Biochemical Characterization of Prion Strains in Bank Voles

Prions exist as different strains exhibiting distinct disease phenotypes. Currently, the identification of prion strains is still based on biological strain typing in rodents. However, it has been shown that prion strains may be associated with distinct PrPSc biochemical types. Taking advantage of the availability of several prion strains adapted to a novel rodent model, the bank vole, we investigated if any prion strain was actually associated with distinctive PrPSc biochemical characteristics and if it was possible to univocally identify strains through PrPSc biochemical phenotypes. We selected six different vole-adapted strains (three human-derived and three animal-derived) and analyzed PrPSc from individual voles by epitope mapping of protease resistant core of PrPSc (PrPres) and by conformational stability and solubility assay. Overall, we discriminated five out of six prion strains, while two different scrapie strains showed identical PrPSc types. Our results suggest that the biochemical strain typing approach here proposed was highly discriminative, although by itself it did not allow us to identify all prion strains analyzed.


Introduction
Transmissible spongiform encephalopathies (TSEs), or prion diseases, are neurodegenerative disorders that afflict humans and others mammals. Prion diseases may be caused by prion exposure (acquired forms), mutations in PRNP gene (genetic or hereditary forms) and sporadic events in which the source of infection has not yet been demonstrated (idiopathic or sporadic forms). They include sporadic and genetic Creutzfeldt-Jakob disease (sCJD and gCJD) in humans, scrapie in sheep and goats and bovine spongiform encephalopathy (BSE) in cattle.
All TSEs are characterized by the accumulation of PrP Sc , the pathological form of host encoded prion protein (PrP C ), considered to be the main or sole component of infectious agent termed prion, according to the protein-only hypothesis [1].
Prions exist as different strains that, when propagated in the same host, exhibit distinct disease phenotypes that persist upon serial transmissions [2,3]. Within the context of the protein-only hypothesis, it has been suggested that prion strain diversity is encrypted in distinct conformations of PrP Sc aggregates [4].
Although the identification of prion strains is still based on biological strain typing in rodents, several studies showed that prion strains can be distinguished based on different biochemical properties of PrP Sc : the electrophoretic features of the protease resistant core of PrP Sc [5][6][7], the relative proteinase K resistance of PrP Sc [8,9], or the physico-chemical behavior of PrP Sc during denaturation [10][11][12]. Indeed, such biochemical discrimination allowed large scale testing of small ruminants TSEs in EU in order to recognize the possible presence of BSE [13][14][15][16][17]. Notwithstanding, no unequivocal relationship between type of PrP Sc and strain has been demonstrated so far, possibly due to technical difficulties which do not allow the direct structural analysis of PrP Sc aggregates, or due to the presence of still unidentified strain-specific co-factors [18].
Further drawbacks may derive by PrP res -based approaches, as these are focused on the PK-resistant core of PrP Sc (PrP res ), while it is becoming increasingly clear that protease-sensitive isoforms of PrP are involved in different animal and human prion diseases [12,[19][20][21][22][23]. These findings led us to develop a new method (CSSA) which does not rely on the protease resistant PrP Sc but allows us to study the conformational properties of both protease sensitive and protease-resistant PrP Sc species [11].
We have shown that bank voles (Myodes glareolus) are susceptible to a wide range of prion sources [24][25][26][27][28][29]. Taking advantage of vole-adapted prion strains, this study aimed at investigating if any different strain is actually associated with distinctive PrP Sc types and if it might be possible to unequivocally identify strain through these biochemical phenotypes. We selected vole-adapted human and animal strains including gCJD and sCJD subtypes, sheep classical scrapie and bovine BSE. PrP Sc from voles was analyzed by semi-quantitative western blot after PK digestion (PrP res ) and after denaturation with increasing GdnHCl concentrations and PrP C /PrP Sc differential centrifugation (conformational stability and solubility assay).

