Molecular Typing of Protease-Resistant Prion Protein in Transmissible Spongiform Encephalopathies of Small Ruminants, France, 2002–2009

The agent that causes bovine spongiform encephalopathy (BSE) may be infecting small ruminants, which could have serious implications for human health. To distinguish BSE from scrapie and to examine the molecular characteristics of the protease-resistant prion protein (PrPres), we used a specifically designed Western blot method to test isolates from 648 sheep and 53 goats. During 2002–2009, classical non-Nor98 transmissible spongiform encephalopathy had been confirmed among ≈1.7 million small ruminants in France. Five sheep and 2 goats that showed a PrPres pattern consistent with BSE, or with the CH1641 experimental scrapie source, were identified. Later, bioassays confirmed infection by the BSE agent in 1 of the 2 goats. Western blot testing of the 6 other isolates showed an additional C-terminally cleaved PrPres product, with an unglycosylated band at ≈14 kDa, similar to that found in the CH1641 experimental scrapie isolate and different from the BSE isolate.

T ransmissible spongiform encephalopathies (TSEs) are a group of fatal neurodegenerative diseases that include scrapie in sheep and goats, bovine spongiform encephalopathy (BSE) in cattle, and Creutzfeldt-Jakob disease (CJD) in humans (1,2). TSEs are characterized by accumulation in the brain of a disease-associated isoform (PrP d ) of a host-encoded cellular prion protein (PrP c ) (3). PrP d , in comparison with the normal prion protein PrP c , clearly differs in secondary and tertiary structures (4,5) and in bio-chemical characteristics (6). Proteinase K (PK) digestion destroys PrP c , but in PrP d it generates a protease-resistant fragment known as PrP res . Most TSE diagnostic methods (e.g., ELISA and Western blot tests) are based on detection of PrP res (7).
The transmissible agent involved in BSE in cattle is known to cause prion diseases in other species under natural conditions (8). BSE can also be experimentally transmitted to sheep and goats, including after oral challenge to test for transmission (9). Because BSE-contaminated meat and bone meal may have been fed to small ruminants, BSE may have been transmitted to sheep or goats. Also, the Scientifi c Steering Committee of the European Commission has hypothesized that the BSE agent might have originated from a scrapie agent in sheep or goats and that these animals may represent a reservoir (10). In view of these data, the European Commission defi ned a strategy to investigate the possible presence of BSE in sheep and goats under natural conditions (11).
The standard for strain typing TSE agents is based on analysis of the phenotypic characteristics of the disease after transmission in laboratory rodents. Biological characterization of the BSE agent in inbred wild-type mice appeared to be reliable, because it showed uniform features in mice (8). However, this approach is time-consuming and costly. The identifi cation of uniform molecular features of PrP res by Western blot in human variant CJD paved the way to a similar approach for detecting possible BSE in small ruminants (12). The molecular criteria defi ned in these studies included electrophoretic mobilities, glyco-sylation characteristics, and immunolabeling with different monoclonal antibodies (13). The last criteria enabled mapping of the protease cleavage site of the PrP protein fragment obtained after PK digestion (14). More recently, the identifi cation of additional C-terminal PrP res products may contribute to discrimination of the different types of CJD (15) or of different scrapie and BSE sources (16,17). Discriminant molecular features of the prion protein can also be investigated by immunohistochemical analysis (18) or ELISA (19). In all of these studies, it was assumed that the strain information was closely associated with the structural features of PrP d .
The Western blot method enabled discrimination of experimental BSE in sheep from most scrapie-affected animals (12,13,(20)(21)(22)(23)(24). Nevertheless, discrimination was more diffi cult with the CH1641 experimental scrapie isolate (21,25), which otherwise clearly differs from BSE by its absence of transmissibility to wild-type mice (26). Similar molecular features to those of CH1641 have been described in a few natural scrapie cases in France (27) and in the United Kingdom (24). We describe the molecular fi ndings obtained for a large series of TSE infections in France identifi ed in small ruminants by active surveillance during 2002-2009 and for CH1641-like isolates in sheep and in 1 goat.

