H-Type Bovine Spongiform Encephalopathy

Abstract

We previously reported that some cattle affected by bovine spongiform encephalopathy (BSE) showed distinct molecular features of the protease-resistant pion protein (PrPres ) in Western blot, with a 1-2 kDa higher apparent molecular mass of the unglycosylated PrPres associated with labelling by antibodies against the 86-107 region of the bovine PrP protein (H-type BSE). By Western blot analyses of PrPres, we now showed that the essential features initially described in cattle were observed with a panel of different antibodies and were maintained after transmission of the disease in C57Bl/6 mice. In addition, antibodies against the C-terminal region of PrP revealed a second, more C-terminally cleaved, form of PrPres (PrPres #2), which, in unglycosylated form, migrated as a ≈ 14 kDa fragment. Furthermore, a PrPres fragment of ≈ 7kDa, which was not labelled by C-terminus-specific antibodies and was thus presumed to be a product of cleavage at both N- and C-terminal sides of PrP protein, was also detected. Both PrPres #2 and ≈ 7 kDa PrPres were detected in cattle and in C57Bl/6 infected mice. These complex molecular features are reminiscent of findings reported in human prion diseases. This raises questions regarding the respective origins and pathogenic mechanisms in prion diseases of animals and humans.

Introduction

Transmissible spongiform encephalopathies (TSE) are fatal neurodegenerative diseases, affecting both humans (Creutzfeldt-Jakob disease; CJD) and animals, such as sheep and goats (scrapie), deer and elk (chronic wasting disease; CWD) and cattle (bovine spongiform encephalopathy; BSE). Although the nature of the infectious agent causing these diseases still remains controversial,1–3 a central event in their pathogenesis is the accumulation of an abnormal form of the host-encoded prion protein, in infected tissues. Under conditions where the normal cellular protein (PrP C) is fully sensitive to proteases, the disease-associated prion protein (PrP^d) is only degraded to a partly resistant fragment (PrP^res) by removal of the amino-terminal end of PrP^d. The sizes of PrP^res fragments and their N-terminal amino acid residues of PrP^res may vary between different prion strains.4–10

Experimental transmissions of these diseases revealed a biological diversity, reminiscent of strains of classical infectious agents such as viruses.11 This has mainly been described in genetically defined inbred wild-type mice by the different features of the disease, which include differences in incubation periods and in the distribution of brain lesions.12 Different and specific features of the PrP^res protein have also been found in mice or hamsters infected with different biological strains of prion diseases, suggesting that biological properties of the infectious agent might be enciphered in the conformation of the protein.4–10

It was believed until recently that TSEs in cattle were caused by a single strain of infectious agent identified as responsible for the food-borne BSE epidemic. Characterization of the infectious agent associated with BSE indeed showed unique features, with well-defined incubation periods and distribution of brain lesions following transmission of the disease to wild-type mice, not only directly from cattle, but also, after natural or experimentally-induced cross-species transmissions.13,14 The surprisingly uniform features of the disease in cattle were also illustrated by analysing the distribution of neurodegenerative brain lesions in affected cattle, at different places and periods of the BSE epidemic.15,16 However, twenty years after identification of the first cases in cattle, the origin of the actual BSE agent remains controversial.17,18 The possible recycling of an infectious agent derived from prion diseases found naturally in other species, such as scrapie in sheep and goats, has been the most famous hypothesis, but the possible existence of a previously unidentified “sporadic” form of prion disease in cattle, as are most cases of Creutzfeldt-Jakob disease in humans, has also been proposed.19

Western blot analyses of PrP^res accumulating in the brains of BSE-infected animals and humans have demonstrated specific molecular features, compared to most other forms of prion diseases, with a low molecular mass of the unglycosylated PrP^res moiety and high proportions of diglycosylated PrP^res.8 Recent studies however have reported cases of prion abnormalities in cattle with distinctly deviant PrP^res features.20,21 We previously described three French cattle isolates essentially characterized by a higher apparent molecular mass of unglycosylated PrP^res (later called H-type isolates) and decreased levels of diglycosylated PrP^res, in comparison to BSE.20 Also, only PrP^res from H-type isolates was labelled by monoclonal antibody P4 with defined PrP^res N-terminus epitope specificity, in contrast with the PrP^res from BSE isolates, suggesting a different cleavage of the protein by proteinase K (PK).20 We recently demonstrated that such H-type isolates of bovine prion diseases were transmissible to wild-type mice, and that these specific features initially described in cattle as distinct from typical BSE, were maintained in mice.22 Similar results were also obtained after transmission in bovine and ovine transgenic mice, favouring the hypothesis that a distinct strain of prion should be involved in H-type BSE.23

We now further describe the existence of complex molecular features of these H-type cases in cattle, by Western blot analyses of PrP^res using a panel of anti-PrP antibodies, and demonstrate the transmissibility of these features after transmission of the disease in C57Bl/6 mice. These data are discussed with respect to their similarities with different prion diseases in humans.