PrP res -Epitope Mapping
PrP res -epitope mapping was performed on protease treated PrP Sc analyzed by western blot with a large panel of mAbs spanning the bank vole PrP, before and after deglycosylation (Figure 1a,b). This sensitive technique allowed the identification of several PrP res fragments which were identified based on the presence or absence of the epitopes examined ( Figure 1c). Overall, we identified five different patterns of PrP res , each characterized by the presence of one or more specific PrP res fragments, that were labeled from A to E (Table 1).  The type A was found only in voles infected with sCJD MM1/MV1 and gCJD E200K, and was characterized by three PrP res fragments of 19, 14 and 11 kDa (Figure 1a,b). The main 19 kDa fragment was detectable equally well by all mAbs, while the 14 kDa fragment was recognized only by SAF84 (Figure 1a,b). Thus 19 kDa and 14 kDa were both C-terminal fragments with a different PK cleavage site at their N-terminus and included the glycosylation sites with the corresponding mono-glycosylated and di-glycosylated bands, resulting in a complex electrophoretic pattern when the membrane was probed with SAF84. In contrast, the 11 kDa fragment was recognized by 6C2, 9A2, 12B2 and SAF32 but not by SAF84, thus indicating that it is an internal fragment cleaved at both the N and C-termini (Figure 1a,b).
The PrP res type B, was found in brain homogenates from bank voles infected with classical scrapie isolates. It was characterized by a single C-terminal fragment of 18 kDa (unglycosylated band): this fragment was recognized by all mAbs, but only partially by SAF32 (Figure 1) indicating that the cleavage site was more C-terminal respect to the type A fragment of 19 kDa.
The type C, characteristic of sCJD MV2 PrP res , showed a molecular weight intermediate between type B and D, which was well detectable with SAF84 and 6C2. With these two antibodies, sometimes the PrP res appeared as a 17-18 kDa doublet (Figure 1a). Type C PrP res was still recognized by mAbs N-terminal to 6C2, although with lower affinity and without showing the PrP res doublet (Figure 1a).
The PrP res type D, found in BSE, was characterized by a single C-terminal fragment of 17 kDa which was well detectable with SAF84 and 6C2, but barely detected by 9A2, and not at all by 12B2 and SAF32 mAbs (Figure 1a,b).
The type E, found only in voles inoculated with sCJD MM2, showed a PrP res characterized by two PK-resistant fragments of 17 and 14 kDa. The main PrP res fragment was a 17 kDa similar to type D, being well recognized only by SAF84 and 6C2. However, compared to type D, it was clearly less glycosylated ( Figure 1a). The 14 kDa fragment was C-terminal and glycosylated too, being similar to the 14 kDa PrP res fragment described above in type A PrP res .
Overall, the analysis of PrP res revealed a close correspondence with each strain, allowing us to discriminate all human-derived strains among them and from animal strains (scrapie and BSE). However, voles infected with two different scrapie strains showed identical PrP res patterns.
Interestingly, the PrP res pattern of scrapie, BSE, and sCJD subtypes, seems well preserved after transmission to voles. As previously reported [24], voles infected with sCJD subtypes faithfully reproduced the PrP res electrophoretic mobilities of human counterparts. Also in this study, with more detailed biochemical analyses, we confirmed that vole-adapted sCJD MM1/MV1 had a lower electrophoretic mobility than sCJD MM2 (19 kDa vs. 17 kDa) due to a different N-terminal PK cleavage site. Moreover, additional C-terminal fragments were found to be associated to sCJD subtypes [32,33] as here found in bank voles inoculated with sCJD MM1/MV1 and MM2. In addition, sCJD MV2 adapted to voles showed an intermediate molecular weight between sCJD MV1/MM1 and MM2 and displayed a unique doublet band similar to that described in human sCJD MV2 cases [34]. Furthermore, as observed in natural isolates, in voles the scrapie PrP res was cleaved by PK more N-terminally than PrP res from BSE adapted to vole (18 kDa vs. 17 kDa), as confirmed by differential detection of 12B2 and SAF32 mAbs.

