The Residue 129 Polymorphism in Human Prion Protein Does Not Confer Susceptibility to Creutzfeldt-Jakob Disease by Altering the Structure or Global Stability of PrPC*

There are two common forms of prion protein (PrP) in humans, with either methionine or valine at position 129. This polymorphism is a powerful determinant of the genetic susceptibility of humans toward both sporadic and acquired forms of prion disease and restricts propagation of particular prion strains. Despite its key role, we have no information on the effect of this mutation on the structure, stability, folding, and dynamics of the cellular form of PrP (PrPC). Here, we show that the mutation has no measurable effect on the folding, dynamics, and stability of PrPC. Our data indicate that the 129M/V polymorphism does not affect prion propagation through its effect on PrPC; rather, its influence is likely to be downstream in the disease mechanism. We infer that the M/V effect is mediated through the conformation or stability of disease-related PrP (PrPSc) or intermediates or on the kinetics of their formation.

post-translational process that involves conformational change. PrP Sc can be distinguished biochemically from PrP C by its partial protease resistance and detergent insolubility.
Although the precise molecular events involved in this conversion remain ill defined, molecular genetic and in vitro studies support the hypothesis that some sort of direct interaction between PrP Sc and either PrP C or some less organized state occurs. This interaction results in the PrP Sc conformation being imposed upon the substrate protein, and the process of conversion is favored by sequence complementarity (8 -13). A key piece of evidence supporting this and the protein-only hypothesis in general is the finding that the large majority of cases of sporadic CJD are homozygous with respect to a common polymorphism at position 129 in the human prion protein, in which either methionine or valine can be encoded (only ϳ49% of the UK population are homozygous with respect to this polymorphism) (9).
Elderly survivors of the kuru epidemic (an acquired prion disease largely restricted to the Fore linguistic group of the Papua New Guinea Highlands, which was transmitted during endocannibalistic feasts) who had multiple exposures at mortuary feasts, are, in marked contrast to younger unexposed Fore, predominantly PRNP 129 heterozygotes. Kuru imposed strong balancing selection on the Fore, essentially eliminating PRNP 129 homozygotes. World-wide PRNP haplotype diversity and coding allele frequencies have suggested that strong balancing selection at this locus occurred during the evolution of modern humans (14).
This polymorphism is crucial to the etiology and neuropathology of prion disease; to date, only individuals homozygous for methionine (Met/Met) have succumbed to variant CJD (3,15,16), and iatrogenic CJD occurs predominately in homozygotes with an excess of valine homozygotes in cases related to exposure to contaminated human pituitary hormones (17). The importance of sequence homology is further emphasized by the observation that in some families with inherited prion diseases, the age of onset is significantly later in individuals heterozygous to this polymorphism (18 -21). Thus, heterozygosity with respect to this polymorphism appears to confer substantial resistance to prion disease.
Perhaps more fundamentally, this polymorphism is also associated with the propagation of distinct human prion "isolates" or strains seen in different CJD phenotypes. Prion strains can be distinguished by their biological properties: distinct incubation periods, clinical features, and pattern of neuropathological targeting in experimental animals. It is now apparent that human prion strains are associated with the different biochemical features of PrP Sc , with differences in PrP Sc conformation and glycosylation that can be serially propagated in transgenic mice expressing human PrP (3,5). The residue 129 HuPrP polymorphism appears to be able to affect the generation of these strain-specific PrP Sc conformers and by inference the phenotype of human prion disease (22,23). Such phenotypic effects of this polymorphism are also seen in inherited prion disease associated with the D178N mutation in PRNP, where this mutation may result in the phenotype of CJD or fatal familial insomnia, depending on the 129 genotype of the mutant allele (24).
Further, there are multiple distinct molecular forms of human PrP Sc identified by molecular strain-typing studies (5,22,25,26). These are thought to represent distinct conformational states of PrP Sc that differ in either the fold of the individual polypeptide chains or in the manner in which the chains are packed in a multimolecular complex or both. These states are manifested at the molecular level by (a) their response to cleavage by proteinase K and subsequent sizing of the resistant segment of polypeptide chain by SDS-polyacrylamide gel electrophoresis and (b) the ratio of diglycosylated, monoglycosylated, and unglycosylated forms found in the protease-resistant PrP Sc material. These PrP Sc types are associated with distinct clinico-pathological phenotypes in affected patients. Types 1-3 are associated with classical CJD, whereas variant CJD is only associated with the type 4 strain (5). The propagation of these strain types is highly dependent upon the codon 129 genotype and hence the primary structure of the prion protein. For instance, the type 1 strain is only seen in methionine homozygotes, whereas the type 3 strain has not been seen on this genetic background. Most strikingly, variant CJD has only been seen to date in homozygous MM individuals.
