Protease resistance of infectious prions is suppressed by removal of a single atom in the cellular prion protein

Resistance to proteolytic digestion has long been considered a defining trait of prions in tissues of organisms suffering from transmissible spongiform encephalopathies. Detection of proteinase K-resistant prion protein (PrPSc) still represents the diagnostic gold standard for prion diseases in humans, sheep and cattle. However, it has become increasingly apparent that the accumulation of PrPSc does not always accompany prion infections: high titers of prion infectivity can be reached also in the absence of protease resistant PrPSc. Here, we describe a structural basis for the phenomenon of protease-sensitive prion infectivity. We studied the effect on proteinase K (PK) resistance of the amino acid substitution Y169F, which removes a single oxygen atom from the β2–α2 loop of the cellular prion protein (PrPC). When infected with RML or the 263K strain of prions, transgenic mice lacking wild-type (wt) PrPC but expressing MoPrP169F generated prion infectivity at levels comparable to wt mice. The newly generated MoPrP169F prions were biologically indistinguishable from those recovered from prion-infected wt mice, and elicited similar pathologies in vivo. Surprisingly, MoPrP169F prions showed greatly reduced PK resistance and density gradient analyses showed a significant reduction in high-density aggregates. Passage of MoPrP169F prions into mice expressing wt MoPrP led to full recovery of protease resistance, indicating that no strain shift had taken place. We conclude that a subtle structural variation in the β2–α2 loop of PrPC affects the sensitivity of PrPSc to protease but does not impact prion replication and infectivity. With these findings a specific structural feature of PrPC can be linked to a physicochemical property of the corresponding PrPSc.

Introduction Transmissible spongiform encephalopathies are fatal neurodegenerative diseases associated with the presence of prions [1]. It is well-established that prions contain aggregates of misfolded cellular prion protein (PrP C ). Such aggregates have been termed "PrP Sc " because they were initially found in scrapie-affected sheep, or alternatively as "PrP res " to denote their extraordinary resistance to proteolytic digestion. Indeed, the carboxy-proximal core of PrP Sc can largely withstand proteolysis with 500 μg/ml proteinase-K (PK) [2,3], and the detection of PK-resistant prion protein is commonly regarded as the definitive diagnostic method for prion diseases in humans and other species.
However, variably protease sensitive prionopathy (VPsPr) has been discovered in humans, characterized by atypical patterns of PrP Sc detected by Western blots [4]. PK-treated PrP from VPsPr shows bands in the range from 8 to 17 kDa, which are not observed in sporadic Creutzfeldt-Jakob Disease (sCJD). Furthermore, recent studies have identified prion diseases with high titers of prion infectivity in the absence of protease resistance [5]. These observations point to heterogeneity in the structural arrangement of PrP aggregates, with looser aggregates being more solvent-accessible and therefore prone to proteolysis.
Here we investigate the phenotypes arising in transgenic mice that express a minimal modification of the residue in position 169, i.e., replacing tyrosine (Y) by phenylalanine (F). It should be stressed, in view of the unexpected results obtained in these experiments, that this variation of the protein structure includes the elimination of a single oxygen atom from the wt structure of mouse PrP. 169F ) and wild-type C57BL/6 mice results in similar phenotypes