Animals and Tissues
Two samples of BSE in small ruminants were used as controls, 1 (SB1) from a sheep experimentally infected by BSE with a brain homogenate from a BSE-affected cow from France (21), and the other (CH41x76) from a goat that had been intracerebrally inoculated with a brain homogenate from cattle BSE from Britain (28).
Two experimental scrapie sources in sheep (provided by N. Hunter, Institute for Animal Health, Edinburgh, UK) were also studied. These were the SSBP/1 scrapie isolate, experimentally maintained by serial passages in sheep (29), and CH1641 that was derived from a Cheviot sheep and then maintained by serial passages in sheep (26).
Natural TSE isolates in small ruminants (Table) were mainly obtained through an active surveillance program in France. In operation since 2002, this program involves the random selection and testing of samples from rendering plants and slaughterhouses (30). Some samples originated from the passive surveillance program, which involves collecting samples from animals showing suspected clinical signs of the disease. Brain stems from all selected animals are subjected to a rapid test, and all reactive samples are then retested in the National Reference Laboratory by confi rmatory methods based on Western blot (31). When a sample is confi rmed as TSE positive, it is classifi ed as similar, or not similar, to Nor98 (atypical scrapie), according to the PrP res molecular profi le, which in atypical scrapie shows 5 major bands (31). Samples that are not similar to Nor98 (classical TSE), i.e, showing a 3-band pattern between 19-30 kDa, are then studied by using a Western blot discriminatory method to identify possible similarities with BSE (11).

Western Blot
All confi rmed classical TSE (non-Nor98-like) cases were analyzed to determine whether the PrP res looked similar to that found in bovine BSE (i.e., showed a lower molecular mass than that found in bovine BSE) or, on the contrary, was similar as in most scrapie cases. PrP res extracts were obtained from animal brains by using Bio-Rad protocol (the TeSeE Western Blot kit, ref: 355 1169; Bio-Rad, Marnes-la-Coquette, France), which is used to confi rm suspected TSE-positive samples and includes PK digestion and rapid protein precipitation steps. Denatured samples were then loaded on two 15% bis-polyacrylamide gels with the same gel plan. After electrophoresis (200 volts for 80 min), the proteins were transferred onto a nitrocellulose membrane (R-Biopharm, St. Didier au Mont d'Or, France). Blocking was performed for 1 h in 5% (wt/vol) milk powder in phosphate-buffered saline (PBS) containing 0.1 % (vol/vol) Tween-20 (PBS-T) for the membrane to be treated with monoclonal antibody Bar233 (1/5000) (144-FGNDYEDRYYRE-155 ovine PrP sequence) (provided by J. Grassi, C.E.A., Saclay, France) or in 3% (wt/vol) bovine serum albumin solution (Sigma, St. Quentin-Fallavier, France) in PBS-T for the membrane to be treated with monoclonal antibody P4 (0.2 mg/mL) (93-WGQGGSH-99 ovine PrPsequence) (R-Biopharm).
Both antibodies were incubated on the membranes for 30 min at room temperature. The membranes were then washed for 20 min in PBS-T and incubated with a solution of streptavidine-peroxidase-conjugated antimouse immunoglobulin (Southern Biotech distributed by Clinisciences, Montrouge, France) in PBS-T for 20 min at room temperature. The membranes were then washed for 30 min in PBS-T and for 5 min in PBS before detection by use of enhanced chemiluminescent substrate (Amersham Biosciences, Orsay, France). The signals were identifi ed on autoradiographic fi lms (Amersham) after a 3-min exposure. Quantitative studies were performed by using Quantity One software (Bio-Rad), and the apparent molecular masses were determined by comparing the positions of the PrP res bands with a biotinylated marker (B2787) (Sigma). All samples were compared, during molecular characterization of fi eld isolates, by expressing the molecular mass of the unglycosylated band in terms of differential molecular mass (dmm). The dmm corresponds to the difference measured between the test sample and the control cattle BSE sample always loaded beside it.