This polymorphism is thus fundamental to prion disease in humans, and therefore an understanding of its influence on PrP folding may provide strong clues as to the molecular mechanisms that underlie prion propagation. Current theories concentrate on the differing abilities for methionine or valine to be accommodated within ␤-sheet structure or the thermodynamic stabilities of both forms of PrP C , both of which are thought to affect the propensity of PrP C to convert to PrP Sc (17,27,28).
As a first step in characterizing this polymorphism, its effect on the three-dimensional structure of the valine form of PrP C has been determined. Here we describe the influence of the polymorphism on the fold stability, solution dynamics, and unfolding characteristics of the normal cellular form of prion protein (PrP C ).

Plasmid Preparation, Protein Expression, and Purification-
The open reading frame of the human PrP gene (PRNP), containing methionine at residue 129, was amplified by PCR using oligonucleotide primers designed to create a unique N-terminal BamHI site and C-terminal HindIII site for directional cloning of the fragment into the expression vector pTrcHisB (Invitrogen Corp.). The primer corresponding to the N-terminal region of PRNP to be expressed was designed to mutate a glycine at codon 90 to methionine and to incorporate a thrombin cleavage site, with the C-terminal primer replacing a methionine residue at 232 with a stop codon. Standard protocols for site-directed mutagenesis were used to create another construct with the valine at position 129.
The ligated pTrcHisB/PRNP construct was used to transform the Escherichia coli host strain BL21(DE3) (Novagen), genotype FЈ ompT hsdS B (r B Ϫ m B Ϫ ) gal dcm (DE3), which was then plated onto Luria-Bertoni (LB) agar plates containing 100 g/ml carbenicillin. Following sequence verification that clones contained the correct construct, cultures were grown for purification using a modification of protocols previously described (29). For cleavage of the fusion protein, thrombin was used at 0.1 unit of thrombin (Novagene) per 1 mg of protein for 16 h at room temperature. Protein concentrations were determined by UV absorption using a calculated molar extinction of 19893 M Ϫ1 cm Ϫ1 at 280 nm.
NMR Sample Preparation and NMR Spectroscopy-Uniformly 15 Nand 13 C/ 15 N-labeled HuPrP(129V) and HuPrP(129M) samples were expressed in E. coli using an EMBL minimal medium recipe with 13 C 6glucose and ( 15 NH 4 ) 2 SO 4 as the sole carbon and nitrogen sources, respectively, and purified as described (29). To ensure that the proteins were free of paramagnetic metal ions, they were unfolded in 6 M guanidine hydrochloride, 100 mM EDTA, pH 8.0, and then dialyzed against 20 mM sodium acetate-d 3 , 3 mM sodium azide, pH 5.55, prior to concentration. NMR spectra were acquired at 303 K on 0.6 -1.2 mM 15 N-and 13 C/ 15 N-labeled samples in 20 mM sodium acetate-d 3 , 3 mM sodium azide, pH 5.55 (in either 90% H 2 O plus 10% D 2 O or 99.9% D 2 O) using Bruker DRX-500 and DRX-600 spectrometers equipped with 5-mm 13 C/ 15 N/ 1 H triple resonance probes. Proton chemical shifts were referenced to TSP. 15 N and 13 C chemical shifts were calculated relative to TSP, using the gyromagnetic ratios of 15 N, 13 C, and 1 H ( 15 N/ 1 H ϭ 0.101329118, 13 C/ 1 H ϭ 0.251449530). NMR data were processed and analyzed on Silicon Graphics Workstations using Felix 2000 software (kindly donated by Accelrys, San Diego).