Prion infection of Tg(MoPrP
To investigate the biological effect of the hydroxyl group at position Y169 (Fig 1A), we generated mice expressing MoPrP 169F using a "half-genomic" Prnp minigene construct [23]. B6D2 hybrid zygotes were microinjected with the hg-MoPrP 169F construct, and four independent transgenic mouse lines were generated. Lines 171 and 173 were selected for mating with PrPdeficient Prnp -/mice [24] over multiple generations, resulting in mice expressing only the transgene (MoPrP 169F ;Prnp -/-) or both the transgene and endogenous PrP C in a hemizygous state (MoPrP 169F ;Prnp +/-).
Line 171 showed the highest transgene expression in the brain, with mice hemizygous for the transgene showing a PrP protein expression slightly higher than C57BL/6 wild-type (wt) mice ( Fig 1B, S1 Fig and S12 Fig), and was therefore chosen for further experiments. Clinical and histological examinations failed to reveal any pathological phenotype or deviation from the characteristics of wt mice up to the age of 2 years and 6 months (S2 Fig). In particular, in 19 month old littermates, myelination of the sciatic nerve was similar to that of hemizygous Prnp +/mice, whereas Prnp -/mice displayed a chronic demyelinating neuropathy (Fig 1C) [25]. Since the latter is a universal phenotype common to all investigated Prnp -/mouse strains, we concluded that MoPrP 169F was physiologically equivalent to wt PrP C .
We then sought to assess whether these mice showed any alteration of prion susceptibility after PrP Sc infection. After intracerebral (i.c.) administration of 3 x 10 5 ID 50 units of prions (Rocky Mountain Laboratory strain, passage 6, henceforth denoted RML), both wt and Tg (MoPrP 169F ) mice developed signs of scrapie with similar latencies (Fig 2A), whereas Prnp -/mice died of intercurrent disease without developing scrapie. Reinoculation of 1% brain homogenates from prion-infected wt and Tg(MoPrP 169F ) mice into Tga20 mice, which express Prnp -/-, wild-type and Prnp -/mice after intracerebral (i.c.) inoculation of brain homogenates from 263K-inoculated Tg81 mice. Mice did not develop any clinical signs of a prion disease within 500 days post inoculation. Letters e, f, g and h within the graph indicate mice that were used for subsequent passaging. (C) Passage of brain homogenates from RML-infected Tg(MoPrP 169F );Prnp -/-(mice a and b), Tg(MoPrP 169F );Prnp +/-(mouse c) and wt Prnp +/+ (mouse d) animals into Tga20 mice evoked a prion disease with median survival times between 79 and 97 days. Tga20 mice reinoculated with noninfectious brain homogenates (NBH) from either Tg(MoPrP 169F );Prnp -/-, Tg(MoPrP 169F );Prnp +/or Prnp -/mice did not show any signs of a prion disease (h.i. = heat inactivated). (D) Same as (C), but passage of brain homogenate from 263K-infected Tg(MoPrP 169F );Prnp -/-(mice e, f and g) animals into Tg81. Tg81 mice inoculated with brain homogenate from mouse "g" developed prion disease with a median survival time of 91 days, whereas for the other brain homogenates a median survival time was not reached.
We next sought to determine whether the MoPrP 169F mutation could have an influence on the species barrier. After intracerebral inoculation with hamster-adapted prions (strain 263K [26]; 3 x 10 6 ID 50 units), mice failed to develop signs of disease within 500 days of observation ( Fig 2B). However, brain homogenates from these animals inoculated into Tg81 mice overexpressing Syrian hamster Prnp [27] occasionally led to prion disease and death, similarly to what was reported for wt mice [26] (Fig 2D and S3B Fig).
We next asked whether the MoPrP 169F mutation would have any effect on antibody-mediated neurotoxicity [28,29]. We therefore injected the monovalent fragment Fab 1 of the antiprion antibody POM1 into the brains of Tg(MoPrP 169F );Prnp -/and C57BL/6 wt mice (S4 Fig). Transgenic and wt mice showed similar neurotoxic effects, which could be suppressed by preincubating Fab 1 -POM1 with a 5-fold molar excess of recombinant murine PrP fragment encompassing residues 90-231.
PrP Sc of RML infected Tg(MoPrP 169F );Prnp -/mice is protease sensitive Treatment of brain homogenates from RML-infected, approximately 1-year-old, terminally sick Tg(MoPrP 169F );Prnp -/mice with PK (25 μg/ml, 37˚C, 30 min) led to almost complete loss of PrP immunoreactivity in Western blots ( Fig 3A and S10A Fig), indicating that very little or no PrP res was formed. Further assessment of PK sensitivity with decreasing PK concentrations showed that PrP res from Tg(MoPrP 169F );Prnp -/mice was sensitive to PK treatment down to a concentration of 6.25 μg/ml (Fig 3B and S10B Fig). In contrast, homogenates of terminally scrapie-sick wt mice showed PrP res even after treatment with a 20-fold higher PK concentration ( S5 Fig and S13 Fig). To assess the presence of shorter PK-resistant fragments that might go undetected by POM1, we also probed the Western blots with a panel of additional anti-PrP antibodies (POM3, POM5, POM6, POM15 and POM19) targeting different regions of the PrP molecule [28]. All antibodies recognized similar differences in PK sensitivity between RML infected wt-and Tg(MoPrP169F);Prnp -/animals (S6 Fig and S14 Fig) and no alternative bands. Tg(MoPrP 169F );Prnp +/mice showed reduced levels of PrP res compared to wt animals, yet showed the same pattern of enhanced electrophoretic motility resulting from N-terminal proteolysis and the typical triplet bands corresponding to unglycosylated, monoglycosylated, and diglycosylated isoforms (Fig 3, S7 Fig, S10 Fig and S15 Fig). Trypsin treatment also resulted in a major reduction of signal in transgenic mice but not in wt PrP C expressing animals ( Fig 3C and S10C Fig). However, thermolysin (TL) treated samples (100 μg/ml) retained PrP immunoreactivity (Fig 3D and S10D Fig), as published previously for PK sensitive PrP [30], albeit without the characteristic shift in gel mobility caused by PK.
To determine whether the increased PK sensitivity of MoPrP 169F prions was heritable (and hence indicative of a new prion strain), we passaged MoPrP 169F prions into Tga20 mice and investigated the properties of the resulting prions in a conformational stability assay [31] ( Fig  3E and S10E Fig). The GdnHCl denaturation curves showed no significant difference (p = 0.8526, unpaired t-test) between RML and MoPrP 169F prions passaged through Tga20 mice (Fig 3E and 3F and S10E Fig). Crucially, infection of Tga20 mice with MoPrP 169F prions caused the generation of PrP Sc with PK resistance. Similar results were seen after passage of 263K-infected transgenic mice into hamster PrP-expressing Tg81 mice (S8 Fig and S16 Fig).
Hence passage of PK-sensitive Tg(MoPrP 169F );Prnp -/prions into Tga20 or Tg81 mice led to reappearance of the original strain properties. PK sensitivity of MoPrP 169F prions is therefore a phenotype due to recruitment of host expressed PrP C into PrP Sc , rather than a prion-encoded strain.

