Effects of different experimental conditions on the PrPSc core generated by protease digestion: implications for strain typing and molecular classification of CJD.

The discovery of molecular subtypes of the pathological prion protein PrPSc has provided the basis for a novel classification of human transmissible spongiform encephalopathies (TSEs) and a potentially powerful method for strain typing. However, there is still a significant disparity regarding the understanding and nomenclature of PrPSc types. In addition, it is still unknown whether a specific PrPSc type is associated with each TSE phenotypic variant. In sporadic Creutzfeldt-Jakob disease (sCJD), five disease phenotypes are known, but only two major types of PrPSc, types 1 and 2, have been consistently reproduced. We further analyzed PrPSc properties in sCJD and variant CJD using a high resolution gel electrophoresis system and varying experimental conditions. We found that pH varies among CJD brain homogenates in standard buffers, thereby influencing the characteristics of protease-treated PrPSc. We also show that PrPSc type 1 and type 2 are heterogeneous species which can be further distinguished into five molecular subtypes that fit the current histopathological classification of sCJD variants. Our results shed light on previous disparities in PrPSc typing, provide a refined classification of human PrPSc types, and support the notion that the pathological TSE phenotype is related to PrPSc structure.

Transmissible spongiform encephalopathies (TSEs), 1 or prion diseases, are a phenotypically heterogeneous group of neurodegenerative disorders that affects humans and animals. Human diseases include Creutzfeldt-Jakob disease (CJD), Gerstmann-Strä ussler-Scheinker disease, kuru, and fatal insomnia (1)(2)(3). CJD, by far the most common human TSE, may occur as a sporadic disease of unknown etiology (sCJD), a disease associated with mutations in the prion protein gene (PRNP), or a proven exogenous infection. The latter group includes variant CJD (vCJD) a distinct disease phenotype that is believed to have been transmitted from cattle to humans through the consumption of contaminated meat (3).
The cellular prion protein (PrP c ), a host-encoded, copper-and membrane-bound glycoprotein of unknown function, has a key role in TSE pathogenesis (4). There is no agent replication or transmission of infectivity in the absence of PrP c expression (5). Moreover, an abnormal, partially protease-resistant isoform (PrP Sc ) of PrP c specifically accumulates in the nervous system during infection and represents the hallmark of the disease (4,6). The conversion of PrP c to PrP Sc is a posttranslational event associated with an increase in ␤-sheet secondary structure in the protein (7,8).
CJD, like scrapie of sheep, comprises a broad spectrum of clinicopathological variants (9). In scrapie transmitted to mice, the most extensively studied TSE, the strain of the agent and the host genotype variability determined by polymorphisms in the coding region of PRNP are the major determinants of disease phenotype (10). Different prion strains transmit diseases to syngenic animals that differ in topography, type of lesion, and intracerebral distribution of PrP Sc . In humans, disease susceptibility and phenotypic expression are influenced by PRNP mutations and by the polymorphism at codon 129 that encodes either methionine (M) or valine (V) (11)(12)(13). Distinct agent strains have been demonstrated (14 -17) but remain to be characterized in full.
Although uncertainties remain on the molecular basis of TSE strains and the relationship between strains and PrP, several lines of evidence indicate that PrP Sc exists in a variety of molecular subtypes showing differences in conformation, glycosylation, protease resistance, and aggregation state (15, 18 -23), which may encode, at least in part, strain-specific properties. Furthermore, evidence from transmission studies indicates that the different PrP Sc types showing distinct physicochemical properties represent genuine biochemical signatures of individual strain-host genotype interactions. Because the analysis of biochemical properties of PrP Sc is much less time-consuming than bioassays in mice, unraveling the physicochemical properties of PrP Sc associated with each TSE strain or phenotype (i.e. PrP Sc "typing") has undoubtedly become of crucial importance for strain typing and molecular classification of TSEs, with wide implications for both disease diagnosis and epidemiologic surveillance.
