|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: LP RN. Performed the experiments: LP SM GV PF EE. Analyzed the data: LP RN. Contributed reagents/materials/analysis tools: MADB CD RG JL. Wrote the paper: LP JL UA RN.
Although proteinacious in nature, prions exist as strains with specific self-perpetuating biological properties. Prion strains are thought to be associated with different conformers of PrPSc, a disease-associated isoform of the host-encoded cellular protein (PrPC). Molecular strain typing approaches have been developed which rely on the characterization of protease-resistant PrPSc. However, PrPSc is composed not only of protease-resistant but also of protease-sensitive isoforms. The aim of this work was to develop a protocol for the molecular characterization of both, protease-resistant and protease-sensitive PrPSc aggregates. We first set up experimental conditions which allowed the most advantageous separation of PrPC and PrPSc by means of differential centrifugation. The conformational solubility and stability assay (CSSA) was then developed by measuring PrPSc solubility as a function of increased exposure to GdnHCl. Brain homogenates from voles infected with human and sheep prion isolates were analysed by CSSA and showed strain-specific conformational stabilities, with mean [GdnHCl]1/2 values ranging from 1.6 M for MM2 sCJD to 2.1 for scrapie and to 2.8 M for MM1/MV1 sCJD and E200K gCJD. Interestingly, the rank order of [GdnHCl]1/2 values observed in the human and sheep isolates used as inocula closely matched those found following transmission in voles, being MM1 sCJD the most resistant (3.3 M), followed by sheep scrapie (2.2 M) and by MM2 sCJD (1.6 M). In order to test the ability of CSSA to characterise protease-sensitive PrPSc, we analysed sheep isolates of Nor98 and compared them to classical scrapie isolates. In Nor98, insoluble PrPSc aggregates were mainly protease-sensitive and showed a conformational stability much lower than in classical scrapie. Our results show that CSSA is able to reveal strain-specified PrPSc conformational stabilities of protease-resistant and protease-sensitive PrPSc and that it is a valuable tool for strain typing in natural hosts, such as humans and sheep.
Transmissible spongiform encephalopathies (TSEs), or prion diseases, are neurodegenerative disorders that afflict humans and others mammals. TSEs may have genetic, infectious, or sporadic origins. Creutzfeldt-Jakob disease (CJD) is the most common TSE in humans and may be sporadic (sCJD), genetic (gCJD), or acquired (iatrogenic CJD). A novel human acquired prion disease, variant CJD (vCJD), appeared from 1995 onwards and was postulated to be caused by consumption of beef from cows infected with bovine spongiform encephalopathy (BSE). The most common forms of TSE in animals, scrapie in small ruminants, BSE in cattle and chronic wasting disease (CWD) in deer, are all acquired. New atypical forms of BSE in cattle, namely BSE-H  and BSE-L or BASE  and atypical scrapie in small ruminants, namely Nor98 , are supposed to be sporadic.
All TSEs are characterised by the accumulation of PrPSc, a misfolded form of the cellular protein PrPC. Besides differing in conformation, the two forms of the prion protein have divergent physical properties: while PrPC is soluble in nondenaturating detergents, rapidly digested by proteases, and rich in α-helical structure, PrPSc is insoluble in detergents, partially resistant to proteolysis and contains mostly β-sheet , , .
A major problem for the protein-only hypothesis, which postulates that prions are composed mainly or exclusively of PrPSc , was to explain the existence of multiple strains of prions. Prion strains are infectious isolates that, when propagated in the same host, exhibit distinct prion disease phenotypes that persist upon serial transmission , , . Thus, prion strains cannot be encoded by differences in PrP primary structure but anyway carry information that are independent from the host . The prion hypothesis equates strains to different self-propagating conformational variants of PrPSc .
Several studies demonstrated that prion strains can be distinguished based on different biochemical properties of PrPSc, thus allowing a molecular strain typing approach to TSEs. These studies are based on the electrophoretic features of the protease resistant core of PrPSc , , , the relative proteinase K resistance of PrPSc , , , or the physico-chemical behaviour of PrPSc during denaturation ,  and strongly support the view that distinct strains show different PrPSc conformations.
A conformational stability assay (CSA), combining guanidine hydrochloride (GdnHCl) denaturation with limited proteolysis using proteinase K, showed that different prion strains may exhibit distinct denaturation profiles . A conformation-dependent immunoassay (CDI) showed the existence of multiple strain-specified PrPSc conformers by quantifying the immunoreactivity of native and denatured PrPSc of eight hamster prion isolates . This technique also showed that prion-infected brains contain both protease-sensitive (sPrPSc) and protease-resistant PrPSc (rPrPSc). More recently, it was shown by CDI that sPrPSc represents as much as 90% of total PrPSc in the brain of patients with sporadic CJD .
These findings along with the recent discovery in humans  and animals  of previously undetected TSEs characterised by relatively protease sensitive PrPSc, highlights the need of molecular strain typing methods able to recognize PrPSc populations based on their physical properties rather than only based on protease digestion.
