Susceptibility of Sheep for Scrapie as Assessed by In Vitro Conversion of Nine Naturally Occurring Variants of PrP

Abstract

Polymorphisms in the prion protein (PrP) gene are associated with phenotypic expression differences of transmissible spongiform encephalopathies in animals and humans. In sheep, at least 10 different mutually exclusive polymorphisms are present in PrP. In this study, we determined the efficiency of the in vitro formation of protease-resistant PrP of nine sheep PrP allelic variants in order to gauge the relative susceptibility of sheep for scrapie. No detectable spontaneous protease-resistant PrP formation occurred under the cell-free conditions used. All nine host-encoded cellular PrP (PrP^C) variants had distinct conversion efficiencies induced by PrP^Sc isolated from sheep with three different homozygous PrP genotypes. In general, PrP allelic variants with polymorphisms at either codon 136 (Ala to Val) or codon 141 (Leu to Phe) and phylogenetic wild-type sheep PrP^C converted with highest efficiency to protease-resistant forms, which indicates a linkage with a high susceptibility of sheep for scrapie. PrP^C variants with polymorphisms at codons 171 (Gln to Arg), 154 (Arg to His), and to a minor extent 112 (Met to Thr) converted with low efficiency to protease-resistant isoforms. This finding indicates a linkage of these alleles with a reduced susceptibility or resistance for scrapie. In addition, PrP^Sc with the codon 171 (Gln-to-His) polymorphism is the first variant reported to induce higher conversion efficiencies with heterologous rather than homologous PrP variants. The results of this study strengthen our views on polymorphism barriers and have further implications for scrapie control programs by breeding strategies.

Scrapie is a fatal and infectious neurodegenerating disease occurring in sheep and goats. The disease belongs to the group of transmissible spongiform encephalopathies (TSEs) or prion diseases found in humans and animals. Creutzfeldt-Jacob disease, Gerstmann-Sträussler-Scheinker syndrome, and fatal familial insomnia in humans and bovine spongiform encephalopathy (BSE) in cattle also belong to this group. Prion diseases are characterized by the accumulation of an “infectious” abnormal protease-resistant isoform (PrP^Sc) of the host-encoded cellular prion protein (PrP^C) in tissues of the central nervous system. Although the exact origin and nature of the causative agent remain unknown, it is thought to transmit and replicate by protein only (16). PrP^Sc molecules form the major, if not the only, component of the agent transmitting the disease (28).

Several polymorphisms in the open reading frame of PrP are associated with differences in phenotypic expression of prion diseases such as incubation period, pathology, and clinical signs. For PrP of sheep, 10 mutually exclusive amino acid polymorphisms at positions 112 (Met to Thr), 136 (Ala to Val), 137 (Met to Thr), 138 (Ser to Asn), 141 (Leu to Phe), 151 (Arg to Cys), 154 (Arg to His), 171 (Gln to Arg or Gln to His), and 211 (Arg to Gln) have been described (Table 1) (1, 2, 6, 12, 13, 20, 25, 33). The polymorphisms at codons 136, 171, and to a lesser extent 154 occur frequently, while the polymorphisms at codons 137, 211, and to a lesser extent 112 are rare. The allelic variant PrP^VRQ (amino acids at positions 136 [Val], 154 [Arg], and 171 [Gln] are indicated in superscript by single-letter amino acid codes; polymorphisms at other codons are indicated separately [Table 1]) is significantly associated with a high susceptibility to scrapie and short survival times of scrapie-affected sheep of many different breeds (6, 9, 14, 20, 21, 25). In contrast, the allelic variant PrP^ARR is significantly associated with resistance to natural and experimental infections with scrapie and BSE in probably all sheep breeds (1, 6, 9, 14, 15, 20). In breeds where the PrP^VRQ allele is rare or absent (for instance, the Suffolk breed), the phylogenetic wild-type (wt) PrP^ARQ allelic variant is associated with increased scrapie susceptibility but with lower penetrance than found for the PrP^VRQ allele (21, 34). Little is known about the association of the allelic variants PrP^{T₁₁₂ARQ}, PrP^{AT₁₃₇RQ}, PrP^AHQ, PrP^ARH, and PrP^{ARQQ₂₁₁} with susceptibility to scrapie.

TABLE 1.