Unfortunately, there is a significant disparity in the literature regarding the existence of distinct human PrP Sc types (23)(24)(25)(26)(27), and this to some extent is also true for animal TSEs (28 -30). Based on differences in gel mobility and N-terminal sequence of the core fragments generated by proteinase K (PK), Parchi et al. (23)(24)(25) originally identified two human PrP Sc types (named type 1 and type 2). Type 1 has a relative molecular mass of 21 kDa and the primary cleavage site at residue 82, and type 2 has a relative molecular mass of 19 kDa and the primary cleavage at residue 97. In other studies, however, the PrP Sc type 1 from codon 129 MM subjects was further distinguished into 2 subtypes showing a less than a 1-kDa difference in mobility (26,27). According to Wadsworth et al. (26) the two PrP Sc type 1 subtypes, they named types 1 and 2, show indistinguishable fragment sizes only when PK digestion is performed in the presence of 20 mM EDTA, thereby representing two distinct conformations acquired by PrP Sc in the presence of metal ions such as copper and zinc. In contrast, the two putative PrP Sc subtypes described by Zanusso et al. (27), also named types 1 and 2 although not comparable with those of Wadsworth et al. (26), would represent two distinct protein conformations because of a different response to pH variations (i.e. one conformation is pH-dependent; the other is not). The rationale for the study of the effect of metal ions and pH on PrP Sc properties lies in the finding that the octapeptide repeat PrP c sequence between residues 51 and 91 is a Cu 2ϩ binding motif that changes its conformation in the presence of copper and that this copper binding stoichiometry is pH-dependent (31,32).
In previous studies on PrP Sc typing in sCJD, we also reported a certain degree of heterogeneity within both type 1 and type 2 samples, particularly among type 1 samples from codon 129 MM subjects (23,24). However, these were subtle biochemical differences compared with the type 1/type 2 difference in relative molecular mass and failed to show a consistent reproducibility or a correlation with the disease pathological phenotype, suggesting that they might be related to the experimental conditions rather than to intrinsic differences in strains. For example, it is noteworthy that the various protocols that have been applied to define PrP Sc properties to date included the use of PK well outside its optimal conditions (i.e. pH optimum above 7.5; optimum temperature above 37°C) (33,34) and, above all, were not standardized with respect to variables that might affect its activity. On the other hand, the characterization of at least five distinct pathological sCJD subtypes (9) argues for the existence of more than two prion strains and two PrP Sc types. Thus, it is conceivable that the application of more sensitive techniques and more rigorous experimental conditions will lead to the distinction between potential artifacts related to sample preparation and disease-specific biochemical differences suitable for a refined classification of CJD PrP Sc types. To reach this goal, shed light on the current PrP Sc typing controversies, and further contribute to the understanding of the molecular basis of TSE strains and phenotypic variability, we examined the effect of different experimental conditions on the characteristics of PrP Sc fragments associated with the sCJD subtypes and vCJD using a high resolution gel electrophoresis system. Particular emphasis has been given to the study of the interplay between the effects of homogenate pH and PK concentration/activity.

EXPERIMENTAL PROCEDURES
Patients and Tissues-We studied 75 sCJD cases and 4 vCJD cases phenotypically characterized in regard to clinical and histopathological features, pattern of PrP deposition, PRNP genotype, and Western blot profile of PrP Sc . Sporadic CJD subtypes were classified according to Parchi et al. (9). They included 40 MM1, 5 MV1, 10 VV2, 10 MV2, 5 VV1, and 5 MM2-cortical. Brain tissues were obtained at autopsy and were kept frozen at Ϫ80°C until use. Brain samples used were from the frontal cerebral cortex, usually the middle frontal gyrus. In 5 MM1 cases tissue was also obtained from the putamen, entorhinal cortex, hippocampus, and amygdala.
