eschenchia coli resulted in proteolytic cleavage after residues 112, 118 and 120, whereas PrP(121-231) is stable against degradation in E. coli, folds cooperatively and reversibly at pH 7 , and is soluble at 1 mM concentration in distilled water between pH 4.0 and pH 8.5 (ref. 6).

This segment also contains six of nine point-mutation sites in mature PrP that have been associated with familial prion diseases, as well as both glycosylation sites of PrP and its single disulphide bond. We therefore chose to use PrP(121-231) for the present NMR structure determination. This choice was also supported by the demonstration that the segment 81-231 of mouse PrP is sufficient for propagation of the prion disease in vivo i5, indicating that the C-terminal part of PrP is of special functional importance. In the context of the structure predictions for prpc (ref. 8), the beta-sheet in the NMR structure of PrP(121-231) is an unexpected feature. Evidence for the identification of the beta-sheet is shown in METHODS. For the production of uniformly 15N-labelled and 15N/l3C-doubly iabeiled PrP(121231), cells of E coli containing plasmid pPrP-C3 were grown at room temperature. Interstrand nuclear Overhauser enhancement (NOE) connectivities are the most direct NMR identifiers of antiparallel pieated-structure. Considering the proposed increase of the beta-sheet content in PrP shown in transition from prpc to PrPs~ (refs 5,17), it is tempting to speculate that the short beta-sheet might be a 'nucleation site' for a conformational transition that could include the loops connecting the beta-sheet to the first helix, which is predominantly hydrophilic and does not show amphipathic character.

A systematic search of the Brookhaven data bank did not lead to the identification of other proteins with folds similar to PrP(121-231), and the relative orientation of the three helices in PrP(121-231) is clearly different from the proposed four-helix-bundle models.

Mapping onto the three-dimensional structure of PrP(121231) of sequence variability in mammalian prion proteins of residues important for the species barrier of prion disease transmission and for predisposition to familial prion diseases, and of biochemical properties of the prion proteins shows the following:

  • Invariant residues in mammalian prion protein sequences are not clustered within the regions of regular secondary structure, but form an important part of the hydrophobic core.
  • The two glycosylation sites at Asn 181 and Asn 197 and the solvent-exposed, single Trp 145 are located on the negatively charged surface of the protein.
  • All six residues of PrP(121231) for which mutation is believed to be associated with inherited prion diseases or predisposition to prion diseases are located in regular secondary structure elements or immediately adjacent to them, but none of these residues is located in the relatively isolated first helix. Three of these residues are part of the hydrophobic core, and three are located on the surface. They may therefore either destabilize the three-dimensional protein structure, or influence its ligand-binding properties.
  • As is generally observed for functionally related proteins from different species, residues in PrP(121-231) that are variant in mammalian prion protein sequences are solvent accessible.
  • The disulphide bond 179-214 is highly shielded from solvent contact in the core of the protein.
  • It has been proposed that the species barrier of prion disease transmission between mice and humans is caused by an altered prpSC binding site in PrPC, which involves residues from the segment 96-167 (ref. 20). There are eight sequence differences between mouse and human PrP within this segment, of which five are contained in PrP(121-231). Four of these differences are located within or adjacent to the first helix, which might thus be part of a single binding site for prpSC (Fig. 3a, b).

    The dipolar character of PrP(121-231) (Fig. lb, c) might stabilize an orientation of prpc with its positively charged surface, which also includes hydrophobic surface patches, towards the cell membrane. Both glycosylation sites, as well as the aforementioned species barrier related potential binding site for PrPsC, would then be located on the opposite, negatively charged surface. Further to these initial observations on possible structure-function correlations, we believe that the NMR structure of PrP(121-231) will provide a basis for more rational design of future in vitro and in vivo experiments on prion proteins and prion diseases.

    3D Views of mouse prion protein
    Upper Left Globular fold and surface properties of PrP(121-231).

    Ribbon diagram of the structure of the mouse prion protein domain PrP(121-231), indicating the positions of the three helices (yellow) and the antiparallel two-stranded beta-sheet (cyan). The connecting loops are displayed in green if their structure is well defined, and in magenta otherwise. The disulphide bond between Cys 179 and Cys 214 is shown in white. The N-terminal segment of residues 121-124 and the C-terminal segment 220-231 are disordered and not displayed. Upper Right: Prion protein backbone chain. Locations of selected residues in the three-dimensional structure of PrP(121-231). The backbone is shown in grey and the orientation of the molecule is as in Fig. 1c. The side chains of amino-acid residues with mutations that have been associated with inherited prion diseases7 are highlighted in red (line (1) in a). The solvent-accessible glycosylabon sites at Asn 181 and Asn 197 are shown in green, and the disulphide bond is shown in yellow. Five residues that may be involved in the species barrier are shown in blue.