Materials and Methods

Cattle TSE isolates.

The H-type BSE isolates used in the mouse transmission experiments in this study included two cases which showed, as previously described with RB1 polyclonal antibody, a higher apparent molecular mass of unglycosylated PrP^res, as well as labelling by P4 monoclonal antibody.20 A typical BSE case was used in comparison in both sets of mouse transmission experiments. Main features of these three cases, identified from active surveillance at rendering plants, were previously reported.21

Mouse inoculations and follow-up.

Four to six weeks old female C57Bl/6 mice (Charles Rivers, L'Arbresle, France) were inoculated intracerebrally, following anaesthesia, with 20 ml per mouse of a 10% (wt/vol) homogenate prepared from cattle brain stem samples in glucose (5%) using disposable blenders. Mice were cared for and housed according to the European guidelines (directive 86/609/EEC) and of the French Ethical Committee (decree 87–848), and were supplied with food and drink ad libitum. Following experimental challenge, experiments were conducted in the Biohazard prevention area (A3) of the author's institution with the approval of the Rhône-Alpes Ethical Committee for Animal Experiments. The mice were then checked twice a week for the presence of clinical signs, such as tremors, plastic tail, coarse coat, abnormal gait, hind limb paralysis, clasping feet, hunched posture, bradykinesia or leanness. The animals were sacrificed by anaesthetic solution overdose when they exhibited any signs of distress or confirmed evolution of clinical signs of prion disease; in a few cases they were found dead. The whole brain of every second mouse was frozen and stored at -80°C before Western Blot analysis and the other brains were fixed in paraformaldehyde 4% for histopathological studies.

Western blot analysis.

(i) Extraction of PrP^res. PrP^res was extracted from bovine brain stem samples using the TeSeE® Western blot Bio-Rad kit (Ref 355 1169) following the manufacturer's instructions. Briefly, 250µl of 20% brain homogenate were incubated with an equal volume of reconstituted proteinase K solution (reagent A + PK) at 37°C for 10 min. After adding 250 µl of reagent B, samples were centrifuged at 15,000 x g for 7 min. The pellets were resuspended and heated for 5 min at 100°C in 50 µl of denaturing buffer (TD4215)(4% sodium dodecyl sulfate, 2% β-mercaptoethanol, 192 mM glycine, 25 mM Tris and 5% sucrose) and centrifuged at 12,000 g for 15 min. Finally, the pellets were discarded and the supernatants were run on a sodium dodecyl sulphate polyacrylamide gel. Mouse brain tissues representing half of the whole brain after sagittal section were homogenized in a 5% glucose solution (10% wt/vol). The homogenates were forced through a 0.4-mm diameter needle before incubation for 1h at 37°C with proteinase K (PK) (10 µg/100 mg of brain tissue). Samples were then incubated for 15 min with N-lauroylsarkosyl (final concentration 10%) and centrifuged at 100,000 rpm for 2h on a 10% sucrose cushion (Beckman TL100 ultracentrifuge). The pellets were resuspended and heated in denaturing buffer as previously described, before SDS-PAGE. (ii) Western Blotting procedures. Proteins were separated by SDS-PAGE on 15% polyacrylamide gels and transferred to 45 µm nitrocellulose membranes (Amersham) in a 25 mM Tris, 192 mM glycine, 10% isopropanol buffer at 400 mA constant for 1h. For immunoblotting, the membranes were blocked for 1h with fat milk 5% (12B2, SAF84), TeSeE® Western blot Bio-Rad kit Blocking Solution (Sha31) or 3% bovine serum albumin in PBS-Tween 20 (0.1%) (PBST)(other antibodies). They were then incubated for 30 min at room temperature with antibodies against different PrP domains, as summarised in Table 1. After washes in PBST, the membranes were incubated 30 min at room temperature with peroxidase-labelled conjugates diluted in PBST against mouse IgG (TeSeE® Western blot Bio-Rad conjugate (1/10) for Sha31 antibody or anti-mouse IgG (H+L) from Clinisciences (ref 1010-05)(1:2500) for other monoclonal antibodies) or against rabbit IgG (1:2500)(Clinisciences ref 41010-05) for polyclonal antibody RB1. Streptavidin (5 ng/ml)(S5512, Sigma) was added to the conjugate solution. The membranes were finally washed three times in PBST and once in PBS, and bound antibodies were detected by enhanced chemiluminescence ECL (Amersham) or Supersignal (Pierce) and visualized either on film (Biomax, Kodak) or directly in an image analysis system for quantitative measurements (Versadoc, Biorad). (iii) Analyses. For quantitative studies of the glycoform ratios, chemiluminescent signals corresponding to the three glycoforms of the protein were measured by using Quantity One (Biorad) software. Glycoform ratios were expressed as mean percentages (± standard deviations) of the total signal from at least three separate gel runs. Apparent molecular masses (M_r) of PrP^res glycoforms were determined as the average of the centre positions of the bands from at least 3 repeated runs, measured in comparison with a biotinylated marker (B2787, Sigma) run on each gel. In experiments for quantitative measures of the apparent molecular masses and glycoform ratios, PrP^res from C57Bl/6-infected mice infected with C506M3 scrapie strain was also used as a control.4 Immunological reactivities of the most N-terminus-specific antibodies (P4, 12B2, SAF32, 4F2) on samples were assessed by visual examination and quantitative image analysis, in comparison with analyses run in parallel on the same samples using antibody Sha31.