Molecular Genetics-Genomic DNA was extracted from blood or frozen brain tissue. Genotyping of the PRNP coding region was performed as described (9).
According to their pK a values, PBS buffers were prepared between pH 5.5 and 8.0, lysis buffers were prepared between at pH 6.7 and 8.0, and the citrate-phosphate buffer was prepared at pH 4.0, 5.0, and 5.5. Because the pH of Tris buffers changes significantly according to the buffer temperature, the lysis buffers were titrated to the desired pH value at 37°C (i.e. the temperature at which protease digestion is performed). Because sodium deoxycholate is known to precipitate at acidic pH (around 6.3) it was not used at pH values lower than 6.5. Experiments exploring the influence of EDTA on PK digestion were performed using 250 mM EDTA stock solutions at pH 7.0 or 8.0.
Samples were treated at 37°C with PK (20 units/mg, Roche Diagnostics), chymotrypsin (1500 units/mg from bovine pancreas; Calbiochem), or cathepsin L (6500 milliunits/mg, 1.1 mg/ml from bovine kidney; Calbiochem) using various concentration/incubation time combinations (concentration range for PK, 7-10,000 g/ml; incubation time range for PK, 1-15 h). PK stock solutions (10 mg/ml or higher) were prepared in storage buffer (50% glycerol, 10 mM Tris, pH 7.5, 2.9 mg/ml CaCl 2 ). Small aliquots were prepared and stored at Ϫ20°C. For each experiment a new aliquot was used. Protease digestion was terminated by the addition of 2 mM phenylmethylsulfonyl fluoride. The pH of tissue homogenates was measured in duplicate using a needle electrode (Hamilton).

Analysis of PrP Sc Type 1 in sCJD MM Subjects
The Effect of Gel Resolution-To examine the gel mobility of human PrP Sc type 1 in more detail than in previous studies we compared the Western blot profiles obtained in 12% Tris-glycine PAGE gels 5.5 cm long (gels commonly used for PrP Sctyping analysis) with those obtained using longer gels (15 cm). In the latter condition, PrP Sc extracted from sCJD MM1 showed a more significant heterogeneity. This related either to the number of bands or to the electrophoretic mobility of the most represented bands (Fig. 1, a and b).
The Effect of pH-When we compared MM1 samples prepared in LB pH 7.4 with those homogenized in LB pH 8.0 we noticed a higher heterogeneity in gel migration in the first condition ( Fig. 1). This observation prompted us to measure the pH of the homogenates. Samples prepared using standard PBS or lysis buffer at pH 7.4 (i.e. buffers that have been used to date for PrP Sc -typing studies) showed unexpected heterogeneous pH values ranging from 6.48 to 7.46 (Table I). Similar pH variations were also detected in homogenates prepared from different areas of the same brain (Fig. 2). Homogenates prepared in water were obviously even more acidic (Table I). Interestingly, there was a significant correlation between the pH value of the homogenate and the immunoblot profile of each type 1 sample (Fig. 1). This was also true for samples obtained from different areas of the same brain (Fig. 2). Thus, samples from the same brain may also show a certain degree of heterogeneity in brain homogenate pH as well as electrophoretic mobility of PrP Sc core fragments.