Allelic variants of sheep PrP

Variant designation		Polymorphic amino acid residue at position:
Allele	Proteina	112	136	137	138	141	151	154	171	211
PrPARQ	PrPC_wt and PrPSc_wt	M	A	M	S	L	R	R	Q	R
PrPT112ARQ	PrPC_112T	T	A	M	S	L	R	R	Q	R
PrPVRQ	PrPC_136V and PrPSc_136V	M	V	M	S	L	R	R	Q	R
PrPAT137RQ	PrPC_137T	M	A	T	S	L	R	R	Q	R
PrPAN138RQb	PrPC_138N	M	A	M	N	L	R	R	Q	R
PrPAF141RQ	PrPC_141F	M	A	M	S	F	R	R	Q	R
PrPAC151RQb	PrPC_151C	M	A	M	S	L	C	R	Q	R
PrPAHQ	PrPC_154H	M	A	M	S	L	R	H	Q	R
PrPARR	PrPC_171R	M	A	M	S	L	R	R	R	R
PrPARH	PrPC_171H and PrPSc_171H	M	A	M	S	L	R	R	H	R
PrPARQQ211	PrPC_211Q	M	A	M	S	L	R	R	Q	Q

^aPrP^ARQ has been designated as phylogenetic wt since all polymorphic variants can be derived from this variant by a single nucleotide or amino acid change. 

^bPolymorphisms published during manuscript preparation and therefore excluded from this study. 

The mechanisms by which the different allelic variants contribute to susceptibility to scrapie are not completely understood. From genetic studies, it can be deduced that naturally occurring variants of PrP, including those associated with a high risk for scrapie, probably cannot induce the spontaneous development of scrapie in sheep (5, 18, 19) as has been found for inherited prion diseases. Furthermore, it has been shown that the scrapie susceptibility-linked polymorphisms at codon 136 and 171 can modulate the efficiency of the cell-free conversion of sheep PrP^C to protease-resistant forms (PrP-res [protease-resistant PrP for which no linkage with infectivity has been demonstrated]) induced by PrP^Sc (4). In this cell-free system, sheep PrP^C_VQ (=PrP^C_136V) and PrP^C_AQ (=PrP^C_wt) are efficiently converted to PrP-res by the homologous PrP^Sc (PrP^Sc_VQ/VQ [=PrP^Sc_136V] and PrP^Sc_AQ/AQ [=PrP^Sc_wt], respectively). In contrast PrP^C_AR (=PrP^C_171R) is poorly converted to PrP-res by PrP^Sc (4, 29). Also, other biological aspects of TSE diseases such as species barriers, polymorphism barriers, and the nongenetic propagation of prion strain phenotypes are reflected in the specificities of PrP conversion reactions in this cell-free system (3, 4, 24, 29). Therefore, the system may be used to gauge the potential in vivo transmissibility of TSEs and to predict the susceptibility of hosts for TSEs (3, 24, 29).

In this study, we used this cell-free system to determine the relative conversion efficiencies of the various sheep PrP allelic variants by PrP^Sc isolated from sheep with different PrP genotypes. Based on the observed conversion efficiencies, we tried to gauge the linkage between nine natural occurring variants of PrP (PrP^ARQ and eight polymorphic variants [Table 1]) and the susceptibility of sheep for scrapie, and to a minor extent we tried to gauge the transmissibility of scrapie between sheep. This relative scrapie susceptibility profile may allow further development and/or justification of specific sheep breeding programs. It may also provide insight into structurally important residues involved in modulation of the conversion efficiencies of PrP^C to PrP^Sc, and it may reveal important sites in the PrP molecule to target in the development of potential therapeutics. The effect of each of the eight polymorphisms on the cell-free conversion efficiencies and the predicted in vivo modulation of scrapie phenotype will be discussed.

MATERIALS AND METHODS

PrP constructs and expression.

The sheep PrP allelic variants PrP^ARQ, PrP^VRQ, and PrP^ARR were cloned, expressed, and characterized as described before (4). The full open reading frames of the sheep allelic variants PrP^{AT₁₃₇RQ}, PrP^{AF₁₄₁RQ}, PrP^AHQ, PrP^ARH, and PrP^{ARQQ₂₁₁} were cloned by PCR amplification and proofread by DNA sequencing essentially as described by Bossers et al. (6). Sheep PrP allelic variant PrP^T₁₁₂ARQ, which was not readily available, was constructed by site-directed mutagenesis using standard methods. After subcloning all of the different sheep PrP allelic variants (EMBL accession no. AJ000679 to AJ000681 and AJ000734 to AJ000739) to eukaryotic expression vector pECV7 (4), the PrP constructs were transfected to N2a/H (Hubrechts, Utrecht, The Netherlands) eukaryotic cells as described elsewhere (4). Single-cell clones that expressed each PrP allelic variant were analyzed for expression by immunoperoxidase monolayer assay and radioimmunoprecipitation (4). The best PrP^C-expressing single-cell clones were selected and/or stored for further labeling experiments. To monitor the stability and validity of the expressed sheep PrP sequences, the genomically integrated sequences were proofread by PCR amplification and subsequent analyzed by denaturing gradient gel electrophoresis after several rounds of subculturing the single cell clones (6).