Active Surveillance Findings during 2002-2009
Since active surveillance of TSEs in small ruminants began in France in 2002, a total of 1,231 small ruminant (1,153 sheep and 78 goats) samples have been confi rmed as TSE-positive by Western blot using Sha 31 antibody; >1.7 million animals have been tested by rapid tests (1,152,065 sheep and 685,634 goats) ( Table). Nearly half of the positive isolates (504) have been identifi ed as Nor98 scrapie isolates, on the basis of detection of 5 major bands, including a prominent ≈10-12 kDa band detected by Western blot (32). After confi rmatory Western blot, insuffi cient quantities of brain tissue were available for Western blot discriminatory testing in 26 samples, which are shown in the Table  as samples not analyzed.
The other classical TSE cases, i.e., non-Nor98 cases (648 sheep and 53 goats), which typically showed a 3-band pattern between 19-30 kDa, have been further characterized by applying a discriminatory Western blot method, described as the Agence française de sécurité sanitaire des aliments (French Food Safety Agency) discriminatory method in the Technical Handbook for National Reference Laboratories (33). This method enables rapid identifi cation of PrP res patterns similar to those found in experimental ovine BSE and is essentially based on comparison of the PrP res molecular mass with that of cattle BSE, and comparative labeling with 2 antibodies, P4 and Bar233, against either the N terminal end or core part of the PrP res protein, respectively.

Molecular Characterization of Experimental Isolates
We fi rst analyzed 4 reference experimental isolates in small ruminants: a sheep and a goat with BSE, 2 sheep infected with SSBP/1 or CH1641 scrapie sources, and a bovid with classical BSE. The mean molecular masses of the diglycosylated (H), monoglycosylated (L), and unglycosylated (U) PrP res bands (using Bar233 antibody) and the P4/ Bar233 differential labeling, as well as the proportions of glycoforms, are shown in Figure 1 and in Figure 2, panel A. These analyses show the lower molecular mass of the unglycosylated PrP res in BSE in sheep (-0.6 kDa) and 1 goat (-0.5 kDa), and in CH1641 scrapie (-0.8 kDa), compared with the cattle BSE ( Figure 1; Figure 2, panel A). In contrast, SSBP/1 showed a higher apparent molecular mass (+0.6 kDa). The molecular masses of the 3 bands obtained for CH1641 (H: -0.8 kDa, L: -0.3 kDa and U: -0.2 kDa) were lower than in BSE in sheep or goats. Differential labeling by Bar233 and P4 antibodies was correlated with the molecular masses of the protease-resistant cores, with high and low Bar233/P4 ratios in SSBP/1 and BSE in sheep, respectively ( Figure 1). Like BSE in sheep, CH1641 showed a decreased signal with P4 compared with Bar 233. bovine BSE showed no labeling with P4 antibody.
Experimental BSE in a goat showed the same molecular features as those observed in experimental BSE in sheep. Both molecular masses, P4 immunolabeling and glycoforms proportions, were involved ( Figure 1; Figure 2, panels A, D).

Molecular Characterization of Natural TSE Sheep and Goat Isolates
We then compared the cattle BSE control with the 701 classical samples available for further analyses (648 sheep and 53 goats). Most of the isolates tested (657) were PrP res positive by discriminatory Western blot using Bar233 core antibody and showed a 3-band pattern, whereas the signal was weaker in 44 samples, preventing the identifi cation and characterization of the unglycosylated band.
We chose to express the molecular masses measures as dmm, which corresponds to the difference of molecular masses measured between the tested sample and the control cattle BSE always loaded beside it. Most (650) samples showed a positive dmm ( Figure 3) and strong labeling with P4 antibody, as previously described for SSBP/1. However, large variations in molecular mass (1.2 kDa) were observed among these samples, a possible clue pointing to biological diversity among scrapie sources; in contrast, only small variations were observed after repeated measures of a same sample with the Western blot method, as shown by the small standard deviations during repeated analysis of reference samples (Figure 2). A minority of the samples (12) showed similar molecular masses to bovine BSE (dmm <0.1 kDa difference). Stronger P4 labeling compared with Bar 233 antibody was observed in all these samples and in the 44 (not defi ned) samples for which the dmm could not  be measured with Bar233 antibody because quantities of extracted PrPres were too low. A few samples (5 sheep and 2 goats) showed a lower PrP res molecular mass than the bovine BSE control, i.e., a negative dmm (Figure 3), combined with poor labeling by P4 antibody, as described for BSE in small ruminants and CH1641 scrapie. These samples were suspected to represent BSE in small ruminants and were thus further characterized by repeated Western blot analyses in comparison with experimental controls.