To analyze in detail the effect of the homogenate pH on the characteristics of PrP Sc fragments generated by PK digestion, we had to stabilize the pH of the homogenate at the desired value. To this aim, we raised the buffer capacity of the lysis buffer, PBS, and citrate-phosphate buffer solutions and obtained the desired effect by increasing 10-fold either the Tris and the phosphate or citrate-phosphate salt concentration. We found that in sCJD MM1 cases the immunoblot profile of the PrP Sc core fragments generated by PK digestion in standard conditions (100 g/ml at 37°C for 1 h) varied significantly according to the homogenate pH and showed a shift in gel mobility between pH 6.0 and 8.0 (Fig. 3). At a pH between 6.0 and 6.7 the protein resolved in four distinct fragments. By increasing the pH values of the homogenate we obtained a progressive disappearance of the slowest migrating peptides associated with a parallel increase in the amount of the fastest migrating fragments together with the appearance around pH 7.2 of a novel fragment that migrated even faster. Finally, an additional (6th) slower migrating fragment was detected after digestion at a very acidic pH (5.0 -5.5) (Fig. 3), whereas no digestion at all was observed at pH 4.0 (data not shown). Thus, the PrP Sc core heterogeneity observed in MM1 subjects using homogenates prepared in standard buffers (LB with 10 mM Tris or standard PBS, pH 7.4), as used in previous studies (9,24,26,36), was clearly related to the variability of the homogenate pH ( Figs. 1-3).
The Effect of PK Activity-Given that the pH optimum for PK is in the basic pH range, we asked whether the multiple PrP Sc fragments generated by protease digestion at acidic pH may be at least in part the consequence of a reduced PK activity. By exposing the samples to different PK concentrations we found that the same heterogeneity among PrP Sc fragments generated by PK digestion at various pH values could be obtained by changing the PK concentration at a given pH value. In particular, we observed the appearance of the "ladder effect" on type 1 samples digested at pH 8.0 by progressively decreasing the PK concentration (Fig. 4). Similarly, a complete disappearance of the PrP Sc fragments of higher molecular mass that are seen after digestion at acidic pH (6.7) was obtained by increasing the PK concentration 10 -15 times (Fig. 4).
To determine whether there is also an effect of pH on PrP Sc irrespective of its effect on PK activity, we used higher PK concentrations, varying them according to pH to compensate for the change in protease activity and avoid the complete PrP Sc digestion. Whereas at pH values below 7.2 PK digestion of PrP Sc extracted from sCJD MM1 showed a relative molecular mass of 21 kDa, at pH values higher than 7.3 the protein migrated about 1 kDa faster (Fig. 5). Both fragments were seen at pH 7.2-7.3. All MM1 (n ϭ 40) subjects analyzed showed this shift in gel migration. The shift in electrophoretic mobility did not disappear when we prolonged the PK incubation time to the almost complete PrP Sc digestion ( Fig. 10 and data not shown), further suggesting a direct effect of pH on PrP Sc independent from its effect on PK activity.
The Effect of Different Proteases-To further verify whether the strong influence of pH on the immunoblot profile of PKgenerated PrP Sc fragments was mainly related to changes in PK activity rather than to pH-induced changes in PrP Sc , we tested the effect of other proteases; that is, cathepsin L, whose optimum activity is at pH 5.5, and chymotrypsin, a serine protease like PK, whose optimum activity is around pH 8. We found that chymotrypsin is comparable with PK with respect to both the pattern of PrP Sc fragments generated and the sensitivity to different pHs (Fig. 6), whereas cathepsin L shows a pH-related effect that is just the opposite of those of PK and chymotrypsin (Fig. 6). Indeed, the concentration of cathepsin L needed to generate a single PrP Sc fragment was much higher at pH 6.8 than at pH 5.5. As with PK, however, the progressive reduction of the enzyme concentration was related to the appearance of multiple PrP Sc fragments (data not shown).
The rise to a fragment of lower relative molecular mass. Considering our results on the effect of pH and that EDTA, according to the protocol of Wadsworth et al. (26), is prepared in a basic solution at pH 8 and then added to the homogenate, we investigated whether a change in pH rather than the EDTA itself could be responsible for the observed shift in gel mobility of PrP Sc . Indeed, our results indicate that this is the case (Fig. 7). We measured the homogenate pH before and after the addition of EDTA and found that the addition of EDTA-prepared stock solution at pH 8.0 to standard PBS or LB, pH 7.4, significantly increased the pH value of the homogenate (about 0.7 pH units) (Fig. 7). Furthermore, we were not able to reproduce the effect of EDTA when the EDTA solution was adjusted to pH 7.0 before being added to the homogenate (Fig. 7). A potential influence of pH on the copper chelation activity of EDTA is unlikely, since the potential-pH diagram for the copper-EDTA system indicates that EDTA forms stable copper complexes over a wide range of pH values (37). In conclusion, our data argue that EDTA at a 20 mM concentration has no direct effect on the pattern PrP Sc type 1 cleavage by PK.