Assignments and Structure Calculations-Backbone resonances (H N , N, C ␣ , CЈ, C ␤ ) of HuPrP(129V) and HuPrP(129M) were assigned using a standard suite of triple resonance NMR experiments (30 -33). Almost complete backbone assignments were determined, the exceptions being the amide protons and nitrogens of residues 167, 169 -171, and 175, for which no resonances were detected. (As discussed below, these residues occupy a loop region between strand II and helix II, which is undergoing conformational exchange, resulting in line-broadening of NMR signals). Side chain 13 C and 1 H resonances were determined through threedimensional 15 N-separated total correlation spectroscopy-HSQC (34), three-dimensional 13 C-separated HCCH-and CCH-total correlation spectroscopy (35) experiments and confirmed using through-space connectivities in NOE spectroscopy spectra. Aromatic ring protons were determined by analysis of through-space connectivities in NOE spectroscopy spectra. NOE distance constraints were derived from threedimensional 15 N-/ 13 C-separated NOE spectroscopy spectra with 100-ms mixing times (36,37) and assigned using in-house macros and UNIX scripts. NOEs were calibrated on the basis of a survey of covalently fixed distances and divided into three groups with upper bounds of 5, 3.8, and 2.8 Å. J HNHA coupling constants were obtained through threedimensional HNHA spectra (38). Backbone and angle constraints were obtained using the program TALOS (39). Only residues classified by the program as "good" were included, and restraints were implemented with bounds of Ϯ 1.5, where and are the average and S.D., respectively, of the dihedral angles, as predicted by TALOS. Hydrogen bond restraints were included for slowly exchanging backbone amide resonances (see below) and implemented as two NOE constraints, d N-O Ͻ 2.6 -3.2 Å and d NH-O Ͻ 1.7-2.3 Å. Three hydrogen bond restraints were used for the short anti-parallel ␤-sheet. These hydrogen bonds were determined on the basis of the above slow amide exchange and observed interstrand NOE connectivities. Structure calculations were carried out for both the C-terminal domain (residues 125-231) and N-and C-terminal domains (residues 91-231) of HuPrP(129V) using the standard NMR anneal and refine scripts provided within CNS (40). One hundred structures were calculated from randomized starting coordinates, with the lowest energy 20 structures being further refined and chosen to represent the available structural ensemble.
Amide Exchange-Hydrogen-deuterium exchange rates (k ex ) were determined by diluting 1 mM 15 N-or 13 C/ 15 N-labeled HuPrP samples with an equal volume of 20 mM sodium acetate-d 3 , 3 mM sodium azide, pH 5.55, dissolved in 100% D 2 O, and acquiring a series of 1 H 15 N HSQC spectra at 303 K. The decay curves of the 1 H 15 N HSQC cross-peaks were fitted to single exponential decays with offset, and protection factors (k ex /k int ) for observable amides were determined using intrinsic exchange rates (k int ) (41). Acquisition of the first experiment began approximately 5 min after mixing, setting a lower limit on the detection of protection factors of approximately 50.
Spin Relaxation Measurements-Spin relaxation measurements (T 1, T 2 , and 15 N{ 1 H} NOEs) were acquired on 1 mM 15 N-labeled HuPrP(129V) and HuPrP(129M) samples as described in Ref. 42. The T 1 data were obtained using 15 where m represents the sensitivity of the unfolding transition to denaturant, and D is the denaturant activity (43).

RESULTS
Three-dimensional Structure of HuPrP(129V)-The molecule used in this paper comprises residues 91-231 of the protein (HuPrP(129V) and HuPrP(129M)). This length of construct was chosen, since it corresponds with the fragment of PrP that is associated with prion propagation (1,2,44). The protein consists of an N-terminal, flexible and disordered tail (residues 91-124), which contains a high affinity transition metal-binding site (45), and a folded, globular domain consisting of residues 125-231. This domain consists of three ␣-helices encompassing residues 144 -154 (helix I), 173-192 (helix II), and 200 -226 (helix III) and a short, anti-parallel ␤-sheet made up of residues 129 -131 and 161-163. A single disulfide bond that links helices II and III holds the core of the protein together. The polymorphism is situated within the first strand of the ␤-sheet (Fig. 1B).
Structure calculations were carried out for both the C-terminal domain (residues 125-231) and N-and C-terminal domains (residues 91-231) of HuPrP(129V). A number of medium range NOE connectivities were observed within the N terminus of the protein (residues 91-124); however, these did not define any secondary structure, a point supported by the lack of 13 C chemical shift perturbation away from random coil values. This result is consistent with the relatively slow spin relaxation of NMR signals in this region, which is indicative of a predominantly unstructured peptide.