Radiolabeling and purification PrP^C.

Single-cell clones that expressed the different sheep PrP^C variants were subcultured 2 days before labeling and labeled at near confluency as described previously (4, 7). Briefly, cells were starved for 30 to 60 min in culture medium containing only 1/10 of the normal concentration of methionine and 7.5 μg of tunicamycin D per ml. After labeling for 90 to 120 min at 37°C with at least 1 mCi of [³⁵S]methionine-[³⁵S]cysteine Tran³⁵S-label (ICN) per 25-cm² flask, the cells were lysed (0.5% Triton X-100, 0.5% sodium deoxycholate, 5 mM Tris-Cl [pH 7.4] [4°C], 150 mM NaCl, 5 mM EDTA) on ice in the presence of protease inhibitors (0.1 mM Pefabloc SC, 0.5 μg of leupeptin per ml, 0.7 μg of pepstatin per ml, 1 μg of aprotinin per ml). Proteins were precipitated from detergent lysates by 4 volumes of methanol at −20°C and subsequently sonicated in DLPC buffer (0.05 M Tris-HCl [pH 8.2], 0.15 M NaCl, 2% [wt/vol], N-laurylsarcosine, 0.4% [wt/vol] lecithin) containing the same protease inhibitor cocktail. ³⁵S-PrP^C was immunopurified by using antibody R521-7 (raised to sheep PrP peptide 94-107), which specifically reacts with sheep PrP^C (4, 11a). Immunocomplexes were collected by using protein A-Sepharose, and bound ³⁵S-labeled proteins were eluted in 0.1 M acetic acid (pH 2.8). If necessary, eluants were concentrated by vacuum evaporation and reconstituted in 50 mM citrate buffer pH 6.0 (G. J. Raymond, personal communication). Yield of radiolabeled PrP^C was measured with a micro-β counter, and eluates were stored on ice until further use.

PrP^Sc purification and analysis.

PrP^Sc was isolated from randomly collected brain tissue (cerebrum) of five different PrP^ARQ/ARQ sheep, four different PrP^VRQ/VRQ sheep, and one PrP^ARH/ARH sheep. All suffered from natural scrapie confirmed by immunohistochemistry. PrP^Sc was purified in the absence of protease inhibitors by ultracentrifugational pelleting from 15 g of Sarkosyl-homogenated brains (4, 8). After pelleting through a 20% sucrose cushion, the pellet was sonicated in 275 μl of phosphate-buffered saline containing 1% sulfobetaine (SB 3-14) and stored in portions at 4°C. After 1 h of digestion with 45 μg of protease K (PK) per ml at 37°C, yields of PrP^Sc were quantified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by SYPRO Orange (Molecular Probes) protein staining. Direct fluorescence within the 20- to 30-kDa region was determined with a STORM-840 (Molecular Dynamics) in comparison to the internal PK band and a standard dilution of bovine serum albumin. In addition, the relative amounts of PrP^Sc were quantified by Western blotting using antibody R521-7 (1:1,500), ECF fluorescence substrate (Amersham-Pharmacia Biotech), and the STORM-840. All isolates contained about 2 to 6 μg of PrP^Sc per g of brain. PrP^Sc yields from PrP^ARQ/ARQ sheep brain (3 to 6 μg/g) were almost always higher than yields from PrP^VRQ/VRQ sheep brain (2 to 4 μg/g). PrP^Sc concentrations were equalized by further dilution in phosphate-buffered saline containing 1% SB 3-14. Isolates of PrP^Sc were stored at 4°C and briefly sonicated prior to use.

Cell-free conversion.