Analysis of PrP Sc Associated with sCJD Subtypes MV1, VV1, VV2, MV2, MM2-cortical, and vCJD
PrP Sc extracted from sCJD MV1 cases showed the same electrophoretic mobility of core fragments, including the pHdependent shift around pH 7.2 of the type 1 from MM cases ( Fig. 8 and data not shown). A shift in gel mobility that was clearly dependent on the homogenate pH and did not disappear FIG. 2. Regional variability in the electrophoretic mobility of PrP Sc type 1 (MM subject); comparison between PK digestions performed at different pH values. Immunoblot analysis of homogenates from five different brain regions of a MM1 subject is shown. a, homogenates were prepared in standard lysis buffer, pH 7.4. b, homogenates were prepared in standard lysis buffer, pH 8.0. All aliquots were digested with 100 g/ml of PK for 1 h at 37°C. The actual pH values of the homogenates are shown at the bottom of each lane. Samples were run in a Tris-glycine PAGE 12% mini-gel (5.5 cm long). PUT, putamen; MFG, middle frontal gyrus; EC, entorhinal cortex; Hipp, hippocampus; AMG, amygdala. Approximate molecular masses are in kilodaltons. The same result (with no obvious correlation between pH and type of brain area) was reproduced with samples from five MM1 subjects.

FIG. 3. Effect of pH on the electrophoretic mobility of PrP Sc type 1 (MM subject).
Immunoblot analysis of frontal cortex homogenates from a MM1 subject is shown. Homogenates were prepared in different solutions with high buffer capacities at pH values from 5.0 to 8.0 (citrate-phosphate buffer at pH 5.0 and 5.5; PBS with a 10ϫ increase in phosphate salt concentration plus 1% Sarkosyl at pH 6.0; lysis buffer with 100 mM Tris at pH Ն 6.7). All aliquots were digested with 100 g/ml of PK for 1 h at 37°C. The actual pH values of the homogenates are shown at the bottom of each lane. Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced at least twice with samples from 5 MM1 subjects.

FIG. 4. Effect of PK concentration on the electrophoretic mobility of PrP Sc type 1 (MM subject).
Immunoblot analysis of frontal cortex homogenates from a MM1 subject is shown. Homogenates were prepared in lysis buffer with 100 mM Tris at pH 6.7 (lane 1-5) and 8.0 (lane 6 -10). Aliquots were digested at 37°C for 1 h using different PK concentrations. Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced at least twice with samples from three MM1 subjects. with increasing PK concentrations was also observed in VV1 subjects. Indeed, PK digestion of VV1 samples at pH 6.9 generated a doublet of about 21 and 20 kDa even when a relatively high PK concentration was used, whereas at pH 8.0 only the fastest migrating fragment was detected (Fig. 8). Thus, the use of a 15-cm-long gel instead of the mini-gel distinguishes between PrP Sc type 1 from VV subjects and PrP Sc type 1 from MM or MV subjects (Fig. 8a).
In contrast, PrP Sc extracted from each of the sCJD subtypes associated with PrP Sc type 2 (VV2, MV2, MM2-cortical) and vCJD showed much less variability in immunoblot profile. Indeed, none of the type 2 CJD subtypes showed a pH-dependent shift in PrP Sc mobility that could be dissociated from PK activity. At a pH below 7.4, additional fragments migrating slower than 19 kDa were generated in all these CJD subtypes by a standard PK concentration (i.e. 100 g/ml); however, they disappeared once the PK activity was increased either by raising the enzyme concentration or by prolonging its incubation time (Figs. 9 and 10 and data not shown). The only exception to this rule was the PrP Sc type 2 from MV subjects, which resolved as a doublet of about 19 and 20 kDa even in the presence of a higher PK activity, indicating that the immunoblot profile of PrP Sc from MV2 subjects is distinct from those of sCJD MM2 or VV2 as well as vCJD (Figs. 8 and 10).