    Bottom: Colors indicate the surface electrostatic potentials, with blue for positive charges, red for negative charges. [Recall that the phospolipid bilayer is negatively charged; the authors suggest the markedly positive side of the prion protein faces this.]

    Sequence of prion protein showing significant residues Above: Residues involved in sequence variations among mammalian prion proteins or associated with the species barrier of prion disease transmission and with inherited prion diseases.

    Residues contributing to the hydrophobic core of the domain are underlined; variable residues among mammalian prion proteins are marked with asterisks. Mutations in human PrP that have been associated with inherited prion diseases (a stop codon at residue 145 is not shown here, nor the Met232Arg mutation [not contained in mature PrP]. All of these residues are identical in wild-type human and mouse PrP.

    The polymorphism at codon 129 in human PrP, where homoygosity appears to increase susceptibility to sporadic CJD7, is marked by italics. Residues in PrP(121-231) forwhich experimental evidence has been presented that they contribute to the species barrier of prion disease transmission between mice and humans.

    1. Griffith, J. S. Natune 215,1043-1044 (1967).
    2. Pnusiner, S. B. Science 252,1515 1522 (1991).
    3. Weissmann, C. Tnends Cell Biol. 4,10-14 (1994).
    4. Weissmann, C. Natune 376, 628-629 (1995).
    5. Pan, K.-M. etal. Pnoc. natn. Acad. Sci. U.S.A. 90,10962-10966 (1993).
    6. Hornemann, S. & Glockshuber, R. J. molec. Biol. (in the press).
    7. Pnusiner, S. B. Anch. NeunJl. 50,1129-1153 (1993).
    8. Huang, Z. et al. Panic. natn. Acad. Sci. U.S.A. 91, 7139-7143 (1994).
    9. Schatzl, H. M., Da Costa, M., Taylor, L., Cohen, F. E. & Pnusiner, S. B. J. molec. Biol. 245, 362
    374 (1995).
    10. Stahl, N. & Pnusiner, S. B. MSEB J. 5, 2799-2807 (1991).
    11. Collinge, J. et al. Natune 370, 295-297 (1994).
    12. Lledo, P.-M. et al. Proc. natn Acad. Sci. U.S.A. 93, 2403-2407 (1996).
    13. Sakaguchi, S. et al. Natune 380, 528 531(1996).
    14. Tobler, 1. et al. Nature 380, 639-642 (1996).
    15. Fischer, M. et ah EMBO J. 15, 1255 1264 (1996).
    16. Wuthrich, K. NMR of Proteins and Nucleic Acids (Wiley, New York, 1986).
    17. Huang, Z., Pnusiner, S. B. & Cohen, F. E. Fold. Design 1,13-19 (1996). 
    18. Bernstein et al. J. molec. Biol. 112 535 542 (1977).
    19. Holm, L & Sander, C. Proteins 19,;65-173 (1994).
    20. Telling, G. C. et al. Cell 83, 79-90 (1995).
    21. Guntert, P., Bnaun, W. & Wuthrich, K. J. molec. Biol. 217, 517-530 (1991).
    22. Comell et al. J. Am. chem. Soc. 117, 5179 5197 (1995).
    23. Honig, B. & Nicholls, A. Science 258,1144 1149 (1994).
    24. Koradi, R., Billeter, M. & Wuthrich, K. J. molec. Graph. 14, 51 55 (1996).
    25. Studier, F. W. & Moffat, B. A. J. molec. Biol. 189,113-130 (1986).
    26. Cavanagh, J., Fairbrother, W. J., Palmer, A. G. & Skelton, N. J. Protein NMR

    A 3-D clue to mad cows
    3D Structure of Normal Prion Protein
    NMR structure of the mouse prion protein domain PrP(121-231)

    A 3-D clue to mad cows

    Roger Highfield ... July 18 ... Daily Telegraph

    THIS IS the three-dimensional atomic structure of the brain protein linked to CJD and mad cow disease that will provide important clues to the cause of these and other spongiform diseases.