Approximately 0.9 μg of PrP^Sc per conversion reaction in siliconized tubes was partially denatured at a final concentration of 2.5 M guanidine HCl (Gdn-GdnHCl) (using 8 to 9 M Gdn-HCl in 50 mM sodium citrate) overnight at 37°C in the presence of protease inhibitors. The partially denatured PrP^Sc was subsequently added to a conversion mix containing about 10,000 cpm of nonglycosylated ³⁵S-labeled PrP^C and conversion buffer to give final concentrations of ≤1 M Gdn-HCl, 50 mM sodium citrate (pH 6.0), 1% N-laurylsarcosine, and 5 mM cetyl pyridinium chloride in a total volume of 34 μl or less. The use of nonglycosylated PrP^C does not seem to alter species specificity, strain specificity, and polymorphism barrier specificity (3, 4, 24, 29). Reactions were mixed regularly and incubated for 2.5 days at 37°C. After incubation, each reaction volume was increased to 100 μl by addition of Tris-NaCl (50 mM Tris-Cl [pH 7.5], 150 mM NaCl) and split 1:10. The 9/10 fraction was digested with 35 μg of PK per ml for 1 h at 37°C, and thereafter PK was inactivated by adding Pefabloc SC on ice. The proteins in both fractions, minus (1/10) and plus (9/10) PK, were precipitated for at least 2 h with methanol in the presence of 20 μg of thyroglobulin as a carrier. Pellets were sonicated and boiled in 4% β-mercapthoethanol–4 M urea containing Laemmli sample buffer and analyzed by SDS-PAGE on a 14% gel (Novex). Percentage of conversion was determined by measuring radiolabel in specific regions of the gels (input PrP^C, 25 to 27 kDa; conversion products, 20 to 22 kDa), using for the initial experiments exposures of X-ray film and for later experiments phosphorimaging (STORM-840). Percent conversion was calculated as conversion signal/start signal × 10% × 100/90. At least two or more independent conversion reactions were performed from each different PrP variant. Normalization was performed within each set of conversions (nine PrP^C variants with one PrP^Sc type) in comparison to the homologous conversion in that particular set. The significance of differences between PrP^C conversion efficiencies was calculated by the linear mixed model regression method.

RESULTS

The nine different naturally occurring sheep PrP^C variants could be readily detected after labeling, immunoprecipitation, and SDS-PAGE by phosphorimaging (Fig. 1, lanes 1 to 9). Due to the absence of glycosylation, a uniform product with a molecular mass of about 26 kDa was present in all lanes, indicating that apart from the single amino acid changes, no differences could be observed with regard to processing and/or partial endogenous proteolysis. In addition, also at the cell culture level, we detected no (microscopic) differences between the nine different variants in biosynthesis and localization of PrP^C in immunoperoxidase monolayer assays. The nine radiolabeled PrP^C variants revealed no PrP-res formation after incubation under cell-free conversion conditions without addition of exogenous PrP^Sc (Fig. 1, lanes 10 to 18). This finding indicates that under these in vitro conditions, no detectable spontaneous formation of PrP-res occurs, which coincides with the requirement of exogenous agent for natural scrapie development (5, 18, 19). In contrast, if partially denatured PrP^Sc is added to the conversion reaction, readily detectable PrP-res formation occurs with the specific shift in molecular mass of about 6 kDa (Fig. 1, lanes 19 to 27).

FIG. 1.

Phosphorimage of an SDS-polyacrylamide gel (indicated in pseudo-intensity staining) showing cell-free conversion reactions of nine different nonglycosylated sheep PrP^C variants by sheep PrP^Sc_136V to PK-resistant forms. Lanes 1 to 9 show approximately 1/10 of the input of PrP^C in the various conversion reactions. Incubation of these variants under cell-free conditions revealed no spontaneous PrP-res formation after PK digestion (lanes 10 to 18). Different amounts of PrP-res were formed upon addition of partially denatured exogenous PrP^Sc (lanes 19 to 27). PK digestion removes about 6 kDa from the N-terminal part of PrP when converted to protease-resistant forms (compare lanes 1 to 9 with lanes 19 to 27). Molecular mass markers (in kilodaltons) are indicated.

Three different PrP^Sc preparations, i.e., PrP^Sc_wt, PrP^Sc_136V, and PrP^Sc_171H, could be obtained from natural scrapie cases homozygous for PrP. Of several independent isolates, about 1 μg of PrP^Sc was partially denatured and subsequently renatured under conversion conditions in the presence of one of each of the nine different ³⁵S-labeled PrP^C molecules. After PK digestion, SDS-PAGE, and phosphorimaging, different amounts of PrP-res could be observed (Fig. 1, lanes 19 to 27). These sets of nine conversion experiments were repeated several times with PrP^Sc preparations isolated from different sheep brains, except for PrP^Sc_171H, of which the various isolates were obtained from the single homozygous PrP^ARH/ PrP^ARH scrapie case available worldwide. The data show that for unknown reasons, the conversion efficiencies between sets vary enormously (Fig. 2) independent of PrP^Sc isolate. However, the relative conversion efficiencies of the nine allelic variants induced by either PrP^Sc_wt, PrP^Sc_136V, or PrP^Sc_171H in multiple and independent experiments were very similar. For example, in all PrP^Sc_136V-induced conversion sets, the PrP^C_136V allelic variant was always converted with highest efficiency, PrP^C_wt was converted with reduced efficiency (P_{different_136V} = 0.06), and PrP^C_171R was converted almost not at all (P_{different_wt} < 0.01). Therefore, we normalized the data for each set of nine conversion experiments to the homologous conversion reaction within that set, usually the strongest conversion reaction (Fig. 3).

FIG. 2.