PrP Sc from Various CJD Subtypes Differs in the Degree of PK Resistance
The observation that the number of fragments obtained when PrP Sc was digested at acidic pH (at a given PK concentration) varied significantly among samples extracted from distinct sCJD subtypes raised the possibility that distinct PrP Sc types differ in the degree of protease resistance. To answer this question, we compared the effect of different PK concentrations at a given pH value. A qualitative analysis revealed that samples extracted from different CJD subtypes show varying degrees of protease resistance. In particular, PrP Sc extracted from VV2 and vCJD subjects showed the highest degree of protease resistance (Fig. 10). In contrast, the type 1 PrP Sc extracted from VV subjects was the most sensitive to PK digestion (Fig. 10).

The Effect of Various Detergents
To evaluate the influence of detergents on PrP Sc and PK activity we compared homogenizing buffers with different compositions. Sarkosyl (1%) prevented the pH-dependent shift in electrophoretic mobility seen in PrP Sc MM1/MV1 (data not shown). In addition, an indirect effect on PrP Sc digestion by PK was seen with Sarkosyl (1%) or sodium deoxycholate (5%) because in these conditions a higher concentration of PK was needed to obtain a complete PrP Sc digestion (data not shown).

DISCUSSION
The results of the present study shed light on previous disparities in PrP Sc typing, provide a refined classification of human PrP Sc types, and have significant implications for both the classification and the epidemiologic surveillance of TSEs in humans and animals.
We have shown that, due to the lack of sufficient buffer FIG. 5. Effect of pH on the electrophoretic mobility of PrP Sc type 1 (MM subject) in the presence of a high PK concentration. Immunoblot analysis of frontal cortex homogenates from a MM1 subject is shown. Homogenates were prepared in lysis buffer with 100 mM Tris at pH values from 6.8 to 7.6. Aliquots were digested with relatively high PK concentrations for 1 h at 37°C. Because the maximum efficiency of the enzyme is at pH 8, to avoid complete digestion of PrP Sc PK concentration was progressively decreased in parallel to the increase in pH values (lane 1-4, 10 mg/ml; lane 5-7, 2 mg/ml; lanes 8 and 9, 1 mg/ml). Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced at least twice with samples from 3 MM1 subjects.
FIG. 6. Effect of pH on the digestion of PrP Sc type 1 (MM subject) by cathepsin L and chymotrypsin. Immunoblot analysis of frontal cortex homogenates from a MM1 subject is shown. Homogenates were prepared at different pH values using different buffers (citrate phosphate pH 4.0 and pH 5.0; PBS with a 10ϫ increase in phosphate salt concentration pH 6.5; lysis buffer with 100 mM Tris pH 6.8). Digestion was performed at 37°C for 1 h with chymotrypsin (100 g/ml) (lanes 1-4) or cathepsin L (cathepsin L, 0.15 l/l) (lanes 5-7). Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced at least twice with samples from 3 MM1 subjects. capacities of standard Tris or PBS buffers, there is a significant heterogeneity in pH among CJD brain homogenates even when they are prepared from the same brain. The homogenate pH in turn influences the size of the PrP Sc core generated by protease digestion due to the combination of at least three factors. (i) Protease digestion of full-length PrP Sc is a step-by-step process yielding fragments with various degrees of resistance, (ii) PK activity is pH-dependent, (iii) in sCJD MM1, MV1, and to a lesser extent in VV1, the size of PrP Sc further changes depending on whether protease digestion is performed at a pH below or above 7.2.