Within the C-terminal domain of the protein, the secondary structure elements are well defined, as reflected by the r.m.s. deviations from the mean structure of Ͻ0.70 Å for backbone heavy atoms and Ͻ1.0 Å for all heavy atoms (Table I). Regions where the level of definition falls are associated with the loop connecting strand 2 of the ␤-sheet and helix 2 (residues 167-171), the loop that connects helices II and III (residues 194 -199), and the C terminus of helix 3 (residues 224 -226) (Fig.  1A). The loop comprising residues 167-171 is characterized by extensive line broadening of NMR signals. This has been reported for a number of prion proteins from different species and is due to conformational exchange between two or more backbone conformations, on the millisecond time scale, which results in broadening of NMR signals. The lack of definition observed within the NMR ensemble for this region of the protein is thus due to lack of constraints, as well as conformational averaging. The increased disorder seen in the other less defined regions, however, appears to be due to conformational equilibria on a faster time scale and is discussed later.
Overall, the three-dimensional structure of HuPrP(129V) derived here has the same tertiary fold as the previously determined structures of mouse (46), hamster (47), bovine (48) and ovine (49) prion proteins and a human protein containing the pathogenic mutation E200K (50). They also closely resemble the NMR and crystal structures of HuPrP(129M) (51, 52) ( Fig. 2A). The r.m.s. deviations for backbone heavy atoms within structured regions are less than 1.2 Å between the NMR structures of the two polymorphs and less than 0.75 Å between the HuPrP(129V) and HuPrP(129M) crystal structures, which is within the uncertainty of the derived structures. Indeed, they coincide remarkably well, given the differences in solution conditions for both samples. Regions where local differences occur are associated with the highest sequence variation between different species and between different HuPrP mutants. These regions comprise loop 167-171, and the C termini of helices II and III, regions for which the greatest conformational variability is also observed, as discussed above. Therefore, the degree of structural definition for these regions of the protein is rather limited.
The site surrounding the polymorphism is better defined, allowing a detailed examination of the effect of the methionine to valine substitution at position 129 (Fig. 2B). For this, we have compared the HuPrP(129V) NMR structure with the high resolution crystal structures of human (52) and sheep prion protein (49), as well as the HuPrP(129M) NMR structure, all of which contain methionine at position 129. Overall, the region  surrounding the polymorphism is remarkably similar between the two forms (Fig. 2B). The methionine and valine side chains are well defined in the NMR structure ensemble, and both occupy similar spatial regions, with no significant deformation of the residues surrounding the polymorphism site with either residue. This is of particular interest, since a D178N mutation is associated with a clinico-pathological phenotype of FFI when residue 129 is methionine and of CJD when it is valine. It has been proposed that the side chains of residues 129 and 178 interact via hydrogen bonds between the side chains of Tyr 128 , Arg 164 , and Asp 178 (24, 53) (although in the averaged NMR structure of HuPrP(129M), the side chain of Arg 164 is not in a position to accomplish this (51)). The side chains of these residues are well defined in the HuPrP(129V) NMR structure, and NOEs between the backbone and side chain resonances of residues 128, 129, 164, and 178 are observed. From the distribution of conformations within the NMR ensemble, it is not possible to determine unambiguously whether the side chains of Tyr 128 , Arg 164 , and Asp 178 are hydrogen-bonded to each other, but these residues are sufficiently close in individual members of the ensemble for this hydrogen-bonding to be present. Hence, the close association of the side chains of residues 128, 164, and 178 is maintained in both polymorphs, and there does not appear to be a significant disruption of the hydrogenbonding network surrounding residue 129 (Fig. 2B). This conclusion is further supported by amide exchange and chemical shift data described below. Thus, the effect of residue 129 in the D178N mutation does not appear to be resolvable through examination of the structure surrounding the polymorphism. Fold Stability and Partially Unfolded Species Populated by HuPrP Polymorphs-Although the three-dimensional structures of the HuPrP polymorphs appear very similar, it is entirely possible that either may access, to varying degrees, alternatively folded states involved in the disease process. To assess this, the folding behavior of both species was determined in detail.