Scattergraph of individual percentages of PrP^C conversion to protease-resistant forms. Each set of conversions (nine PrP^C variants induced by a single PrP^Sc type performed at the same time) is indicated by a single marker. The nine different sheep PrP^C variants are converted with different efficiencies by either PrP^Sc_136V, PrP^Sc_wt, or PrP^Sc_171H. Mean conversion percentages (±standard deviation) are indicated.

FIG. 3.

Sets of conversions (a plus b plus c), defined as all nine different PrP^C variants incubated with one type of PrP^Sc, were normalized to the homologous conversion reactions. Mean normalized conversion efficiencies (±standard deviation) and the number of independent conversion reactions of each PrP^C variant are indicated.

Within each conversion set, remarkable differences in the convertibility of PrP^C into PrP-res are visible. In each PrP^Sc set, a trend in decreasing convertibility can be observed. The order of this trend in the three different sets is summarized in Fig. 4. From these three different PrP^Sc sets, we created a consensus in which the rare allelic variants PrP^{AT₁₃₇RQ} and PrP^{ARQQ₂₁₁} were excluded (Fig. 4). This consensus indicates the relative order in which these allelic variants are predicted to be linked to susceptibility for scrapie. The most efficient converting variants by each of the three types of PrP^Sc were PrP^C_136V, PrP^C_wt, and PrP^C_141F, while PrP^C_112T, PrP^C_154H, and PrP^C_171R converted with low efficiency. The variants PrP^C_171H and PrP^C_211Q converted with intermediate efficiency. Some remarkable differences between the first two PrP^Sc sets exist (compare Fig. 3a and b). One observed difference is the switch in convertibility of PrP^C_136V and PrP^C_wt with homologous and heterologous PrP^Sc (Fig. 3a and b, bars 1 and 3). Another point of interest is the relatively high conversion rate of PrP^C_141F, which has conversion properties similar to those of PrP^C_wt (Fig. 3a and b, bars 1 and 5; P_{different_wt} = 0.77). The PrP^C_137T variant behaves like PrP^C_136V in PrP^Sc_wt- and PrP^Sc_171H-induced reactions (Fig. 3b and c, bars 3 and 4; P_{different_136V} = 0.88 and 0.80, respectively), while this variant is inconvertible by PrP^Sc_136V (Fig. 3a, bar 4; P_{different_wt} < 0.01 while P_{different_171R} = 0.37).

FIG. 4.

Relative conversion efficiencies of the various sheep PrP^C variants by each PrP^Sc. PrP^C allelic variants are indicated by their single amino acid polymorphisms and ranked within each type of PrP^Sc in order of decreased conversion efficiency. *, virtually absent PrP allelic variants in vivo; PrP^AT_137RQ and PrP^{ARQQ₂₁₁} were excluded.

The smaller data set of the convertibility of various PrP^C variants by PrP^Sc_171H shows also a shift in the relative conversion efficiencies of the nine PrP variants. Due to the limited availability of PrP^Sc_171H, only one to three data points could be generated for each PrP^C variant. The results show, however, that the heterologous variants PrP^C_wt, PrP^C_136V, and PrP^C_137T have higher conversion efficiencies with PrP^Sc_171H than the homologous PrP^C_171H variant itself (Fig. 3, compare bars 1, 3, and 4 with bar 8 in each panel). More importantly, the convertibility of the PrP^C_154H and PrP^C_171R variants in the PrP^Sc_171H set remained low (Fig. 3c, bars 6 and 7).

DISCUSSION

In a previous study (4), we demonstrated that the conversion efficiencies of three sheep PrP variants (PrP^C_136V, PrP^C_wt, and PrP^C_171R) reflected their linkage with scrapie susceptibility. In this study, we measured the conversion efficiencies of nine different PrP allelic variants with three different types of PrP^Sc. Based on these results, we tried to gauge the linkage of PrP alleles with relative susceptibility for scrapie. Of these nine allelic variants, those with polymorphisms at either codon 112, 137, 141, 154, 171, or 211 have not been significantly associated with a particular scrapie phenotype.

The nine different natural variants of sheep PrP^C, incubated with different types of PrP^Sc, converted to PrP-res in vitro with different efficiencies. The most efficient converting PrP^C variants were PrP^C_136V, PrP^C_wt, and PrP^C_141F. The conversion efficiencies of variants PrP^C_154H, PrP^C_171R, and to a minor extent PrP^C_112T were consistently low in all reactions induced with the different types of PrP^Sc. From the perspective of breeding strategies, of great interest are the low-convertible variants PrP^C_154H and PrP^C_171R for positive selection and the variants PrP^C_136V, PrP^C_wt, and PrP^C_141F for negative selection. The presence in the same flock of several PrP variants which are linked to resistance may turn out to be safer in scrapie control programs based on breeding strategies and may contribute to maintaining a more genetically diverse sheep population.