In this study we aimed at developing a new conformational stability assay based on the differential solubility of PrPC and PrPSc , , that we called CSSA (conformational stability and solubility assay). We have previously transmitted CJD and scrapie isolates to bank voles , , which showed to be a valuable tool for the biological characterization of the most common forms of sCJD, gCJD and scrapie. We took advantage of these vole-adapted strains in order to evaluate the potential of CSSA for strain discrimination.
We first set up experimental conditions allowing the most advantageous separation of PrPC and PrPSc, and thus performed the conformational stability assay by measuring PrPSc solubility in homogenates treated with increasing concentrations of GdnHCl in the absence of proteinase K. Indeed, insoluble PrP was inversely correlated to GdnHCl concentration, and dose-response curves allowed estimation of the concentration of GdnHCl able to solubilise 50% of PrPSc.
Additionally, we extended the study to natural isolates of sheep scrapie and human sCJD cases to investigate the potential of CSSA for strain discrimination in natural hosts. Our results show that this method is valuable for the biochemical typing of strains in voles and it is also a promising tool for molecular analysis of natural prion isolates.
It has been previously reported that normal PrP is composed of full-length PrP (FL-PrP) as well as of 2 C-terminal fragments derived from physiological cleavage at the α and β sites: C1, which is the most represented, derives from α PrP cleavage at position 111/112, while C2 is usually barely detectable and is cleaved around the octarepeat region . Interestingly, α cleavage disrupts the conserved neurotoxic and amyloidogenic region comprising residues 106–126 of PrP, preventing the generation of PrPSc, while β cleavage occurs upstream this conserved region. Accordingly, C2 was reported to be enriched in diseased brains and insoluble in nondenaturing detergents, similarly to PrPres .
In order to determine if C1 and C2 were detectable in voles, we analysed healthy and diseased brain homogenates either before or after deglycosylation, with SAF84, whose epitope is present in both fragments, and 12B2 which recognises an epitope that is present only in C2 (Fig. 1A). In normal brain homogenate (NBH), FL-PrP was accompanied by substantial amounts of C1 and lower amounts of C2. In contrast, in voles infected with the Italian scrapie isolate SS7 (SBH)  C2 was strongly enhanced while C1 was not easily detectable (Fig. 1B and C). Furthermore, the PrPres fragment generated after PK digestion in SBH was similar to C2 (Fig. 1B and C). Finally, in SBH PrP dimers were present which were completely absent in NBH.
In order to develop a conformational stability assay based on the differential solubility of PrPC and PrPSc we first set up the experimental conditions which enabled the most advantageous separation of PrPC from PrPSc. This was obtained through a conventional procedure based on centrifugation in the presence of detergents.
By varying concentrations of different detergents, times of centrifugation and centrifugal force (see Material and Methods), we found that treatment with 2% sarcosyl followed by centrifugation at 20.000 g for 1 h enabled an optimal separation of PrPC from PrPSc.
In these conditions, >95% of the total PrPC in NBH was soluble and was found in the supernatant fraction (Fig. 2A, left panel) while in SBH most of PrP was sedimented (Fig. 2A, right panel). Unfractionated SBH contained high amounts of PrPres which exclusively sedimented to the pellet fraction (Fig.2A). Insoluble PrP from SBH was indeed mostly PK-resistant (~90% of insoluble PrP), while soluble PrP was PK-sensitive. All fractions from NBH were devoid of PrPres (Fig. 2A).
Furthermore, soluble and insoluble PrP fractions from SBH displayed slightly different banding patterns which suggested that C2 was mainly insoluble, while soluble PrP contained FL-PrP and the C1 fragment (Fig. 2A). Moreover, PrPSc dimers were clearly segregated in the insoluble fraction (Fig. 2A). The differential solubility of C1 and C2 in SBH was confirmed by the analysis of deglycosylated PrP species (Fig. 2B), which showed that C1 was completely soluble while C2 was mainly sedimented in the pellet fraction (compare “S” and “P” lanes in Fig. 2B, left panel). With the aim to mimic a situation comparable to that expected in animals with pre-clinical disease, we also studied the differential solubility of C1 and C2 after mixing equal amounts of NBH and SBH (Fig. 2B). Indeed, under these conditions the PrPC content is increased compared to SBH, as can be seen by the higher proportion of C1 in mixed NBH/SBH compared to SBH alone (compare “Tot” lanes in the two panels of Fig. 2B). Also in this condition, C1 was completely soluble and the pellet fraction was enriched in C2 (compare “S” and “P” lanes in Fig. 2B, right panel).
Finally, we investigated the efficacy of our solubility assay for separating PrPC and PrPSc in voles infected with other, non scrapie-derived prion strains (Fig. 2C). For these experiments we used voles infected with MM1 and MM2 sCJD . In both strains a substantial amount of PrP was found in the insoluble fraction after detergent treatment and centrifugation. As already observed in SBH, the banding patterns of insoluble and soluble PrP were distinct, suggesting a specific precipitation of PrP species associated to disease, namely C2 and PrP dimers (compare “S” and “P” lanes in Fig. 2C). Furthermore, after PK digestion, PrPres was strongly enriched in the pellets and virtually absent in the supernatants.