PNGase treatment.

Deglycosylation was performed using PNGase F (kit P07043, BioLabs). Denatured samples of PrP^res in TD4215 buffer (1–2 µl) were mixed with denaturing buffer from the PNGase kit, G7 buffer, NP40 and PNGase according to the manufacturer's instructions. After incubation at 37°C for 1 h, samples were ready for Western blot analysis following appropriate dilution in TD4215 buffer.

Results

PrP^res characterization in cattle TSE isolates.

We characterized the H-type cattle BSE isolates, in comparison with typical BSE, using a panel of antibodies against different regions of the PrP protein (Table 1). Results showed a quite different Western blot pattern according to the antibodies used (Fig. 1A–C).

The first group of antibodies (group A), which include antibodies recognizing the most N-terminal region of the PrP^res (amino-acids 62–107), exhibited an unusual three band pattern at ≈28, 23, and 19 kDa in H-type BSE, whereas the staining of PrP^res in typical BSE, in contrast, was very poor (Fig. 1A). In addition, low levels of a small ≈7 kDa PrP^res fragment could be detected following overexposure of the membranes (arrowhead in inset of Fig. 1A).

A second group of antibodies (group B), against epitopes in the 110–164 amino-acids region, recognized the same PrP^res pattern with almost the same apparent molecular masses (M_r) as the group A antibodies. The group B antibodies however did also stain the three PrP^res bands in typical BSE. The PrP^res pattern recognized by group B antibodies in the H-type isolates was however characterized by a 1–2 kDa higher M_r of each of the three PrP^res bands in comparison to those of the PrP^res in typical BSE (Fig. 1B).

A third group of antibodies (group C), against the more C-terminal part of the PrP protein (154–236), showed a strikingly different pattern in H-type isolates from that obtained with the group A and B antibodies (Fig. 1C). This pattern was characterized by (i) the presence of four bands at 28, 23, 19 and 14 kDa (respectively referred as bands I to IV), of which those at 23 (II) and 19 (III) kDa are quite broad (ii) the predominance of the band II, in contrast to the typical BSE isolate that showed a clear predominance of the diglycosylated PrP^res band (see bracket in Fig. 1C), and (iii) the width of the band III (see arrow in Fig. 1C), with diffuse labelling below the position of the well-defined 19 kDa band as recognised by groups A or B antibodies. Taken together, in H-type BSE, the group A and group B antibodies show similar PrP^res binding to a triple-band PrP^res population in the M_r 19–30 kDa range, whereas the group C antibodies reveal the presence of more complex pattern with four main bands, including also a lower band at 14 kDa. Such an unusual pattern with the additional ≈14 kDa band was not observed using SAF84 (group C) antibody in any of more than 200 cases otherwise characterized, after analysis using 6H4 (group B) antibody, by PrP^res molecular weights of typical BSE.