Decreased high-density aggregates in RML-inoculated Tg(MoPrP 169F ) mice
To better understand the physical and morphological properties of the pathological MoPrP 169F in RML-infected mice, we analyzed brain homogenates by density gradient ultracentrifugation followed by Western blotting of each fraction [34]. In uninfected mice, most PrP was present in the low-density fractions (1)(2)(3)(4)(5), whereas all RML inoculated mice showed additional PrP signals in high-density fractions (10-20) (Fig 5A, S9A Fig, S11 Fig and S17 Fig). In contrast to RML-inoculated wt Prnp +/+ and Tg(MoPrP 169F );Prnp +/mice, animals expressing only the transgenic PrP showed a strong reduction of high-density PrP aggregates. PK digestion confirmed the presence of PrP res in the high-density fractions (S9B Fig and S18 Fig) and reduced amounts of PrP res in Tg(MoPrP 169F ).
The relative quantity of PrP signals was calculated, with the highest amount of dense PrP deposits in RML-inoculated wt Prnp +/+ mice followed by transgenic mice with a hemizygous Prnp background (Fig 5B). In transgenic mice without endogenous PrP C expression, only weak bands of PrP deposits in the high-density range could be detected. These findings indicate that Tg(MoPrP 169F );Prnp -/mice contain reduced amounts of compact PrP aggregates in the brain. (A) Western blot analysis of brain homogenates from RML-inoculated Tg(MoPrP 169F ) mice showed increased sensitivity to PK digestion (25 μg/ml PK for 30 min at 37˚C) compared to wt Prnp +/+ mice. (B) A gradient of different PK concentrations showed that PrP Sc in the brain homogenates from Tg(MoPrP 169F );Prnp -/was sensitive to PK concentrations up to 6.25 μg/ml. (C) Same as (B) for trypsin (Try) digestion. Brain homogenates (10%) from Tg(MoPrP 169F );Prnp -/mice showed a reduction of PrP signal after trypsin treatment up to a concentration of 2.5μg/ml. (D) Same as (B) for thermolysin (TL) digestion. Thermolysin showed no significant effect on PrP Sc in the brain homogenate from Tg(MoPrP 169F );Prnp -/at a concentration of 100μg/ml (70˚C, 30 min), which is in contrast to 10% brain homogenate from wild-type Prnp +/+ mice inoculated with NBH, where no PrP could be detected at this concentration. (E) Western blot analysis of brain homogenates from Tga20 mice infected with prions from Tg(MoPrP 169F );Prnp -/or wt mice. Homogenates were treated with increasing GdnHCl concentrations and PK-digested. (F) Quantitative analysis of the Western blot data shown in panel E. Each data point is the mean of three biological replicates ± SD. WT and PrP 169F prions showed similar GdnHCl stability (p = 0.8526, unpaired t-test). The antibody POM1 was used for detection. Molecular sizes are indicated in kDa. NBH: noninfectious brain homogenate. doi:10.1371/journal.pone.0170503.g003