Molecular Studies of TSE Isolates with Low Molecular Mass PrP res
These experiments allowed repeated measures of PrP res molecular masses and P4/Bar233 differential labeling and estimation of the proportions of PrP res glycoforms. Western blot analyses were also performed with the C-terminal SAF84 antibody to identify possible additional C-terminal PrP res cleavage products, as described for the CH1641 experimental isolate (17).
The 3 PrP res glycoforms of the 5 sheep isolates showed similar molecular masses close to BSE in sheep or CH1641 scrapie (Figure 2, panel B; Figure 4, panel A). These similarities of PrPres migration were associated with intermediate labeling with P4 but nevertheless varied considerably (up to 3×) according to the isolate (data not shown). Analyses of the glycoform proportions showed lower proportions of the diglycosylated band, compared with BSE in sheep, and also similarity to CH1641 in this respect (Figure 2, panel E). All 5 isolates showed an additional ≈14 kDa PrP res band, after use of the SAF84 antibody ( Figure 4, panel C), that in 4 isolates represented 30 ± 7% to 34 ± 7% of the total signal after repeated analyses and PNGase deglycosylation. Among these sheep isolates, the 06-412 sample showed higher molecular masses and proportion of diglycosylated band. Nevertheless, the 5 natural isolates thus appeared similar to the CH1641 experimental scrapie isolate.
For the 2 goat samples with low molecular mass PrP res , the molecular characteristics of the CH636 isolate were indistinguishable from BSE in goat (CH41x76), with regard to PrP res molecular masses ( Figure 5, panel A) and the respective proportions of the diglycosylated/monoglycosylated bands (70 ± 9%/22 ± 6% and 71 ± 8%/20 ± 4%, respectively) (Figure 2, panel F) when Bar233 antibody was used, and reduced labeling by P4 antibody compared with Bar233 ( Figure 5, panel B). A similar pattern was observed with the other goat sample 08-357, but the levels of the diglycosylated band were even higher (83 ± 2%). However, Western blot analyses with SAF84 antibody showed that these 2 goat samples clearly differed and had an additional ≈14 kDa PrP res band only in the 08-357 isolate, as in the CH1641 experimental scrapie isolate. This ≈14 kDa PrP res band was represented in the same proportions as in the CH1641-like sheep isolates. This sample could not be clearly distinguished from BSE in goat according to other molecular criteria.