The conversion efficiencies of the allelic variants PrP^C_136V, PrP^C_wt, and PrP^C_171R are consistent with the earlier in vitro observations of a smaller conversion data set (4). Although the PrP^Sc_wt in the latter study was less potent in inducing conversions of PrP^C_wt and PrP^C_136V, a comparable but less drastic switch in the convertibility of PrP^C_wt and PrP^C_136V was observed if conversions were induced by either PrP^Sc_wt or PrP^Sc_136V. The codon 112 polymorphism Met to Thr reduces the convertibility of PrP compared to PrP^C_wt in conversion reactions induced by different types of PrP^Sc (P = <0.04). This reduced convertibility is consistent with observations that this allelic variant in homozygous form is found only in healthy sheep (at low frequency). Only a few heterozygous PrP^VRQ/PrP^{T₁₁₂ARQ} or PrP^ARQ/PrP^{T₁₁₂ARQ} scrapie-infected sheep of the Ile-de-France or Japanese Suffolk breed have been described (22, 25). In these cases, the other alleles are probably dominant and result in scrapie development. In the theoretical three-dimensional structure of sheep PrP, this polymorphism is located in the highly flexible N-terminal region (Fig. 5). The larger side chain of Met, the polarity of Thr, or the shorter hydrogen bond between amino acids at codons 110 and 112 may influence the stability of PrP^C and/or the interaction with PrP^Sc since both side chains are directed toward the potential interaction site.

FIG. 5.

Three-dimensional representation of all polymorphic residues in sheep PrP. Sheep PrP fragment (amino acids 93 to 234) was modeled by using SwissModel (27) and the three-dimensional structures (PDB) of recombinant hamster PrP (2PrP) and mouse PrP (1AG2) as templates. The N terminus (N), C terminus (C), α-helices 1 to 3, and the separate polymorphic positions are indicated.

The allelic variant PrP^C_137T demonstrates conversion properties similar to those of PrP^C_136V in PrP^Sc_wt- and PrP^Sc_171H-induced conversion reactions (P_{different_136V} = 0.88 and 0.80, respectively). In PrP^Sc_136V-induced conversion reactions, however, the codon 137 polymorphism results in inconvertibility (P_{different_wt} < 0.01; P_{different_171R} = 0.37). Probably the neighboring polymorphisms at codons 136 and 137 interfere with the potential interaction site between PrP^C and PrP^Sc (Fig. 5). No in vivo data for this allelic variant are available since the allele is rare (only one healthy but culled sheep has been found to have this allele [6]). This allele is therefore irrelevant for breeding strategies and was excluded from the consensus linkage with scrapie susceptibility in Fig. 4.

The polymorphism Phe at codon 141 (PrP^C_141F) had minor reduction effects on the convertibility to PrP-res, and this variant reacted in the same order of magnitude as PrP^C_wt (P_{different_wt} = 0.77). This seems to be consistent with observations that sheep having the PrP^VRQ/PrP^{AF₁₄₁RQ} genotype have somewhat longer survival times than sheep having the PrP^VRQ/PrP^ARQ genotype, while both have longer survival times than PrP^VRQ/PrP^VRQ sheep in a flock of Swifter sheep where PrP^VRQ is linked to highest scrapie susceptibility and shortest survival times (6). Based on in vitro conversion efficiencies, the PrP^{AF₁₄₁RQ} allele is expected to be associated with high susceptibility in flocks where PrP^ARQ is associated with highest scrapie susceptibility. From the similar conversion efficiencies of PrP^C_wt and PrP^C_141F, but also from amino acid sequence comparisons between different species, the L141F polymorphism is expected to act fairly neutrally. In addition, L141M in hamster and mouse, L141V in brush-tailed possum, and L141I in human and gorilla are also found without any demonstrated linkage to TSEs, which indicates that some degree of evolutionary amino acid freedom at this position of the PrP is allowed. However, the neighboring polymorphism in goat PrP at codon 142 (Ile to Met) is associated with decreased susceptibility to experimental infections with scrapie and BSE (15), suggesting that the loop between β-sheet 1 and α-helix 1 (Fig. 5) is not without importance.