Discussion
One major goal of prion research is to better understand the relationship between PrP Sc structure and pathogenesis. A powerful approach towards this goal relies on developing variants of PrP C with well-defined structural features and molecular dynamics, and analyzing the consequences of their in vivo expression under conditions of health and disease. The design of such in vivo experiments is supported by the availability of atomic-resolution PrP C structures [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]. Here we focused on the β2-α2 loop because its conformation affects cross-species transmission of prions and is even associated with the spontaneous generation of prions [35][36][37].
The replacement of tyrosine with phenylalanine at position 169 amounts to the removal of a single oxygen atom from the tyrosine residue. NMR studies showed that the conformation of the β2-α2 loop is unaffected by this amino acid exchange, whereas substitution by alanine or glycine results in a major rearrangement of the 3 10 -helical β2-α2 loop to a type-I β-turn conformation [20]. The Y169F mutation decreases the melting temperature by only 2.6˚C [21], whereas for the variants with alanine or glycine at position 169 the melting temperature was lowered by ca. 10˚C [21]. The substitution of Y169 by glycine in Tg(MoPrP 169,170,174 ) [35] mice also prevented the development of the spontaneous TSE that was observed in Tg(MoPrP 170,174 ) mice expressing the double mutation S170N and N174T [36,37]. Tg(MoPrP 169,170,174 ) mice additionally showed a markedly enhanced transmission barrier [35], suggesting that residue 169 plays a crucial role in prion conversion, as was also demonstrated by in vitro conversion experiments [38].
After inoculation with two different prion strains, RML and 263K, mice expressing the Y169F mutant developed prion infectivity similarly to wt mice. Moreover, prions extracted from the brains of these mice were transmissible to Tga20 and Tg81 mice, respectively. This indicates that the Y169F variant prion protein is fully competent to enable the acquisition and multiplication of prions as well as the transmission of de novo synthesized prions to additional hosts.
In addition to typical clinical signs of scrapie, infected Y169F mice showed neuronal vacuolation and extensive astrogliosis similar to wt mice, and RML-infected mice developed polythiophene-stainable plaques [32]. Moreover, injection with Fab 1 -POM1 induced similar toxicity in Y169F transgenic mice and wt mice. Finally, expression of MoPrP 169F suppressed the progressive demyelination seen in Prnp -/mice [25]. We conclude that both the physiological and pathogenic properties of MoPrP 169F are indistinguishable from those of wt MoPrP.
Despite their unabated pathogenicity, MoPrP 169F prions were surprisingly sensitive to PK and trypsin, and were less dense/smaller when compared to RML prions. Rather than representing a strain shift, these traits were host-encoded, and the original strain properties of RML re-emerged after passage of MoPrP 169F into Tga20 mice (Fig 6). Furthermore, co-expression of PrP 169F did not conspicuously interfere with the conversion of wt PrP C into PrP Sc . This is in contrast to other PrP variants which often exert a dominant-negative effect on prion ;Prnp +/-, Prnp -/and Prnp +/+ mice, as well as Tga20 and Tg81 mice inoculated with Tg(MoPrP 169F );Prnp -/brain homogenates. Hallmarks of prion disease including vacuolation (hematoxylin and eosin; HE), astrogliosis (glial fibrillary acidic protein; GFAP) and microglial activation (activated microglial marker; IBA1), were detected in terminally sick Tg(MoPrP 169F ) and RML-infected wild-type Prnp +/+ mice. RML-infected wild-type (wt) mice showed the typical diffuse synaptic pattern of SAF84 positive signals (first row), and Tg(MoPrP 169F );Prnp +/showed some plaque-like deposits (asterisk; third row). Few weakly stained PrP Sc deposits were visible in Tg(MoPrP 169F );Prnp -/and no PrP signals in Prnp -/brains (rows 2 and 4). After passage of Tg(MoPrP 169F );Prnp -/brain homogenate into Tga20 mice, synaptic and plaque-like PrP deposits were observed in the cortex (5 th row). Passage of 263K-infected Tg(MoPrP 169F );Prnp -/into Tg81 mice resulted in a prion disease with PrP plaque formation in the cortex (6 th row). Scale bar: 50 μm. (B) Frozen section (10 μm) of Tg(MoPrP 169F );Prnp -/brain showed LIN5050 stained aggregates. Fluorescence emission spectra were recorded in 5 regions (M0-M4) at 500-800 nm and compared to the surrounding tissue (M5-M6) (magnification 100x). (C) Same as (B) but from a Tg(MoPrP 169F );Prnp +/mouse (magnification 40x). propagation [39][40][41]. Therefore, a subtle structural variation in the β2-α2 loop of PrP C affected sensitivity of PrP Sc to proteases and reduced the amounts of large/compact aggregates arising after prion infection. Although the overall structure of monomeric MoPrP 169F is similar to that of wt PrP C , the Y169F mutation causes the loss of the side-chain hydrogen bond formed between Tyr169 and Asp178, which stabilizes the π-stacking interactions formed between the aromatic residues Tyr169, Phe175, and Tyr218 [21,42,43]. MD simulations for PrPs carrying pathological mutations demonstrated that a loss of the π-stacking interactions could lead to a rearrangement of residue 169 to a more solvent exposed trans conformation, which affects the distance between the β2-α2 loop and α3 helix and leads to a more flexible α3 helix [21,42,43]. Further simulations also showed that this effect can be determined by the presence of either methionine or valine at position 129, as Tyr128 forms a hydrogen bond with Asp178, which can be affected by the kind of amino acid at position 129 [43]. Hence, a plausible explanation might be that a modulation of this network in PrP C also affects the aggregational PrP Sc states, leading to looser packing and increased PK sensitivity of MoPrP 169F aggregates. This is also in agreement with the feature of MoPrP 170,174 prions being mostly PK-sensitive after passaging them into Tga20 mice [36], as the conformation and dynamics induced by residue 170 also plays a crucial role in the preservation of this network [44].
The physical basis of PK resistance is likely to reside in steric crowding, which would limit contacts with solvent water and the protease. The removal of the hydroxyl group at residue 169 could thus ease the accessibility of water and proteolytic enzymes, and loss of packing density resulting from the elimination of the-OH group could provide a rationale for the results of density-gradient centrifugation ( S9A Fig and S17 Fig). The reduced density of MoPrP 169F prions is also in agreement with other studies, which demonstrated that PK-sensitive forms of PrP Sc are found in lower-density fractions. This could also be rationalized by the assumption that PK-sensitive forms consist of smaller aggregates of PrP molecules [45][46][47]. Our findings thus yield another example of a prion disease that generates infective PK-sensitive aggregates, as was previously seen in several familial human prion diseases and related mouse models [37,[48][49][50][51][52].
In conclusion, our findings demonstrate that a subtle change in the prion protein covalent structure, i.e. the removal of a single oxygen atom, has a significant effect on a physicochemical property of PrP Sc without abolishing its infectivity. Further studies using structurally welldefined PrP variants may yield direct insights into prion generation, thereby advancing the development of novel diagnostic and therapeutic procedures.