First, the equilibrium unfolding behavior of both polymorphs was examined; the equilibrium denaturation profiles of both species are shown in Fig. 3A. The amide and aromatic CD curves show that both molecules denature in a single co-operative transition, without the formation of any populated intermediate species during equilibrium unfolding. The equilibrium constant between the native and unfolded states derived from these unfolding curves for both proteins are, within error, the same, at ϳ5000. This finding is similar to that obtained when the methionine of the mouse prion protein is substituted with valine (54).
Although this indicates that both proteins have the same overall stability for the folded state and that the residue 129 polymorphism does not appear to be acting on prion disease by altering the fold stability of PrP C , this does not exclude the possibility of other folding intermediates that are transiently populated in the folding process being present to differing degrees in both forms of the protein.  1. B, amide protection factors (k ex /k int ) in HuPrP(129M) and HuPrP(129V) for those residues with measurable protection. (Those additional residues that appear protected in HuPrP(129V) are due to peak overlap in HuPrP(129M), which precluded measurement of those peaks.) The protection factor corresponding with the equilibrium constant between the native (N) and unfolded (U) states of the protein is plotted as a dashed line (log 10 K (N/U) ). Regions of highest protection surround the disulfide bond that links helices II and III and that forms a core of residual structure in what is believed to be the unfolded state of the protein. C, chemical shift differences of C ␣ nuclei between HuPrP(129V) and HuPrP(129M) at pH 5.55 and 30°C. Regions of secondary structure in HuPrP are denoted by arrows (␤-strands) and cylinders (␣-helices). The chemical shift differences of residue 129 have been corrected for the intrinsic chemical shift differences of methionine and valine in an extended ␤-strand (60).
To investigate this, the rates of hydrogen/deuterium exchange for backbone amide resonances were determined, which has been used to identify regions of the protein that remain hydrogen-bonded until the protein becomes unfolded (Fig. 3B) (55). The observed amide exchange protection data shows that the stabilities of secondary structure elements are very similar in both polymorphs. The regions of highest stability within both forms of the protein are located in the three helical regions of the protein that associate to form the core of the molecule. For both forms of PrP C , the protection factors show the stability of these secondary structure elements to be equivalent to the equilibrium constant between native and unfolded states of the protein. The only regions of the protein where this is not the case are the region centered around the disulfide bond that links helices 2 and 3 and the first strand of the ␤-sheet, where the polymorphism resides. The former displays anomalously high protection factors, due to residual local structure in the unfolded state of the protein (55). The latter region, however, displays protection factors that are below the detectable limit (ϳ50 under the conditions of the present experiment). However, protection factors for residues 161 and 163, which are on the other strand of the ␤-sheet, exhibit similar protection factors in both forms of PrP C . Thus, although we can say that substitution of valine for methionine has not significantly increased the stability of hydrogen bonds surrounding residue 129 to a level where protection is observable (ϳ50), it does not appear to have significantly affected the stability of hydrogen bonds associated with the opposite side of the ␤-sheet. We can therefore conclude that for both forms of HuPrP, there are no significantly populated species, either folding intermediates or alternatively folded states in which the three helical and second ␤-strand core of the protein is disrupted, and that the effect of the polymorphism on PrP C (if any) is likely to be outside of this core region.
Some secondary structure elements of the protein show very little protection from exchange. For example, the C-terminal regions of helix II (residues 186 -194) and helix III (residues 219 -226) and the first strand of the ␤-sheet exhibit protection factors below the detectable limit, consistent with a significant level of conformational flexibility in these regions of the protein. In the case of the ␤-sheet, this conclusion is supported by weak, broadened 1 H 15 N HSQC signals observed from this region (residues 129 -131 and 161-164), whereas for the helical regions, although downfield-shifted C ␣ chemical shifts characteristic of ␣-helical structure are observed, the downfield shift is reduced, consistent with these regions being in equilibrium with more unfolded structures (data not shown).