Collectively, these findings strongly suggest that under the experimental conditions described above we were able to specifically precipitate PrPSc in brain homogenates from voles infected with different prion strains.
The near complete separation of PrPC from PrPSc allowed us to develop a procedure for biochemical strain typing based on the conformational stability of PrPSc after exposure to GdnHCl. The conformational solubility assay was set up by measuring PrPSc solubility in positive brain homogenates treated for 1 hour with increasing concentrations of GdnHCl.
As expected, in SBH PrPSc was solubilized by increasing GdnHCl concentrations (Fig. 3A and B). Indeed, with concentrations of GdnHCl equal or greater than 1.5 M, PrPSc was partially solubilised and was progressively found in the supernatant instead of the pellet (compare Fig. 3A and 3B). With 3.5 M GdnHCl, virtually all PrP from SBH was found in the soluble fraction (Fig. 3A and 3B). The solubilization of PrPSc was guanidine-dependent and dose-response curves enabled estimation of the concentration of GdnHCl able to solubilise 50% of PrPSc ([GdnHCl]1/2). The [GdnHCl]1/2 value was similar when estimated either in the pellet or in the supernatant fractions (Fig. 3C). In contrast, PrP from NBH remained soluble within the range of GdnHCl concentrations tested (Fig. 3D). Denaturation of PrPSc was complete after incubation with GdnHCl for 1 hour, as very similar denaturation curves were obtained when the treatment was extended up to 4 hours (Fig. S1). When the denaturation curves where measured in replica blots with mAbs recognising different PrP species, namely SAF32 and SAF84 (see scheme in Fig. 1A), we obtained similar [GdnHCl]1/2 values, which suggest that C2 and FL PrPSc share the same conformational stability (Fig. S2).
We then studied the relationship between insolubility and PK-resistance of PrPSc during denaturation. In order to investigate if denaturation equally affects insolubility and PK-resistance, we compared denaturation curves derived from the same SBH, either obtained by insoluble PrPSc (CSSA) or by PK-resistant PrPSc (CSA). As shown in Fig. 4, the curves of insoluble PrPSc (Fig. 4A) and PrPres (Fig. 4B) didn't show differences (Fig. 4C), suggesting that insolubility and PK-resistance were equally susceptible to denaturation by GdnHCl. This finding was further confirmed in experiments aimed at investigating whether, after denaturation, solubilized PrPSc could partially preserve its resistance to proteinase K. Indeed, after denaturation with 3 M GdnHCl, soluble PrP was fully susceptible to protease digestion (Fig. 4D).
Under the experimental conditions described above, we investigated the potential of CSSA for differentiating prion strains.
As reported in previous studies ,  sCJD, gCJD and some scrapie isolates present distinct and specific patterns of transmission in voles, based on survival times, lesion profiles, PrPSc deposition and PrPres biochemical properties. We have shown that voles infected with MM1/MV1 sCJD and E200K gCJD isolates were characterised by a PrPres fragment of ~19 kDa, MM2 sCJD showed a PrPres fragment of ~17 kDa, while natural scrapie isolates and murine scrapie ME7 were characterised by a PrPres fragment of ~18 kDa, intermediate between types 1 and 2 CJD.
Since these PrPSc types showed specific molecular characteristics after PK digestion, we investigated the conformational stability of PrPSc in voles infected with MM1 (n=4), MV1 (n=3) and MM2 (n=5) sCJD, E200K gCJD (n=3) and scrapie SS7 (n=4).
Interestingly, CSSA revealed distinct denaturation profiles (Fig. 5), with [GdnHCl]1/2 values ranging from 1.5 M to 3 M (Table 1). Voles infected with MM1 sCJD, MV1 sCJD and E200K gCJD showed the highest resistance to denaturation, with mean [GdnHCl]1/2 values of 2.77 M, 2.88 M and 2.88 M, respectively. In contrast, PrPSc from voles infected with MM2 sCJD was the most sensitive (mean [GdnHCl]1/2 value of 1.63 M), while scrapie infected voles showed intermediate [GdnHCl]1/2 values (2.10 M). We then combined the individual curves within each group in order to compare the denaturation profiles of the 5 groups (Fig. 5B). Scrapie SS7 ([GdnHCl]1/2 value of 2.10±0.02) and MM2 sCJD ([GdnHCl]1/2 value of 1.61±0.03) showed strain specific denaturation profiles, being significantly different from all other groups (SS7 vs MV1, P=0.0037; SS7 vs MM1, P<0.0001; SS7 vs E200K, P<0.0001; SS7 vs MM2, P<0.0001; MM2 vs MV1, P=0.0022; MM2 vs MM1, P<0.0001; MM2 vs E200K, P<0.0001). In contrast, MM1 sCJD, MV1 sCJD and E200K gCJD were not significantly different among them (shared [GdnHCl]1/2 value of 2.79±0.06, P=0.52) possibly representing a distinct strain, in agreement with previous findings which showed comparable phenotypes of disease in voles infected with these 3 isolates .