The observed low convertibility of the PrP^C_154H variant confirms previous speculations that this variant might also induce a certain degree of resistance to scrapie development (P_{different_171R} = 0.06, 0.15, and 0.71 for PrP^Sc_136V, PrP^Sc_wt, and PrP^Sc_171H, respectively). However, to our knowledge only two natural cases of scrapie in a PrP^AHQ/PrP^AHQ sheep have been found in the United Kingdom in a flock where PrP^ARQ is associated with highest susceptibility (M. Dawson, personal communication). This observation and the data for experimental challenge of sheep with the PrP^AHQ allele show that this allele does not induce such an absolute resistance to TSEs as the PrP^ARR allele (10, 31). This seems to be consistent with the slightly increased in vitro conversion efficiencies of the PrP^C_154H variant compared to the PrP^C_171R variant in vitro (P_different = 0.06). After intracerebral infections with the CH1641 scrapie isolate (predominant PrP^Sc_wt) or BSE, sheep with the PrP^AHQ/PrP^AHQ genotype, but not with the PrP^ARR/PrP^ARR genotype, developed scrapie with some unusual features (31). The 10-fold reduction in convertibility of PrP^C_154H compared to PrP^C_136V in the PrP^Sc_136V-induced reactions and the 5-fold reduction in convertibility of this PrP^C_154H variant compared to PrP^C_wt in the PrP^Sc_wt-induced reactions suggests that in vivo the PrP^AHQ allele may induce more resistance in flocks where PrP^VRQ is associated with highest scrapie susceptibility rather than in flocks where PrP^ARQ is associated with highest scrapie susceptibility. The codon 154 polymorphism is located in α-helix 1 (Fig. 5), and its charged side chain may interfere with the potential nucleation site between the two smaller β-sheets toward α-helix 1. In addition, this polymorphism induces a charge inversion comparable to that found in the resistant PrP^C_171R variant, indicating that the two polymorphisms may destabilize the hydrophobic core and/or the dipolar character of PrP by similar mechanisms.

The allelic variant PrP^C_171H has a significantly reduced convertibility compared to PrP^C_wt in conversion reactions induced by PrP^Sc_wt and PrP^Sc_171H (Fig. 3, bars 1 and 8; P < 0.01 and P = 0.02, respectively). The codon 171 Gln-to-His polymorphism is found mainly in Texel and some smaller cross-breeds. Several scrapie sheep with the PrP^VRQ/PrP^ARH genotype have been described (2), and thus far only a single scrapie-infected PrP^ARH/PrP^ARH sheep of the Dutch Zwartbles breed has been found (unpublished results). Based on the in vitro conversion data and observations that this allele in the homozygous PrP^ARH/PrP^ARH setting is almost absent in sheep with scrapie, it may be that the PrP^ARH allele introduces a certain degree of resistance to scrapie at least in a homozygous setting but has a more neutral role in heterozygous PrP^ARH genotypes.

The codon 211 polymorphism (Arg to Gln), located inside α-helix 3 (Fig. 5), reduces convertibility compared to the PrP^C_wt variant (P < 0.01). Polymorphisms at almost the same location in human PrP are all associated with spontaneous Creutzfeldt-Jacob disease or Gerstmann-Sträussler-Scheinkersyndrome (polymorphic codons 211 Arg to His, 213 Val to Ile, 215 Gln to Pro, or 220 Gln to Arg [sheep PrP numbering]) (17, 26, 30, 35). Under the cell-free conditions used, however, no detectable spontaneous PrP-res formation occurred in the PrP^C_211Q reaction (Fig. 1, lane 18). Although only a single (culled) sheep with this PrP^ARQQ211 allele has been found (2), this polymorphism is found as a common polymorphism in goats (H. M. Schatzl, unpublished results), and its introduction in sheep therefore might actually have its origin in goats. Since goats and sheep have identical wt PrP sequences, it is predicted that this polymorphism is associated with reduced susceptibility in goats as well.

Only three different types of PrP^Sc could be isolated from sheep homozygous for PrP^ARQ, PrP^VRQ, and PrP^ARH with natural scrapie. Since PrP^Sc could not be obtained from sheep homozygous for one of the other six PrP genotypes, no in vitro conversion data could be generated for these PrP^Sc types. Nevertheless, PrP^Sc from other homozygous types of PrP^Sc like PrP^Sc_141F or PrP^Sc_154H are of great interest. Conversion reactions using, for instance, PrP^Sc_154H may elucidate whether allelic variants linked to resistance remain inconvertible even when induced with homologous PrP^Sc. Such conversion reactions are important for the justification of future scrapie control programs which introduce TSE-resistant PrP genotypes. The set of nine conversion reactions induced by PrP^Sc_171H indicate that the most efficient reaction is not always the homologous one (Fig. 3c).