Analysis of PrP^res after deglycosylation by PNGase treatment first confirmed the higher PrP^res apparent molecular mass (≈19 kDa) in H-type BSE compared to typical BSE (Fig. 2B), as well as the more intense labelling using group A antibodies (Fig. 2A). In addition the group C antibodies specifically recognized an additional ≈14 kDa in H-type BSE (Fig. 2C). This strongly suggests that PrP^res from H-type isolates consists of a mixture of two populations - PrP^res#1 (producing the ≈ 19 kDa unglycosylated band) and PrP^res#2 (producing the ≈ 14 kDa unglycosylated band). Regarding PrP^res#2, (i) a diglycosylated form derived from the ≈ 14 kDa band would overlap with the monoglycosylated band of PrP^res #1, and (ii) a monoglycosylated form derived from the ≈ 14 kDa band could lead to the more diffuse labelling observed below the position of the ≈ 19 kDa unglycosylated band of PrP^res#1. As shown in (Fig. 1D), this interpretation is consistent with the unusual and complex Western profiles specifically observed with group C antibodies prior to deglycosylation.

PrP^res characterization in C57Bl/6 mice infected with H-type and typical BSE isolates.

The molecular features associated with H-type isolates from cattle were further studied in brain from wild-type mice (C57Bl/6) inoculated with either of the two bovine H-type isolates, in comparison with a typical BSE isolate. Most C57Bl/6 mice were positive for PrP^res by Western blot. However, the PrP^res levels were significantly lower in mice infected with H-type isolates than in BSE-infected mice since, when group B antibodies were used, four times as much brain tissue from H-type infected mice was needed per lane to obtain equivalent levels of PrP^res signals to those found in mice infected with typical BSE (Fig. 1F). In these latter conditions, and in relation to the recognition of an additional PrP^res form (PrP^res #2) with SAF84, the detected PrP^res signal was strongly increased in H-type infected mice compared to BSE-infected mice (Fig. 1G).

PrP^res from Western blot-positive mice was further characterized for comparison with the molecular features previously described in cattle, by analysing the PrP^res staining patterns with antibodies representative of the 3 above-defined groups. Using the same immunological reagents as in studies with bovine tissues, analysis of the ≈14 kDa region of the gel was hampered by an unspecific immunoreactivity in this region, which was also detected in uninfected mouse brain (Fig. 1H). This was not linked to PrP since it is also detected in the absence of anti-PrP primary antibody (Fig. 1H) and from PrP 0/0 mouse brain tissue (data not shown).

With group A antibody 12B2, the H-type mouse PrP^res was stained, but that of typical BSE only very poorly (Fig. 1E) and when the group B antibody Sha31 was used, the PrP^res in C57Bl/6 mice inoculated with H-type isolates revealed a 1–2 kDa higher M_r for the three PrP^res bands compared to those observed in BSE-infected mice (Figs. 1F and 3A). The differences in M_r of the PrP^res bands observed with group B antibodies were maintained after deglycosylation by PNGase (Fig. 2E). These data are consistent with a PrP^res #1 population as observed in cattle with group A and B antibodies, and show that the PK resistance of the PrP region recognized by group A antibodies was specifically preserved in mice infected with H-type isolates, whereas it was largely removed from BSE-infected mice. In addition to the triple band pattern of PrP^res, a conspicuous ≈7 kDa band was also identified using 12B2 and even more strongly with Sha31 (Fig. 1E and F), which could also be identified after deglycosylation (Fig. 2E). The labelling in the ≈14 kDa region that was clearly apparent with 12B2 antibody remained consistent with the unspecific labelling found in normal mice, even in the absence of anti-PrP primary antibody.

Using group C antibodies (SAF84 and R524) the same distinct 4-band pattern was observed in mice as described in cattle (compare Fig. 1G with Fig. 1C). Although the ≈ 14 kDa in mice might at least in part result from unspecific labelling, this pattern was characterized by the two other salient features of the SAF84 WB profile in H-type BSE in cattle, namely (i) the predominance of band II (≈40% of signal of bands I–IV together, compared to ≈30% for band I) (Fig. 1G) (see bracket in Fig. 1G) and (ii) the width and diffuse aspect of band III (see arrow in Fig. 1G). All together these data are consistent with the presence of PrP^res #2 in mouse brains, as described in cattle. Accordingly, an ≈14 kDa PrP^res form, corresponding to unglycosylated PrP^res #2, was indeed clearly detected with group C antibody SAF84 after PNGase deglycosylation, in addition to the unglycosylated PrP^res #1 at ≈19 kDa (Fig. 2F).