Since we were interested in exploiting our conformational solubility assay (CSSA) for strain discrimination in natural prion diseases, we analysed 3 of the isolates used as inocula for transmission to voles, namely the human MM1 and MM2 sCJD isolates and the sheep scrapie SS7.
At first we tested the efficiency of separation of PrPSc from PrPC in human and sheep brain homogenates, under the same experimental conditions used for vole brain homogenates. In all samples, PrPSc was enriched in the pellet, as showed by the segregation of PrPres in the insoluble fraction and by the distinct banding patterns shown by soluble and insoluble PrP (Fig. S3). In negative sheep brain homogenates the fraction of PrP segregating in the pellet was somewhat higher than in vole brains (>10% in some experiment). However, this sedimented PrPC was insensitive to GdnHCl and thus did not interfere with our assay (Fig. S4).
When tested by CSSA, MM1 sCJD, MM2 sCJD and SS7 scrapie revealed different susceptibilities to denaturation (Fig. 5A and B). Interestingly, the rank order of conformational stability and the [GdnHCl]1/2 values observed closely matched those found in the vole counterpart, being MM1 sCJD the most resistant ([GdnHCl]1/2=3.31 M), followed by scrapie SS7 ([GdnHCl]1/2=2.23 M) and MM2 sCJD ([GdnHCl]1/2=1.63 M).
In order to test the ability of CSSA for studying the conformational stability of protease-sensitive PrPSc, we took advantage of the recently described atypical scrapie strain, Nor98 , which has been shown to induce the accumulation of relatively protease-sensitive PrPSc . As previously observed for all other strains investigated, insoluble and soluble Nor98 PrPSc showed different segregation of PrP species (Fig. 6). Interestingly the 12 kDa PrPres fragment, which is a characteristic feature of Nor98 PrPres , , was observed before PK-digestion and segregated with insoluble PrP, in a manner similar to what was observed for C2 in all other strains. However, in Nor98 only a minimal fraction of insoluble PrP was protease resistant (Fig. 6), while PrPres represented >90% of insoluble PrP in classical scrapie (Fig. 6). These findings confirm that in Nor98 samples disease-associated PrP is mostly PK-sensitive, although accompanied by low amounts of genuine PK-resistant PrPSc.
We then studied the conformational stability of Nor98 (n=5) and classical (n=4) Italian field scrapie cases of different PrP genotypes (Table 2). Classical scrapie samples gave denaturation curves very similar to that of SS7, with [GdnHCl]1/2 values ranging from 1.96 to 2.31 (Fig. 7 and Table 2). Nor98 samples gave remarkably similar denaturation profiles, independently of PrP genotype, and showed high sensitivity to GdnHCl denaturation (Fig. 7), with [GdnHCl]1/2 values ranging from 1.26 to 1.43 (Table 2). Within group comparison didn't show significant differences among classical scrapie (shared [GdnHCl]1/2 value of 2.09±0.11, P=0.52) and Nor98 (shared [GdnHCl]1/2 value of 1.36±0.04, P=0.50). Comparison of the combined classical scrapie and Nor98 denaturation profiles (Fig. 7B) gave statistically significant different [GdnHCl]1/2 values (P<0.0001).
The main pathogenetic event in TSEs involves the trans-conformation of PrPC to PrPSc, which leads to the accumulation of detergent-insoluble and partially protease-resistant aggregates. Thus, resistance to PK digestion and insolubility are the hallmarks of PrPSc. However, previous studies on the molecular characterization of PrPSc were mainly focused on the PK-resistant core of PrPSc, due to the difficulties in differentiating PrPC from PrPSc in infected brain homogenates. Notwithstanding, it is becoming increasingly clear that insoluble but protease-sensitive isoforms of PrP are involved in different animal and human prion diseases , , , , , , . These PrP isoforms have been detected by several methods, including conformation-dependent immunoassay , , , immunological capture , , differential centrifugation , , thermolysin digestion  and cold PK digestion .
Among the above-mentioned techniques able to detect sPrPSc, the CDI was also reported to distinguish eight hamster prion strains based on PrPSc conformation, by plotting the ratio of antibody binding to denatured/native PrP as a function of the concentration of PrPSc . However, CDI depends on the availability of monoclonal antibodies able to recognise buried epitopes in PrPSc, such as the 3F4. We developed a protocol for the molecular characterization of PrPSc aggregates which does not rely on their protease resistance but is based on a conventional procedure of differential centrifugation in the presence of detergents and on the solubilization of PrPSc aggregates upon denaturation with increasing concentrations of GdnHCl. Our findings show that CSSA is reliable and straightforward, and that it is able to discriminate PrPSc conformers associated with different TSE strains. Compared to CDI, we believe that CSSA offer some interesting advantages: i) it does not depend on the antibody used and can be thus exploited to compare PrPSc conformational stability in different species, ii) the proof that CDI is able to discriminate prion strains in natural hosts is still lacking, while here we show the potential of CSSA to be used in human and sheep natural isolates.