Generation of Tg(MoPrP 169F ) mice
Half-genomic PrP vector was genetically modified to exchange tyrosine of PrPY169 against phenylalanine by using the Quick change lightning site-directed mutagenesis kit (Stratagene) and the primer set fwd: AGG CCA GTG GAT CAG TTC AGC AAC CAG AAC AAC. Rev: GTT GTT CTG GTT GCT GAA CTG ATC CAC TGG CCT. PCR was performed according to the company´s protocol. The linearized transgenic construct (NOT1/ Sal1) was injected into 296 fertilized Prnp +/+

Fig 5. Brains of RML-infected Tg(MoPrP 169F );Prnp -/harbor reduced amounts of high density PrP. (A)
Western blot analysis of differentially fractionated brain homogenates from RML-infected Tg(MoPrP 169F );Prnp -/and Tg(MoPrP 169F );Prnp +/mice showed reduced signals of high density PrP aggregates compared to RML-infected wild-type (wt) mice (fractions 10 to 20; OptiPrep 7 to 28%). Wild-type Prnp +/+ mice inoculated with noninfectious brain homogenate (NBH) displayed only signals in the low density fractions (Fractions 2 and 4). Every second fraction was analyzed by Western blotting. PrP was detected by the anti-PrP antibody POM1. Molecular sizes are indicated in kDa. (B) Quantification of the PrP signal, using two technical duplicates per mouse (Western blots are shown in Fig 5A and S9 Fig; mean ± SEM).