Currently, there is a significant disparity in the literature regarding the existence of distinct human PrP Sc types. After the original identification by Parchi et al. (24,25) of two major PrP Sc types (named type 1 and type 2) in sporadic human TSEs, two additional classifications of human PrP Sc types have been proposed (26,27). According to Wadsworth et al. (26) PrP Sc type 1 from MM subjects can be distinguished into 2 subtypes based on a 0.5-kDa difference in the relative molecular mass of PrP Sc . However, when pK digestion is performed in the presence of at least 20 mM EDTA, the two PrP Sc subtypes show the same electrophoretic mobility, with an estimated shift of about 1.100 and 0.650 kDa for the first and second type, respectively. Zanusso et al. (27) also proposed that the Parchi et al. (24) PrP Sc type 1 includes two protein subtypes having distinct conformations and strain-specific properties. However, at variance with Wadsworth et al. (26), they distinguished the two PrP Sc subtypes based on their sensitivity to pH variations rather than on their presumed degree of binding to metal ions. According to the study by Zanusso et al. (27), about 60% of type 1 samples show a pH-dependent shift in electrophoretic mobil-ity, whereas the remaining 40% appear unaffected by pH variations.
In our original descriptions of human PrP Sc types (23,24) we also noticed a certain degree of PrP Sc heterogeneity among subjects belonging to the same sCJD subtype (i.e. MM1, VV2, etc.), particularly among MM1 subjects. However, at variance with the above-mentioned studies, we did not attempt to further subclassify the human PrP Sc types based on the observations that (i) the same degree of variability was also seen in samples taken from the same brain, and ii) this molecular heterogeneity did not correlate with any distinct phenotypic feature.
By demonstrating that the heterogeneity of human PrP Sc type 1 within specific groups (MM or MV subjects) strictly depends on pH variations among CJD brain homogenates, the present study provides a reasonable technical explanation for the PrP Sc molecular heterogeneity associated with sCJD MM1 or MV1. Thus, we believe that there is no basis at present for claiming that more than one CJD subtype or CJD strain is associated with the molecular combination codon 129 MM genotype and PrP Sc type 1. Indeed, our results clearly show that, once the experimental conditions (i.e. homogenate pH, PK concentration, type of detergents) are carefully controlled, the electrophoretic mobility of PrP Sc type 1 from MM subjects is highly homogenous. In this respect, it is also noteworthy that experimental transmission of prions from pure MM1 subjects to non-human primates and mice have given to date homogenous results in terms of incubation time or lesion profile (14,16,17). These results are in keeping with the original classification of Parchi et al. (9) and strongly suggest that only a single major agent strain is associated with the sCJD MM1 subtype.
We cannot provide a definite explanation for the finding of Zanusso et al. (27), who observed a pH-dependent shift in electrophoretic mobility in only about 60% of MM1 subjects. It is of significance, however, that they did not perform any direct measurement of actual pH in the homogenates. Thus, they may unwittingly have compared PrP Sc core fragments generated at different pHs. It is also noteworthy that Wadsworth et al. (26), in perfect agreement with our results, reported a shift in electrophoretic mobility in all their sCJD MM subjects with either PrP Sc type 1 or type 2 (i.e. corresponding to our type 1) when they were treated with 20 mM EDTA at pH 8 before PK digestion. However, according to our results the change in PK cleavage depends on homogenate pH variations rather than on EDTA.