Glycoform analyses with Sha31 antibody showed comparable proportions (≈40% of the 19–30 kDa signal) of the diglycosylated and monoglycosylated bands in H-type BSE (Fig. 3B), the ratios of which were closer to those observed in a mouse-adapted scrapie strain (C506M3) than in mice infected with typical BSE (≈55% of diglycosylated band).

Discussion

Our study provides new information regarding the molecular characterization of one of two deviant phenotypes described in cattle diagnosed with BSE.20,21 In contrast with the previously described uniformity of BSE in cattle, this phenotype was initially revealed by the higher apparent molecular mass of the unglycosylated PrP^res, compared to that of typical BSE. This molecular mass difference is determined by an extended N-terminus of PrP^res that only occurs in the deviant phenotype, as evidenced by the presence of epitope 101-WGQGGTH-107 of bovine PrP.4,20,22,24,25 This higher apparent molecular mass of PrP^res has now been confirmed using 11 other antibodies (groups B and C) which recognize different epitopes along the central or C-terminal part of the PrP protein, and with three other antibodies (group A) against its more N-terminal end, all of which showed a prominent binding to the PrP^res from H-type isolates.

Some of the antibodies used during the study (group C), revealed an unexpected feature of H-type isolates leading to the identification of a second, previously unrecognised form of PrP^res (14–24 kDa)(PrP^res #2), in addition to the previously described PrP^res (19–30 kDa)(PrP^res #1). This appears as a major new molecular feature that distinguishes H-type form typical BSE. Although both the size of PrP^res #2 and its specific labelling only with the most C-terminus-specific antibodies suggest that it differs from PrP^res #1 by a more C-terminal cleavage, the data also imply the presence of glycosyl-groups. The coexistence of different PrP^res cleavage products, including different glycosylated fragments, was also recently reported in the recently described form of atypical scrapie or Nor98, which is characterized by very unusual PrP^res Western blot features.26 Although N-terminal sequencing of PrP^res was not considered in this study, epitope mapping suggests a cleavage of the PrP region around amino-acid 154–163. This situation is similar to that recently described in several forms of Creutzfeldt-Jakob disease, including sporadic,27 iatrogenic28 and genetic29 CJD in which the presence of C-terminally cleaved fragments (11–13 kDa) were revealed by using C-terminus-specific antibodies. A similar finding was also reported in cases of human Gerstmann-Sträussler-Sc heinker (GSS) syndrome but only in those patients in which the ≈21–29 kDa PrP^res was also detected.30

Beside the two isolates here described that were transmitted to C57Bl/6 mice, similar molecular features were also obtained in five other H-type isolates identified in France at abattoir or at slaughterhouse, but also in similar cases recently identified from active surveillance in other countries (Netherlands, Germany, Poland, Sweden, USA, Switzerland)(1 case in each of these countries).31–33 Interestingly the case identified in Switzerland had been identified in a miniature zebu (Bos indicus) kept in a zoological park, which allowed detailed investigations of the case in very good sampling conditions.33 Detailed study of this case also showed, in agreement with Western blot results, that PrPd in the central nervous system was labelled by immunohistochemistry using P4 antibody, which was not the case in a typical BSE case under the conditions used. All together our data also clearly show that the presence of lower molecular weights PrP^res forms is not linked to the quality of the cattle brain samples, which in some cases may be autolysed, since similar features were found in all the samples examined in the natural host (13 cases), including some samples of very good quality.