Since only one scrapie sheep with the PrP^ARH/PrP^ARH genotype was found, limited PrP^Sc was available for in vitro conversions. Although the data set for PrP^Sc_171H is limited, some interesting observations could be made. The relative conversion efficiencies of the various PrP^C variants by PrP^Sc_171H reflected most of the features of reactions induced with the PrP^Sc_wt or PrP^Sc_136V, for instance, the reduced convertibility of the PrP^C_154H and PrP^C_171R variants (P < 0.01), the reduced convertibility of the PrP^C_112T variant (P < 0.01), the comparable characteristics of PrP^C_141F and PrP^C_wt, and the overall high convertibility of the PrP^C_wt and PrP^C_136V variants. A rather unusual feature, however, is the relatively high convertibility of PrP^C_wt, PrP^C_136V and to lesser extent PrP^C_141F compared to the homologous PrP^C_171H variant (Fig. 3c, bars 1, 3, 5, and 8). This PrP^Sc_171H variant seems to be the first ever described variant of PrP^Sc inducing higher conversion efficiencies in heterologous PrP^C variants than in homologous variants. The molecular mechanism underlying this effect is still unclear, but it might be that as an intrinsic property, PrP^C_171H has less capability to convert to protease-resistant PrP. This PrP^Sc_171H conversion set therefore emphasizes that some polymorphic PrP^C variants (like PrP^C_171R and PrP^C_154H) are more or less prone to convert than others, independent of PrP^Sc type.

Some conversion reactions showed a large variation in conversion efficiency between independent PrP^Sc isolates (Fig. 3a, bar 1; Fig. 3b, bars 3 and 6). This variation might be linked to PrP^Sc isolate quality and/or it might be linked (hypothetically) to the PrP allele which in that particular flock is associated with high scrapie susceptibility. For example, PrP^Sc_wt isolated from PrP^ARQ/PrP^ARQ sheep residing in a flock of sheep in which the PrP^VRQ allele is associated with high scrapie susceptibility might be different from PrP^Sc_wt isolated from sheep in a flock in which the PrP^ARQ allele is associated with highest scrapie susceptibility. These structural differences between PrP^Sc isolates may cause shifted conversion profiles between the various PrP^C variants. For instance, PrP^Sc_wt from a PrP^VRQ background might convert PrP^C_136V better and PrP^C_wt less than PrP^Sc_wt from a PrP^ARQ background. This might indicate that if scrapie develops in sheep residing in a flock in which a different PrP allele is linked to highest risk of scrapie, the scrapie agent has overcome the polymorphism barrier (defined as a single amino acid mismatch between PrP^C and PrP^Sc reducing the conversion efficiency) but may not be completely adapted to the PrP genotype of the new host, and thus the isolated PrP^Sc might still be carrying “PrP^VRQ information.” Adaptation effects in strains when transmitted to different species have been theoretically predicted by numerical integration to require at least two or three passages for stabilization (23). Data on the isolation, transmission, and characterization of the CH1641 scrapie isolate indicate that these adaptation effects may also occur in vivo (11, 14). If these adaptation effects between the different PrP genotypes in vivo indeed occur, then it becomes even more important to know the PrP alleles in the direct environment of sheep before assessing their risk for scrapie development. Whether this hypothesis is true needs further investigation.

All of the different sheep PrP polymorphisms (except codon 211) are located in, or their side chains are directed to, the potential interaction site between PrP^C and PrP^Sc (Fig. 5). This interface, which has been predicted to be located between the two smaller β-sheets toward α-helix 1, might interact very specifically in the process of self recognition between PrP^Sc and PrP^C (PrP^C_136V is most efficiently converted by PrP^Sc_136V, while PrP^C_wt is most efficiently converted by PrP^Sc_wt). The polymorphism at codon 136 (Ala or Val), which influences convertibility or polymorphisms like Pro to Leu in human PrP at codon 102 (32), may modulate the interaction (recognition) between PrP^C and PrP^Sc. Such a recognition site might be a potential location in PrP for interference with PrP^Sc formation or TSE replication by targeting antibodies or synthetic peptides.

In summary, this study has found a clear effect of different polymorphisms in sheep PrP^C and to a lesser extent in sheep PrP^Sc on the in vitro convertibility of PrP^C to protease-resistant isoforms. The conversion data indicate that sheep with the PrP^VRQ, PrP^ARQ, and PrP^{AF₁₄₁RQ} alleles are at high risk for scrapie development. The alleles of PrP^AHQ, PrP^ARR, and to minor extent PrP^{T₁₁₂ARQ} seem to be linked to resistance to scrapie. Therefore, these alleles seem to be most suitable for introducing genetic resistance to scrapie by breeding strategies using positive selection for PrP^ARR and PrP^AHQ and negative selection for PrP^VRQ, PrP^ARQ, and PrP^{AF₁₄₁RQ}. It should be mentioned, however, that in vivo other factors such as dose, strain of TSE, route of infection, and efficiency of delivery are likely to be important in determining scrapie susceptibility and transmissibility. Future studies including BSE, sheep-passaged BSE, or PrP^Sc of other species may reveal which animals are most prone to carry or become infected with a cross-species TSE infection.