In previous studies , ,  a conformational stability assay (CSA) able to discriminate PrPSc conformers was set up by measuring the extent of loss in protease resistance as a function of increased exposure to GdnHCl. This method proved to be very helpful for molecular strain typing in different species , ,  and was also recently exploited for investigating some basic mechanisms of prion replication . The protocol that we developed is conceptually similar to CSA, as it derives information on the conformational stability of PrPSc aggregates. Indeed, we showed that CSA and CSSA gave very similar denaturation profiles in vole-adapted scrapie brain homogenates (see Fig. 4). However, we believe that CSSA represents a step forward in prion molecular typing and offers several advantages compared with previously used protocols. With CSA, in fact, susceptibility to PK is exploited to distinguish denaturated from native PrPSc. However, different prion strains may have distinct susceptibilities to PK, while CSA uses the same PK concentration to derive the level of PrPSc denaturation in different strains. Furthermore, when performing CSA it is necessary to dilute the GdnHCl to allow the activity of proteolytic enzyme. However, it is known that PrPSc unfolding can be a partially reversible phenomenon and, as reported, the dilution of the denaturant could restore the original protease-resistance of PrPSc . With CSSA this problem was circumvented by avoiding any change in denaturant concentration during the assay. Indeed, the denaturation step is followed by the centrifugal separation of soluble and insoluble fractions, which is performed under conditions identical to the denaturation step.
Most importantly, CSSA allowed the characterization of protease-sensitive PrPSc, and thus the direct comparison of the conformational stabilities of TSE strains associated with protease-sensitive and protease-resistant PrPSc. We investigated the potential of CSSA for strain typing of ovine field isolates of Nor98, which are characterised by high amounts of insoluble sPrPSc [29, present paper]. Our findings seem very encouraging, since Nor98 samples displayed a distinct denaturation profile, although the isolates derived from sheep of various PrP genotypes and presumably at different stage of the disease. Furthermore, Nor98 samples were easily discriminated from classical scrapie, which represents a different strain. Thus CSSA enabled characterization of the conformational stability of a protease sensitive strain, for which methods based on proteolysis would have characterised only a minimal amount of PrPSc present in the brain homogenate. This is promising in view of recent studies ,  showing that in sCJD and vCJD isolates, as much as 90% of PrPSc in the brains was estimated to be sPrPSc. Furthermore the recent discovery of protease sensitive prionopathy (PSPr), a previously unrecognised human prion disease , might suggest that prion diseases characterised by protease-sensitive PrP isoforms are more frequent than previously thought. Brain homogenates from PSPr cases were reported to contain mainly a protease sensitive form of insoluble PrP, accompanied by low amounts of typical protease-resistant PrP , similarly to our findings in Nor98 samples. It would be interesting to investigate the potential of CSSA for characterising these human protease-sensitive disease-related isoforms of PrP.
Another interesting feature of CSSA is that, by avoiding PK treatment, it allowed the characterization of FL-PrPSc, including the N-terminus which is cleaved upon PK digestion. Besides FL-PrP, in vivo trimmed PrPSc fragments were selectively precipitated and thus CSSA also conveyed information on their conformational stability. These fragments comprised C2 C-terminal fragments in scrapie and sCJD isolates, as well as the 12 kDa internal fragment characteristic of Nor98, and may be easily distinguished from FL-PrP, also by means of differential antibody binding (data not shown; see for example Fig. S2). This is interesting because it enables a comparison of the conformational properties of PrPSc aggregates made-up of FL-PrP or in vivo trimmed PrPSc fragments. In situ epitope-mapping of PrPSc, indeed, showed that FL and C2 PrPSc aggregates have different cellular localizations . Furthermore, by analysing thermolysin-resistant PrPSc in sheep scrapie and BSE isolates, Owen and colleagues have recently shown the potential of C2 fragments for strain typing .
The main potential drawback of CSSA might be the incomplete separation or PrPC and PrPSc. Herein, we developed a protocol able to minimise this problem, notwithstanding 1–4% of PrPC in brain homogenates from healthy voles was found in the pellet under our working conditions. There are however several lines of evidence suggesting that this potential drawback did not interfere with CSSA results. We have shown that brain homogenates from clinically affected voles contain low levels of soluble PrP (bona fide PrPC) (see Fig. 2). Based on the results obtained with NBH, less than 5% of PrPC is expected to be found in the pellet; on the other hand, in diseased brains bona fide PrPC was 5–30% of total PrP. From these considerations, it can be argued that the PrPSc/PrPC ratio in the pellet should be higher than 2 orders of magnitude, which represents the working range of CSSA. Furthermore, we obtained clear-cut differential solubility of PrPC and PrPSc even after increasing the PrPC/PrPSc ratio in the homogenate, by mixing SBH with NBH (Fig. 2). Most importantly, we found that the sedimented PrPC fraction was insensitive to denaturation (Fig. S4) and thus it does not interfere with CSSA.