Furthermore, we then used selected antibodies to analyse the PrP^res features following transmission of the disease in C57BL/6 mice, and showed that the salient features previously recognized in cattle were faithfully maintained in mice. We have already reported the higher apparent molecular mass of PrP^res #1 in C57Bl/6 mice, associated with presence of the WGQGG epitope.22 We now report the additional recognition in these mice of the PrP^res #2 form, as identified in cattle, by group C C-terminal antibodies. As in H-type infected cattle, this resulted in a stronger detection of total PrP^res by SAF84 antibody in H-type BSE than in typical BSE. Western blot analyses in mice also showed the presence of an additional ≈7 kDa band recognised by antibodies 12B2 and Sha31 (groups A and B respectively), i.e., corresponding to the most N-terminal region in comparison to those recognized by group C antibodies. A similar fragment could also be detected in the H-type cattle brain stems, but clearly at lower levels (compare left and right lanes in inset of Figure 1A. It is presently unknown whether this is associated with the bovine-to-mouse transmission. It should also be emphasized that the studies of bovine brains were limited to the brain stem, whereas the PrP^res analyses in mice were performed on the whole brain. The precise mechanisms involved in PrP^res formation in such cases remain to be elucidated, and a proposed interpretation of the PrP^res cleavage is shown in Figure 4. Such a ≈7 kDa PrP^res fragment, non-reactive with group C C-terminus-specific antibodies, might potentially represent an N-terminal fragment of PrP^res #2 which would roughly correspond to the region 85–156 of bovine PrP. However, the identification of low molecular weight PrP^res fragments cleaved at both C- and N-terminal ends of the prion protein is reminiscent of a major feature of Gerstmann-Sträussler-Scheinker (GSS) syndrome in humans, in which this may even be the only detected PrP^res fragment in the absence of PrP^res 27–30.30,34,35 Sequencing of the open reading frame in the H-type isolates in cattle has already shown that such cases are not necessarily associated with changes in the PrP sequence20,31–33 as is the case in human GSS. Interestingly, it should be noted that recently the identification of a proteinase K resistant N- and C-terminally truncated form of PrP^res has also been reported in atypical scrapie.26,37 Although this PrP^res fragment appears to be a signature of this recently described form of prion disease in sheep,38 it is not associated with a specific genetic change.26,39

It has been stated that the abnormal PrP deposits detected by immunohistochemistry in C57Bl/6 mice infected with H-type isolates are only present as amyloid plaques, in contrast to BSE in such mice in which only diffuse deposits are found.22,40 PrP amyloids are also consistently detected in human GSS, in contrast to most other forms of Creutzfeldt-Jakob disease.41 Unfortunately, due to the lack of fixed material, histopathological studies could not be considered in these initial cattle samples. It is thus unknown whether these remarkable features in mice reflect a property of the pathological process also present in the original host, since amyloid formation was shown to be strongly influenced by the primary structure of PrP.42,43 In GSS, the presence of a C-terminally cleaved 8 kDa PrP^res fragment was associated with the presence of multicentric amyloid deposits, with which it colocalizes.29,44 The high propensity of synthetic PrP peptides corresponding to N- and C-terminally cleaved PrP, such as the human 85–148 sequence, to form amyloid fibrils was also demonstrated in vitro.45,46

Similarities of PrP^res in H-type BSE in cattle with that found in different human prion diseases, including both CJD in its different forms and GSS, raise crucial questions regarding the molecular mechanisms involved in PrP^res accumulation in such cases in cattle. Among human prion diseases, both human GSS and genetic CJD could be transmitted experimentally, in some cases, to animals such as rodents or primates expressing only a wild-type prion protein.47–50 These latter data clearly indicate that a host expressing a normal form of the prion protein can develop such forms of prion diseases, although molecular analyses of the specific features associated with these forms of human diseases have been poorly described in experimental models.

Our data obtained from one of the deviant phenotypes of prion diseases in cattle (H-type) showed some remarkable molecular similarities with human prion diseases. Although the origin of such cases is still unknown, these findings raise important questions17–19 regarding the respective origins and pathogenic mechanisms of natural prion diseases in humans and in animals prion diseases.

Abbreviations

BSE	=	bovine spongiform encephalopathy
CJD	=	Creutzfeldt-Jakob disease
CWD	=	chronic wasting disease
GSS	=	Gerstmann-Sträussler-Scheinker
H-type BSE	=	high type bovine spongiform encephalopathy
PK	=	proteinase K
PNGase	=	peptide N-glycosidase F
PrPd	=	disease-associated prion protein
PrPres	=	protease-resistant prion protein
PrPSc	=	scrapie prion protein
TSE	=	transmissible spongiform encephalopathy