Our finding of decreased pH values in brains from patients with sCJD deserves further comment. The observation raises the interesting question of whether the acidic pH is somehow related to the disease pathogenesis. Previous studies document a decline in post-mortem human brain pH values that positively correlate with age at death and agonal state severity (38,39). Moreover, we found that in non-human primates experimentally infected with prions who did not suffer from an agonal FIG. 8. Distinction of subtypes of PrP Sc type 1 and type 2. Immunoblot analysis of frontal cortex homogenates from MM1, MV1, VV1, MM2-cortical, VV2, MV2, and vCJD subjects is shown. Homogenates were prepared in lysis buffer with 100 mM Tris at pH 6.9 (a) or pH 8.0 (b). Aliquots were digested with 2000 g/ml of PK for 1 h at 37°C. Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced twice with samples from at least four subjects of each group.
FIG. 9. Effect of different PK concentration/pH combinations on the electrophoretic mobility of PrP Sc type 2. Immunoblot analysis of frontal cortex homogenates from an MM2-cortical subject is shown. Homogenates were prepared in lysis buffer with 100 mM Tris at pH 6.7 (lanes 1-2), 7.0 (lanes 3-5), 7.4 (lanes 6 -8), and 8.0 (lanes 9 -11). Aliquots were digested at 37°C for 1 h using different PK concentrations. Samples were run in a Tris-glycine PAGE 15% gel (15 cm long). Approximate molecular masses are in kilodaltons. The same result was reproduced at least twice with samples from 3 MM2-cortical subjects. state, the brain pH was much less acidic (data not shown). Further studies on brain metabolites in vivo, however, will be required to provide a definitive answer to this issue.
Our data also shed light on the current understanding of the molecular basis of human TSE strains and phenotypic variability. Several lines of evidence indicate that the pathological heterogeneity of human TSEs is related to the physicochemical properties of protease-resistant prion protein (24,25). However, whether each phenotype is indeed related to a specific PrP Sc tertiary or quaternary structure is still unknown. In sporadic CJD five phenotypes have been characterized, but only two major PrP Sc types have been consistently reproduced based on difference in the relative molecular mass of PrP Sc (9). By using gels with improved resolution and different experimental conditions, we have now shown that PrP Sc type 1 in codon 129 MM or MV subjects has distinct physicochemical properties from the type 1 in the VV homozygotes. In addition, only in MM1/MV1 and VV1 samples did PrP Sc show a pH-dependent shift in gel mobility. Our data also show that distinct physicochemical properties characterize the sCJD subtypes associated to PrP Sc type 2. For example, PrP Sc from MV subjects with kuru plaques is clearly distinguishable from the other type 2s because it uniquely resolves as a doublet, even when a relatively high pK concentration is used. Furthermore, we have shown that PrP Sc from VV2 subjects as well as vCJD display a significantly higher protease resistance than the PrP Sc from MM2-cortical cases. Thus, analysis of protease resistance may allow the distinction between VV2 and MM2-cortical cases without PRNP genotyping and further helps, in combination with the study of PrP Sc glycoform ratio (36), to distinguish vCJD from other CJD MM2 subtypes. Taken together, these data demonstrate that there are specific PrP Sc properties associated with each sCJD pathological phenotype as well as vCJD, supporting the notion that at least in human TSEs the pathological phenotype is related to differences in PrP Sc molecular structure.
By improving the molecular distinction between sCJD subtypes, our data also have important implications for the future epidemiological studies into the etiology of sCJD. Indeed, the epidemiology of sCJD to date has been analyzed considering sCJD as a single entity, which might have prevented the discovery of risk factors associated with single or subgroups of sCJD subtypes. Finally, it is reasonable to believe that our study will also contribute to standardizing and harmonizing PrP Sc typing among laboratories, which will represent the basis for the new epidemiologic approach. In this regard we wish to underline the protocol, which in our opinion gives the best visualization of the heterogeneity in electrophoretic mobility (typing) among human PrP Sc subtypes (Fig. 8a). This includes (i) a homogenizing solution with strong buffer capacity such as our LB with 100 mM Tris, (ii) protease digestion at pH 6.9 using a relatively high PK concentration (i.e. 2 mg/ml), and (iii) the use of 15% Tris-glycine gels with a 15-cm long running gel.