We investigated the potential of CSSA for strain discrimination by analysing vole-adapted strains deriving from human sCJD, gCJD and sheep scrapie which we have previously shown to give distinct strains in voles , . In previous studies we have shown that voles infected with MM1/MV1 sCJD and E200K gCJD isolates showed identical transmission patterns and were characterised by an unglycosylated PrPres fragment of ~19 kDa, while MM2 sCJD showed an unglycosylated PrPres fragment of ~17 kDa. In contrast, all Italian scrapie isolates studied so far , , as well as the murine scrapie strain ME7 , were characterised by the accumulation of an unglycosylated PrPres with a molecular weight (~18 kDa) intermediate between CJD types 1 and 2, upon transmission to voles. These different PrPres types derive from different PK-cleavage sites of PrPSc, which in turn are believed to reflect distinct conformations of PrPSc aggregates. With CSSA we showed that these three PrPres types are actually characterised by PrPSc aggregates with distinct susceptibilities to denaturation by GdnHCl. Indeed, PrPSc from voles infected with MM1 sCJD, MV1 sCJD and E200K gCJD, characterised by 19 kDa PrPres, showed the highest resistance to denaturation, while sCJD MM2, characterised by 17 kDa PrPres, was the most susceptible and scrapie SS7, that was characterised by 18 kDa PrPres, displayed intermediate susceptibility. Interestingly, the within group variability of [GdnHCl]1/2 values was very low (see table 1) and allowed statistical comparisons among the different groups, which strengthened the view that [GdnHCl]1/2 values reflect strain-specific rather than individual PrPSc properties.
It has been recently suggested that the conformational stability of PrPSc is directly proportional to the length of the incubation time in mice . On this point, it is worth noting that our results in voles, although based on only 5 isolates, seem to contradict this conclusion. Indeed, the lowest conformational stability was associated with prions with the longest survival times, i.e. MM2 sCJD with a survival time of ~330 days post-inoculation (dpi), while scrapie SS7 (survival time of ~90 dpi) and MM1/MV1 sCJD (survival time of ~130 dpi) showed higher conformational stabilities and shorter survival times compared to MM2 sCJD. Further studies with an extended panel of isolates are needed to investigate if the direct relationship between conformational stability and incubation time observed in mice holds true also for vole prions.
We also explored the potential of CSSA for strain typing of natural prion diseases. To this aim we analysed three of the human and sheep isolates which were used for bioassay in vole. Interestingly, we found that the two isolates from MM1 sCJD and MM2 sCJD patients could be easily discriminated by their conformational stabilities, with MM1 sCJD displaying lower susceptibility to denaturation compared to MM2 sCJD ([GdnHCl]1/2 values of 3,31 M for MM1 sCJD and 1,63 M for MM2 sCJD). These findings compare well with previous studies by CSA, which showed that PrPSc associated with MM1 sCJD was ~ 2-fold more stable than that of MM2 sCJD, with [GdnHCl]1/2 values of 2,76 M and 1,63 M for MM1 sCJD and MM2 sCJD . Of course, this needs confirmation in a larger set of isolates.
Furthermore, this approach allowed us to compare the conformational stability of prions in their natural host and after transmission to voles. It has been previously shown that prion strains can either maintain their biological properties or mutate upon propagation in a new host species . More recently it was reported that a change in conformation was accompanied by the emergence of a new prion strain during interspecies transmission, while the conformational stability of PrPSc was preserved when a strain “bred true” in the new host . Our findings show that the rank order of conformational stability of sCJD and scrapie isolates was generally preserved upon adaptation in voles (see Fig. 5). In particular, the conformational stability of MM2 sCJD was identical before and after transmission in voles. We have also previously reported that MM2 sCJD, and to a lower extent MM1 sCJD, encountered a very low transmission barrier to voles when compared to several other prion strains . These observations may suggest that MM2 and possibly MM1 sCJD faithfully propagated their strain properties upon transmission in voles. The “conformational selection” model, postulates that host PrPC primary structure influences which of the portfolio of possible PrPSc types are thermodynamically preferred during propagation . In this model, the transmission barrier is determined by the degree of overlap between the subset of PrPSc types allowed or preferred by PrPC in the host and donor species. It can be thus speculated that the vole PrP sequence is prone to adopt some human PrPSc conformations and we have previously shown that this property may reside in the presence of some peculiar amino acid at relevant positions of vole PrP , .
Finally, we exploited CSSA for discriminating sheep field prion isolates, showing that this method is valuable for strain discrimination in natural hosts. Indeed, the conformational stability of prions was strongly associated with the strain, either classical scrapie or Nor98, and did not depend on individual properties such as age, clinical stage or PrP genotype (see table 2). Although preliminary, these findings suggest that CSSA might be exploited as complementary approach for increasing the discriminatory power of strain typing in small ruminant TSEs. Studies with natural and experimental TSEs, including sheep BSE and CH1641-like isolates, are underway.