Figures and Tables

Figure 1 Molecular discrimination of H-type and typical BSE in cattle and in C57Bl/6 infected mice. Western blot detection of PrP^res in cattle (A–C) or C57Bl/6 infected mice (E–G), from H-type isolate (right lane in each panel) or typical BSE (left lane in each panel), using 12B2 (A and E), Sha31 (B and F) or SAF84 (C and G) monoclonal antibodies. In each Western blot, typical BSE is shown in the left lane and H-type BSE in the right lane. Brain equivalent quantities loaded per lane were 100 µg (left lanes) and 250 µg (right lanes) for cattle samples and 1.2 mg (left lanes) and 4.8 mg (right lanes) for mouse samples. Inset of (A) shows the N-terminal PrP^res fragment (≈7 kDa) in cattle (left lane) and mouse (right lane). (D) Interpretation of the Western blot profile observed with SAF84 monoclonal antibody. (H) Unspecific immunoreactivity of anti-mouse Ig conjugate on samples prepared from normal brain of C57Bl/6 mouse.

Figure 2 Molecular analysis of PrP^res after deglycosylation in C57Bl/6 infected mice. Western blot analysis of PrP^res from cattle (A–C) or C57Bl/6 mice (D–F) infected with H-type (lanes 3 and 4) or typical BSE (lanes 1 and 2), after deglycosylation by PNGase treatment. Monoclonal antibodies used for PrP^res detection were 12B2 (A and D), Sha31 (B and E) or SAF84 (C and F). Brain equivalent quantities loaded per lane were (i) from cattle, for typical BSE, 50 (lane 1) or 2,6 (lane 2) mg, and for H-type BSE, 50 (lane 3) or 6 (lane 4) mg, and (ii) from mice, for typical BSE, 0.5 (D and F) or 0.3 (E) mg, and for H-type BSE, 2.5 (D and F) or 1.3 mg (E).

Figure 3 Quantitative molecular features of PrP^res in C57Bl/6 infected mice. Molecular masses (A) and glycoform ratios (B) (means ± standard deviations) of PrP^res detected with Sha31 antibody, from C57Bl/6 mice infected with H-type or typical BSE isolates. PrP^res of mice infected with an experimental scrapie (C506M3) strain4 was used as a control.

Figure 4 Schematic representation of proteinase K-resistant PrP fragments identified from H-type or typical BSE. Putative regions for proteinase K cleavage are indicated, as assessed by immunoreactivities with anti-PrP antibodies, and categories of epitope locations (groups A, B and C).

Table 1 Antibodies used to characterize H-type and typical BSE

Antibodies	Sequences Recognized in the Bovine PrP Protein	Concentration	Ref.
Group A
SAF32	62 - QPHGGGW - 92	10.000 (ascitic fluid)	51
4F2	62 - QPHGGGW - 92	100 ng/ml	57
12B2	101 - WGQGG - 105	340 ng/mL	58
P4	101 - WGQGGSH - 107	100 ng/ml	52
Group B
9A2	110 - WNK -112	200 ng/ml	58
RB1	110 - WNKP - 113	1/2500 (rabbit antiserum)	53
F89/160.1.5	148 - PLIHFGSD - 155	200 ng/ml	54
Bar233	152 - FGSDYEDRYYRE - 163	1/10.000 (ascitic fluid)	51
L42	156 - YEDRYY - 161	200 ng/ml	58
Sha31	156 - YEDRYYRE - 163	Kit Biorad TeSeE sheep/goat 1/10	51
6H4	156 - YEDRYYREN - 164	100 ng/ml	55
Group C
12F10	154 - SDYEDRYYRE - 163	100 ng/ml	51
SAF84	175 - RPVDQY - 180	75 (mouse) or 150 (bovine) ng/ml	51
94B4	198 - HTVTTTTK - 205	70 ng/ml	24
F99/97.6.1	229 - YQRE - 232	540 ng/ml	54
R524	233 - SQAY - 236	1/500 (rabbit antiserum)	56

Acknowledgments

We greatly acknowledge the technical assistance of Dominique Canal, Jérémy Verchère and Johann Vulin for the Western blot experiments and of Emilie Antier and Clément Lavigne for the follow-up of animal experiments, as well as Diana Warwick for careful reading of the manuscript. We also acknowledge collaborative work on H-type cases with A. Buschmann and M. Groschup (FLI, Germany), M. Polak (NVRI, Poland), D. Gavier-Widen (SVA, Sweden), J. Richt (USDA, USA) and T. Seuberlich (NC, Switzerland). This work was supported by grants from the Neuroprion European network of Excellence (FOOD-CT-2004-506579 EUROSTRAINS project) and from GIS ”Infections à prions.“