Quantification of transgene expression using Fö rster Resonance Energy Transfer (FRET)
A FRET based assay was established for the purpose of quantifying transgene expression by measuring PrP levels of 10% brain homogenates from Tg(MoPrP 169F ); Prnp -/-, wt Prnp +/+ , wt Prnp +/and Prnp -/mice. Europium (Eu 3+ ) donor and allophycocyanin (APC) acceptor fluorophores were coupled to anti-PrP holoantibodies POM1 and POM2 recognizing the globular domain and the octarepeats, respectively. The donor POM1-Eu 3+ conjugate is excited at wavelength 340 nm and transfers energy to the acceptor conjugate POM2-APC when the distance between acceptor and donor is <10 nm. POM2-APC then emits light at wavelength 665 nm, which can be measured with a suitable time-resolving spectrofluorimeter. For PrP level detection in homogenates, the Eu 2+ -POM1 and APC-POM2 antibody pair was added, measured immediately using a FRET reader and normalized to total protein.

Animal experimentation, animal welfare and ethics statement
All animal experiments were conducted in strict accordance with the Rules and Regulations for the Protection of Animal Rights (Tierschutzgesetz and Tierschutzverordnung) of the Swiss Bundesamt für Lebensmittelsicherheit und Veterinärwesen BLV. All animal protocols and experiments performed were specifically approved for this study by the responsible institutional animal care committee, namely the Animal Welfare Committee of the Canton of Zurich (permit numbers 41/2012). All efforts were made to minimize animal discomfort and suffering.

Prion inoculation
Transgenic mice (F3), Prnp -/mice or B6 Prnp +/+ mice were anaesthetized using isofluorane and inoculated with 30μl of 0.1% brain homogenate, diluted in sterile PBS/ 5% BSA, from Rocky Mountain Laboratory infected CD1 (3x10 5 ID 50 units) or 263K (3x10 6 ID 50 units) injected SHaPrP mice (Tg81). Inoculated mice were monitored every other day and actions were taken to minimize animal suffering and distress as indicated in S1 Table. To euthanize mice CO 2 inhalation was used on the day of appearance of terminal clinical signs of scrapie. For confirmation of infectivity, the brains of inoculated mice were homogenized and heat inactivated for 25 min at 80˚C. Heat inactivated and non-heat inactivated brain homogenates were diluted in PBS/ 5% BSA to a final concentration of 1%wt/vol. and injected (30μl) into Tga20 mice (for RML inoculated animals) or Tg81 mice (for 263K treated transgenic mice).

Sample preparation
Terminally sick mice were sacrificed and organs were taken. The brain was sectioned sagittally, and one hemisphere was fixed in 4% formalin. A coronal section of brain containing hippocampus was placed in HANKS buffer and the rest was snap frozen in liquid nitrogen. Brain homogenates from snap frozen samples (10% wt/vol. in 0.32M sucrose diluted in PBS) were used for further analyses.

Conformational stability assay
Brain homogenates (10% w/v) from Tga20 mice inoculated with brain homogenates from RML infected B6 wt Prnp +/+ or Tg(MoPrP 169F );Prnp -/mice were dissolved in equal volumes of solubilization buffer (4% sarcosyl, 100 mM Tris HCl, pH7.4,) and incubated for 1h at 37˚C. Aliquots of 20 μl were treated with 20 μl of GdnHCl solutions at final concentrations ranging from 0 to 2 M and incubated for 1h at 37˚C. Final GdnHCl concentration was adjusted to 0.4 M in equal volumes. Samples were treated with PK at a final concentration of 25μg/ml for 30min at 37˚C. Protease activity was blocked by adding 50% volume of protease inhibitor cocktail (complete Mini, Roche) dissolved in 10ml deionized H 2 O. A fourfold excess of methanol for protein precipitation was added and incubated overnight at -40˚C. Samples were centrifuged for 1h at 20.000 g and 4˚C. The pellets were resuspended in 4x loading dye (NuPAGE LDS sample buffer, Thermo Fisher Scientific) and analyzed on a 4-12% Bis-Tris SDS polyacrylamide gel (NuPAGE, Invitrogen) followed by Western blotting as described above.