Throughout the manuscript we used “prion” to refer to the infectious agent of TSEs, “PrPSc” to refer to the abnormal disease-associated PrP isoform, “rPrPSc” to refer to the protease-resistant fraction of PrPSc, “sPrPSc“ to refer to the protease-sensitive fraction of PrPSc and “PrPres” to refer to protease-resistant PrPSc fragments deriving from rPrPSc trimming by proteinase K.
Human brain tissues were from cerebral cortices of two sCJD cases (MM1 and MM2 types) previously characterised by vole bioassay .
The medulla oblongata of SS7 scrapie was obtained from an ARQ/ARQ Sarda sheep with clinical scrapie reported in Italy in 1997 and previously characterised by vole bioassay . All others sheep brain tissues were from Italian field cases detected by active and passive surveillance and characterised by discriminatory western blotting for molecular strain typing and by complete sequencing of the PRNP open reading frame, according to published protocols , . The details of these TSE cases are reported in Table 2.
Brain tissues were obtained from voles infected with vole-adapted human and sheep TSEs as previously reported , . The passage number and mean survival times ± standard deviation of the vole groups from which the samples used in the present study derived were as follows: sCJD MM1, second passage, 129±8 dpi; sCJD MV1, second passage, 128±15 dpi; sCJD MM2, second passage, 339±27 dpi; sCJD MM2, third passage, 323±41 dpi; gCJD E200K, second passage, 143±12 dpi; scrapie SS7, third passage, 93±5 dpi.
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.
The experimental conditions for PrPC/PrPSc separation were set up in vole brain homogenates by studying the effect of different detergents, centrifugal force and time of centrifugation.
Brain homogenates (3% to 12% w/v) were added with equal volumes of different buffers (100 mM Tris-HCl at pH 7.4 containing Sarcosyl 4% or 2%; 100 mM Tris-HCl at pH 7.4 containing NaDoc 1%, NP40 1%; 100 mM Tris-HCl at pH 7.4 containing Triton X-100 2%) and incubated for 1 hour at 37°C with gentle shaking. Then samples were centrifuged at 10000 to 20000 g for 1 h or 2 h. The obtained pellets were re-suspended with 100 mM Tris-HCl (pH 7.4) containing the relevant detergent. The experimental conditions then used throughout the paper in vole, sheep or human brain tissues included solubilisation in 100 mM Tris-HCl at pH 7.4 containing sarcosyl 2% and centrifugation at 20000 g for 1 h. For each of the different experimental conditions tested, equivalent aliquots of brain homogenate before centrifugation, along with supernatant and pellet fractions, were analysed by western blot either with or without PK digestion.
Aliquots of brain homogenates (3% to 6% w/v) were added with an equal volume of 100 mM Tris-HCl (pH 7.4) containing sarcosyl 4% and incubated for 1 h at 37°C with gentle shaking. Aliquots of 100 µl were treated with 100 µl of guanidine hydrochloride (GdnHCl) solutions with a final concentration ranging from 0 to 4.0 M. GdnHCl stock solutions were prepared from an 8 M solution (Pierce) diluted in water. After treatment with GdnHCl for 1 h at 37°C with gentle shaking, samples were centrifuged at 20000 g for 1 h at 22°C. Pellets were re-suspended in 90 µl NuPage LDS Sample Buffer (Invitrogen) and 10 µl NuPage Sample Reducing Agent (Invitrogen). Aliquots of supernatants were precipitated with 4-fold volume excess of pre-chilled methanol 30 min at −20°C, centrifuged at 15000 g for 30 min at 4°C and then were re-suspended in 90 µl NuPage LDS Sample Buffer (Invitrogen) and 10 µl NuPage Sample Reducing Agent (Invitrogen). Supernatant and pellet fractions were analysed by Western blotting.
Individual denaturation curves were analyzed and best-fitted by plotting the fraction of PrPSc remaining in the pellet as a function of GdnHCl concentration, and using a four parameter logistic equation (GraphPad Prism). In order to fit denaturation curves for each prion strain, the mean fraction of PrPSc remaining in the pellet ± SD were plotted. Statistical comparison of [GdnHCl]1/2 values were made by comparing the best-fit value for each data set with GraphPad Prism. This was performed by either fitting each data set independently or doing a global fit with a shared [GdnHCl]1/2 value, and then the results were compared with an F test. The simpler model was selected unless the extra sum-of-squares F test had a P value<0.05.
Aliquots of the same brain homogenate were treated in parallel according to CSSA and conformational stability assay (CSA) protocols. The CSA was performed as described , with minor modifications. Aliquots of brain homogenates (6% w/v) were added with an equal volume of 100 mM Tris-HCl (pH 7.4) containing sarcosyl 4% and incubated for 1 h at 37°C with gentle shaking. Aliquots of 50 µl were added with 50 µl of GdnHCl to give a final concentration ranging from 0 to 4.0 M. After 1 h of incubation at 37°C all samples were diluted to a final concentration of 0.4 M GdnHCl. Proteinase K (50 µg/ml) was added and the samples were incubated for 1 h at 37°C with gentle shaking. The reaction was stopped with 3 mM PMSF (Sigma). Aliquots of samples were added with an equal volume of isopropanol/butanol (11 v/v) and centrifuged at 20000 g for 5 min. Pellets were re-suspended in NuPage LDS Sample Buffer (Invitrogen) and were analysed by Western Blotting.