PrP res Analysis
Brain homogenates (20% w/v) were prepared in 100 mM Tris-HCl with Complete protease inhibitor cocktail (Roche) at pH 7.4. The homogenates were either used directly or stored at −20 °C. After adding an equal volume of 100 mM Tris-HCl containing 4% sarkosyl, the homogenates were incubated for 30 min at 37 °C with gentle shaking. Proteinase K (Sigma-Aldrich) was added at a final concentration of 250 μg/mL and then the samples were incubated for 1 h at 55 °C with gentle shaking. The reaction was stopped with 3 mM PMSF (Sigma-Aldrich, St. Louis, MO, USA). Aliquots of samples were added with an equal volume of isopropanol/butanol (1:1 v/v) and centrifuged at 20,000 g for 5 min. The supernatant were discarded and the pellets were re-suspended in denaturing sample buffer (NuPage LDS Sample Buffer and NuPage Sample Reducing Agent, Invitrogen, Carlsbad, CA, USA) and were analyzed by Western Blotting.
Deglycosylation was performed by adding 18 μL of 0.2 M sodium phosphate buffer (pH 7.4) containing 0.8% Nonidet P40 (Roche) and 2 μL (80 U/mL) of N-Glycosidase F (Roche) to 5 μL of denaturated samples and by incubating overnight at 37 °C with gentle shaking. Samples were then analyzed by Western blotting.

Conformational Stability Analysis
As previously described [11], the conformational stability was analysed by CSSA. Briefly, aliquots of brain homogenates (6% w/v in 100 mM TrisHCl, pH 7.4) were added with an equal volume of 100 mM TrisHCl (pH 7.4), sarcosyl 4% and incubated for 1 h at 37 °C with gentle shaking. Then, aliquots of each sample were incubated for 1 h at 37 °C with different concentrations of GdnHCl (Pierce) to obtain final concentrations ranging from 0 to 4 M. After GdnHCl treatment samples were centrifuged at 20,000 g for 1 h at 22 °C and the pellets were re-suspended in denaturing sample buffer (NuPage LDS Sample Buffer and NuPage Sample Reducing Agent, Invitrogen) and analysed by WB. The dose-response curves allowed us to estimate the concentration of GdnHCl able to solubilize 50% of PrP Sc (GdnHCl 1/2 ). Individual denaturation curves were analyzed and best-fitted by plotting the fraction of PrP Sc remaining in the pellet as a function of GdnHCl concentration, and using a four parameter logistic equation (GraphPad Prism).

Conclusions
This study confirms that by using appropriate biochemical methodologies it is possible to obtain indirect information on the strain-specific conformation of PrP Sc , such to allow a strain typing approach with a quite high discriminatory power. Indeed, five out of six prion strains analyzed were characterized by specific PrP Sc types, clearly distinguishable by the combined analyses of PrP res fragments and PrP Sc conformational stability. As such, our data further support the close relationship between TSE strains and PrP Sc properties [4]. However, even within our experimental setting in which tested samples were represented by known, rodent-adapted prion strains, it was not possible to discriminate all prion strains analyzed solely based on the biochemical characterization of PrP Sc . Indeed, all vole adapted scrapie isolates showed indistinguishable PrP Sc types, although belonging to 2 different vole-adapted scrapie strains, endowed with different survival times ( Figure 2) and neuro-pathological phenotypes (data not shown). If this simply reflects our inability to detect conformational differences by an indirect biochemical approaches, or has a broader implication in term of the relationships between PrP Sc conformations and prion strains is currently under investigation.
Recently we have applied the biochemical PrP Sc typing methods here developed to the study of atypical human and small ruminant prions [39]. Interestingly, we confirmed a very high ability to discriminate PrP Sc types, even in prion diseases known to be enriched in PK-sensitive PrP Sc aggregates, such as VPSPr and Nor98 [23,39]. Indeed, it was possible not only to discriminate PrP Sc types based on the prion disease, but also to associate specific PrP Sc types to different PrP mutations in Gerstmann-Strä ussler-Scheinker (GSS).
Ongoing studies with other vole-adapted strains and alternative biochemical approaches will hopefully allow to further increase the discriminatory power of PrP Sc typing. Although the lack of discrimination of the two scrapie strains, we believe that our study proposes a relatively easy and straightforward biochemical approach useful to refine the biochemical strain typing in animal models and in natural prion isolates.