    Two professors at the Swiss Federal Institute of Technology unveil in the current issue of Nature the structure of an important part of the mouse variety of the "prion" protein, thought to be necessary for the disease to take grip. Profs Kurt Wthrich and Rudi Glockshuber have discovered the twists and turns in the atomic structure of the prion protein by the use of a technique called nuclear magnetic resonance spectroscopy.

    "Discovering the structure provides a new basis for biological and medical research into prions and prion diseases," said Prof Glockshuber. The prion idea is controversial, since it marks a biological heresy - an infective agent without any genetic material. Proponents believe that a normal prion protein can be converted, irreversibly, into an abnormal prion protein, and that this might happen when an abnormal protein comes into contact with a normal one.

    Brain damage results from a molecular domino effect, when this single abnormal molecule can convert normal versions to the abnormal - insoluble - form, leading to the build-up of deposits. The Swiss study will cast light on this puzzle, in particular help scientists find the structural differences between the different forms of the protein.

    The investigators took a close look at part of the ubiquitous, cellular form of the protein thought to be converted to the abnormal form in disease. They showed that this domain contains a structure, known as the beta-sheet, in addition to the expected structure called an alpha-helix, which is linked with the development of the infectious form.

    Disease-causing mutations lie within the regions of regular structure or directly adjacent to them, suggesting that the mutations may destabilise the benign form of the prion protein, aiding its conversion to the infectious agent.

    3D Structure of Normal Prion Protein Reported

    Deseret News Web 11 July 1996

    ZURICH, Switzerland (Reuters) -- Swiss scientists said Wednesday they had made a major breakthrough in research into causes of mad cow disease by discovering the three-dimensional shape of prions, a protein found in nerves and the brain. [This could allow strukctural and functional roles to be assigned to individual amino acids and alleles in the prion protein.]

    The chemical structure of prions, one of more than 100,000 different proteins in the human body, has been known for some time but the physical form remained a mystery until now. Scientists at Zurich's Federal Institute of Technology (ETH) said mapping prions would allow researchers to hunt for changes in the structure of the proteins that may cause the disease, bovine spongiform encephalopathy (BSE), and a related brain-wasting disease in humans.

    ''Discovering the structure provides a new basis for biological and medical research into prions and prion diseases,'' biophysics professor Kurt Wuethrich told a news conference. Wuethrich and molecular biology professor Rudolf Glockshuber, who headed the research at ETH, said their findings would appear in a letter to the science journal Nature due for publication Wednesday.

    The two researchers said their work was based on a widespread hypothesis that a mutated type of prion causes BSE in cows, its human form Creutzfeldt-Jakob disease, and related diseases.

    The ETH researchers were able to map the twisting shape of prions by using nuclear magnetic resonance, a technology that pinpoints the position of separate compunds making up the prion and therefore allows a composite picture to be drawn.

    The prions mapped were normal ones taken from mice and are believed to be nearly identical to human prions. Abnormal prions were not examined because of technical barriers. However, Wuethrich and Glockshuber said having an accurate map of a normal prion would allow researchers to replace or alter each of its segments to discover whether a change produces disease.

    ''Further research could then possibly show what kind of medicines would have to affect what part of a prion in order to treat the disease,'' Wuethrich said. ''Maybe this could lead to designing molecules to replace prion defects,'' Glockshuber added. Both scientists said it was too early to say whether their findings would lead to a breakthough in treatment of degenerative brain diseases.

    NMR structure of the mouse prion protein domain PrP(121-231)

    Nature 382, 180-182 (1996)
    R Riek, S Hornemann, G Wider, M Billeter, R Glockshuber & K Wuethrich

    Spongiform encephalopathies such as scrapie, bovine spongiform encephalopathy and Creutzfeldt-Jakob disease are thought to be caused by a modified pathological form of a 'prion protein' present in normal individuals. Point mutations in certain parts of the prion protein may accelerate the conversion process. Riek et al. have taken a close look at the domain in question. Their study of the mouse prion protein using nuclear magnetic resonance (NMR) spectroscopy shows that this domain contains beta-sheet in addition to the expected alpha-helix. Disease-causing mutations lie within the regions of regular structure or directly adjacent to them, suggesting how the transition to the disease-causing form of the protein might occur.