The higher conformational variability in the C termini of these helices in comparison with the core of the protein is further supported by small chemical shift differences between the M and V polymorphs. These chemical shift differences are small but lie outside the uncertainty in measurement. The changes in C ␣ chemical shifts indicate that there is a slight stabilization of the C terminus of helix III, and destabilization of the C terminus of Helix II in HuPrP(129V) compared with HuPrP(129M). In previous studies, chemical shift differences in these regions of the protein between different PrP constructs have been variously ascribed to small pH differences between samples or to transient interactions between the folded domain and the unstructured N terminus of the protein (51,56). The chemical shift differences observed here lie outside those caused by uncertainty in the precise pH of the solutions and indicate a difference in the distribution of stability throughout the molecule, implying a possible role for the polymorphism. Chemical shift changes immediately adjacent to the polymorphism are, however, harder to interpret, since there are inev-itably changes that are associated with the change in chemical composition. Indeed, the sequence surrounding a residue can cause a change in C ␣ chemical shift on the order of Ϯ0.2-0.3 ppm in unfolded proteins, which is larger than the chemical shift changes observed adjacent to the polymorphism (57). Any changes in conformation reported by these chemical shift perturbations are small, since they do not affect the observed structure or stability of the protein and reflect small changes in the population distribution of the structural ensemble, which fall within the resolution of the solution structure.
Spin Relaxation Measurements-In order to further characterize the conformational dynamics of both polymorphs and in particular the regions of low amide protection and conformational variability, both constructs were probed using 15 N spin relaxation measurements (Fig. 4), which provide information on flexibility within the protein. The steady-state 15 N{ 1 H} NOEs and the longitudinal and transverse relaxation times (T 1 ( 15 N) and T 2 ( 15 N)) for the folded, well defined, regions of HuPrP(129V) are characteristic of a globular protein of this size. In particular, residues within the ␤-sheet region, where the polymorphism resides, do not exhibit dynamic properties significantly different from the rest of the protein, despite the low amide protection factors and broadened 15 N 1 H HSQC signals. Residues 190 -199, however, which link helices II and III, display smaller but still positive 15 N{ 1 H} NOE values (and slightly increased T 2 values), indicating that these residues are indeed relatively more flexible than the structured parts of the protein, resulting in the decreased definition in the NMR structure ensemble for this region. The degree of flexibility for this region is still significantly less than the N and C termini of the protein (residues 91-124 and 227-231), which display large negative 15 N{ 1 H} NOE values indicative of molecular motions on the subnanosecond time scale, consistent with previous observations that they are unstructured. The increased flexibility within the C terminus of helix III appears to extend to residue 220 and is consistent with the lack of amide exchange and the reduced downfield C ␣ chemical shifts.
Overall, the spin relaxation properties of both polymorphs are very similar. In particular, the region surrounding the residue 129 polymorphism shows remarkably little difference in spin relaxation values for both constructs, with slight changes in the 15 N{ 1 H} NOE for HuPrP(129V) suggestive of only a slight increase in conformational mobility surrounding the polymorphism. DISCUSSION This study set out to compare the conformation and stability of the methionine and the valine forms of human PrP C and was stimulated by the dramatic effect of the codon 129 variations both on susceptibility to and phenotype of human prion disease. Individuals with an MV genotype are highly protected against developing sporadic and acquired prion disease. This effect can be explained by the principle of molecular complementarity combined with a gene dosage effect. In other words, if there is a spontaneous conversion of PrP C to PrP Sc and only identical molecules can be recruited in the propagation process due to structural complementarity, then a homozygote will have a double dose of available substrate; hence, the probability of initiation and/or the rate of propagation will be enhanced.
The above arguments establish the general point that the residue 129 polymorphism confers different molecular properties on the prion protein and that these properties strongly influence both susceptibility to disease and the conformation of PrP Sc . It follows that an understanding of the mechanism underlying this methionine/valine effect would provide valuable insights into the process by which the cellular conformation of the protein is converted to the disease-related PrP Sc state. To this end, we addressed the question of whether the influence of the M and V residues at this position results from changes in the conformation and stability of PrP C . Our results establish a close similarity of the valine and methionine forms of the protein in the PrP C conformation. No obvious effect on the side chain packing or dynamics in the region surrounding the mutation are discernible in the structural ensemble calculated from NOE-derived constraints, and chemical shift data for this region also show remarkably little difference between the two sequences. In addition, the polymorphism does not affect the global stability of the molecule, as evidenced by the lack of effect on the denaturant-induced unfolding transition and on the hydrogen exchange behavior. The latter measurements also probe the propensity for local regions of the protein to unfold or for the protein to adopt less stable, alternative conformations. In these respects, also, the methionine and valine forms behave identically.
From these observations, it seems likely that the influence of residue 129 on prion disease results from its effect on the conformational properties of PrP Sc , such that the type 1 and type 4 conformations are formed only by the methionine variant, whereas type 3 is favored by valine and the type 2 conformation is formed equally well by either.