Discussion
We investigated the PrP res molecular features of one of the largest series of natural TSE isolates from sheep and goats analyzed so far in Europe. Approximately 1.7 million small ruminants were subjected to a rapid test; among these, 1,153 sheep and 78 goats originating from 992 fl ocks were confi rmed as TSE-positive on the basis of identifi cation of PrP res in the brain stem. Another large study using 1,247 sheep originating from 450 fl ocks in Great Britain has been reported (24). Similar studies have been conducted in other European countries such as Germany (34), the Netherlands (14,) and Italy (23), but fewer TSE-positive animals were reported. In our series, animals with classical cases represented 53% of the TSE-affected animals. The molecular features of most of these isolates (99%) were comparable to those previously described for most scrapie cases, in studies in Europe or France (21). In all of these cases, PrP res from sheep and goats showed clearly distinct features from BSE, based on the identifi cation of a higher molecular mass of unglycosylated PrP res , associated with strong labeling by the P4 antibody that recognizes the N terminal end of the protein. However, our observation of large individual variations in this PrP res molecular mass implies that a possible relationship with the biologic diversity of TSE agents, which has been described after transmission of scrapie to inbred wild-type mice (35), would be questionable. Only a few animals (5 sheep and 2 goats) in this large series of TSE-affected animals showed molecular characteristics that, in comparison with the usual features of scrapie, could be consistent with the known BSE signature in small ruminants. These samples represented all the suspected isolates that were identifi ed by offi cial surveillance in France during 2002-2009. An essential molecular criterion defi ned from the observation of PrP res BSE compared with scrapie was the low molecular mass of the unglycosylated band in PrPres BSE, associated with a decreased PrP res signal lower with an N terminal antibody than with a core antibody (13,22,36). After identifi cation of these molecular features in a few small ruminants, only 1 (CH636) of the 2 cases here described in goats, identifi ed as TSE positive in 2002, has been shown to be infected by the BSE agent after bioassays in mice (37). Another goat in the United Kingdom identifi ed by an immunohistochemical discriminatory method as TSE-positive in 1990 showed characteristics that were indistinguishable from BSE (18). These results clearly indicate that in a situation characterized by a decrease in the number of cases in cattle in all countries in Europe, the possibility of fi nding BSE in small ruminants is now remote.
The other unusual isolates showed molecular characteristics that were partly similar, not only to BSE in small ruminants with a low molecular mass of PrP res and faint labeling with P4 antibody, but also to the CH1641 experimental scrapie isolate. However, detailed immunohistochemical investigations of CH1641 showed subtle differences in the cleavage site of the protein compared with BSE in sheep (38). As previously described, after transmission in ovine transgenic mice (17), the slightly lower PrP res molecular mass in CH1641-like isolates, as recognized in the CH1641 experimental isolate (13),was confi rmed by Western blot, at least in sheep (Figure 2, panel B). However, these differences in molecular mass are more easily identifi ed on the diglycosylated band. Compared with BSE in small ruminants, lower proportions of this diglycosylated band were found in sheep, whereas the 08-357 goat sample showed very high levels of this diglycosylated band, which would be consistent with BSE. Experimental transmissions of BSE in sheep have shown that, to a certain extent, the PrP res molecular features could be infl uenced by different factors, such as serial passages in sheep (39) or sheep genotypes (36), although slight variability did not compromise the discrimination with scrapie. Furthermore, all these CH1641like natural isolates in sheep and goats clearly differed from BSE by the presence of an additional, C-terminally cleaved, PrP res product specifi cally recognized by a Cterminal antibody (SAF84), as previously described for the CH1641 experimental scrapie isolate (17). Baron  goat case represents, to our knowledge, a spreading of the known species range for natural CH1641 infection. At least in sheep, for which 4.34 cases per million sheep tested were identifi ed in this study, the frequency of CH1641like scrapie was notably higher compared with other rare TSEs in ruminants such as atypical BSEs, which showed a frequency of 0.76 per million cattle tested during 2001-2007 (40). Thus, large-scale testing of animals would be required to identify these rare TSE isolates. Similar iso-lates were only identifi ed in sheep in the other large series reported from 450 fl ocks in Great Britain (2 cases in 1 fl ock) (23) and in a previous study of 214 TSE-infected sheep in France (2 cases in 1 fl ock) (27). However, an underestimation of the frequency of such cases cannot be fully excluded. PrP res features are assessed by analyzing a single homogenate prepared from a brain fragment from the animal. Stack et al. described a case in sheep that appeared as CH1641-like after repeated Western blot analysis of a brain stem sample, whereas previous analysis of the caudal medulla at the time of submission had shown the usual scrapie profi le (24). Immunohistochemical test-  ing of 2 CH1641-like cases in sheep showed, that unlike BSE, PrP d could be clearly identifi ed by using P4 antibody in some of the brain stem nuclei and in lymphoid tissues (27). Finally, on the basis of identifi cation of low levels of C-terminal PrP res product in ovine transgenic mice infected with usual scrapie isolates, we hypothesized that a CH1641-like component might be present as a minor component in these scrapie cases that showed usual molecular features (17). All these data raise the question of the existence of possible mixtures of TSE agents in these particular CH1641-like isolates.