Electrophoresis and Western blotting were performed as previously described . Samples were denatured by adding NuPage LDS Sample Buffer (Invitrogen, Carlsbad, California, United States) and NuPage Sample Reducing Agent (Invitrogen), and heating for 10 min at 90°C. After centrifugation at 10000 g for 5 min each sample was loaded onto 12% bis-Tris polyacrylamide gels (Invitrogen). After electrophoresis and Western blotting on PVDF membranes (Immobilon-P; Millipore, Bedford, MA, USA), the blots were processed by SNAP i.d.™ Protein Detection System (Millipore) as described by the manufacturer instructions.
The monoclonal antibodies used, their epitopes on sheep PrP and the working dilutions were as follow: SAF84, PrP residues 167–173, 1.2 µg/ml; L42, PrP residues 148–153, 0.28 µg/ml; 12B2, PrP residues 93–97, 2.4 µg/ml and SAF32, PrP octarepeat, 2.4 µg/ml. Horseradish peroxidase-conjugated anti-mouse immunoglobulin (Pierce Biotechnology, Rockford, Illinois, United States) was used as secondary antibody (113000).
The membranes were developed with an enhanced chemiluminescence method (SuperSignal Femto, Pierce). Chemiluminescence signal was detected with the VersaDoc imaging system (Bio-Rad) and was quantified by QuantityOne software (Bio-Rad).
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) di N-Glycosidase F (Roche) to 5 µl of denaturated samples and by incubating overnight at 37°C with gentle shaking. Samples were then analysed by Western blotting as described above.
Effect of denaturation time on CSSA. Dose-response curves of insoluble PrP from brain homogenates of voles infected with scrapie SS7 (top panel) and MM1 sCJD (bottom panel) after treatment with increasing concentrations of GdnHCl for 1, 2 or 4 hours. Denaturation curves were best-fitted by plotting the fraction of PrP remaining in the pellet as a function of GdnHCl concentration. SS7 and MM1 didn't reveal differences based on time of treatment and showed the same [GdnHCl]1/2 values at 1, 2 and 4 hours (2.1 M for SS7 e 3 M for MM1 sCJD).
(0.13 MB TIF)
CSSA with different mAbs. A: Representative western blots of insoluble PrP from vole infected with E200K gCJD after denaturation with increasing concentrations of GdnHCl. Replica blots were probed with SAF84 (top) and SAF32 (bottom), as indicated on the left of the blot. The concentrations of GdnHCl were: 0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 and 4.0 M. In each lane 0.2 mg TE were loaded. B: Dose-response curves derived from the blots in panel A, obtained by plotting the fraction of PrP remaining in the pellet as a function of GdnHCl concentration and best-fitted and using a four-parameter logistic equation.
(0.57 MB TIF)
Separation of soluble and insoluble PrP fractions from human and sheep isolates. Western blot analysis of soluble and insoluble PrP from brain homogenates of sheep with classical scrapie and human with MM1 sCJD. Samples were centrifuged at 20000 g for 1 h in presence of 2% Sarcosyl, and supernatants (S) and pellets (P) were analysed with (+) or without (−) PK treatment. Aliquots of samples before centrifugation (Tot) were analysed too. TE loaded per lane were 0.2 mg for classical scrapie and 0.15 mg for MM1 sCJD. Scrapie membrane was probed with SAF84 and MM1 sCJD membrane was probed with L42.
(0.48 MB TIF)
Separation of soluble and insoluble PrP fractions from sheep normal brain homogenates. A: Western blot analysis of soluble and insoluble PrP fractions from a representative sheep negative brain homogenate. Total PrP (Tot) and PrP from supernatant (S) and pellet (P) fractions were analysed with (+) or without (−) PK treatment. TE loaded per lane were 0.2 mg. B: Western blot analysis of insoluble PrP from sheep negative and scrapie brain homogenates after denaturation with increasing concentrations of GdnHCl, either after normal (top panel) or long (bottom panel) exposure times. The concentrations of GdnHCl were: 0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 and 4.0 M. TE loaded per lane were 0.12 mg. A and B: Membranes were probed with SAF84.
(0.69 MB TIF)
The authors would like to thank Paolo Frassanito and Shimon Simson (Istituto Superiore di Sanità) for animal care, and Consiglia Parisi (Istituto Superiore di Sanità) for administrative management of projects and editorial help.
Competing Interests: The authors have declared that no competing interests exist.
Funding: This work was supported by grants from the European Union (Neuroprion Network of Excellence CT-2004-506579) and the Italian Ministry of Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.