Protein density gradient analysis
For each tested sample, 400 μg of brain homogenate diluted in PBS was solubilized in an equal volume of solubilization buffer [50 mM HEPES (pH 7.4), 300 mM NaCl, 10 mM EDTA, 2 mM dithiothreitol (DTT), 4% (wt/vol) dodecyl-d-maltoside (Sigma)] and incubated for 45 min on ice. Sarkosyl (N-lauryl sarcosine; Fluka) was added to a final concentration of 2% (wt/vol), followed by additional 30 min incubation on ice. Four hundred microliters of the samples were loaded on a 3.6 ml continuous 7 to 28% density gradient (OptiPrep, Sigma), with a final concentration of 25 mM HEPES (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM DTT, and 0.5% sarkosyl. Samples were centrifuged at 52,000 rpm for 90 min at 4˚C (with a Discovery M150SE Micro-Ultracentrifuge and S52-ST swinging bucket). Fractions of 200 μl were collected and 20 μl per fraction was analyzed on 4-12% Bis-Tris SDS polyacrylamide gel (NuPAGE, Invitrogen) followed by Western blotting as described above. The signal was quantified with the freeware analysis program image studio lite (Li-cor).

Histological analyses
Formalin fixed brain hemispheres were decontaminated in formic acid for 60 min to eliminate prion infectivity. After additional fixation in formalin, the tissue was paraffin embedded. Paraffin sections (3 μm) were stained with H&E. Immunohistochemical stainings for GFAP (1:13000 DAKO) and IBA1 (1:1000) were performed using standard methods. For SAF84 immunostains, sections were incubated in 98% formic acid for an additional 6 min after deparaffinization and washed in distilled water for 30 min. Sections were treated with citrate buffer (pH 6.0) for 3 min at 100˚C. After adapting to room temperature, sections were incubated in Ventana buffer, and stained with the NEXEX immunohistochemistry robot (Ventana Instruments) using an iVIEW DAB Detection Kit (Ventana). After incubation with protease 2 (Ventana) for 16 min, sections were incubated with anti-PrP SAF-84 (SPI-Bio, A03208, 1:200) for 32 min and counterstained with hematoxylin.

LCP staining on frozen sections
Brain samples, frozen in HANKS buffered salt solution, were cut (10 μm) and dried for 1h at room temperature. Tissue was fixed on slides with 100% ethanol for 10 min and subsequently washed in deionized water. Slides were preincubated for 10 min in PBS followed by staining with LIN5050 diluted in PBS ([4 μM] final concentration) for 30 min at room temperature. Sections were embedded with fluorescent mounting medium (DAKO) and stored at 4˚C.

Nerve fibre preparations
Sciatic nerves were fixed in 2% glutaraldehyde supplemented with 0.1 M sodium phosphate buffer at pH 7.4 and processed following standard procedures. At least 8 axons per nerve were teased on a glycerine gelatine covered glass slide.
Structural image. The structural representation in Fig 1A was   In (A) 1% brain homogenate from RML inoculated mice was intracerebrally injected (30μl) into Tga20 mice (mouse "a" and "b" correspond to Tg(MoPrP 169F );Prnp -/-, mouse "c" to Tg(MoPrP 169F );Prnp +/and mouse "d" to wt Prnp +/+ (see Fig 2A). (B) 1% brain homogenate from Tg(MoPrP 169F );Prnp -/-(mouse e, f and g; see Fig 2B) and Prnp -/-(mouse h; see Fig 2B) inoculated with 263K were intracerebrally passaged into Tg81 mice. Heat inactivated noninfectious brain homogenate from Tg(MoPrP 169F );Prnp +/or Tg(MoPrP 169F );Prnp -/mice was used as control. PK Western blot analysis of brain homogenates from RML-infected Tg(MoPrP 169F );Prnp -/mice probed with a panel of different anti-PrP antibodies (POMs), covering different epitopes of PrP, compared to brain homogenates from RML-infected wt Prnp +/+ mice. All antibodies detected increased sensitivity to PK digestion and no shorter PK-fragments for MoPrP 169F prions. POM5 which recognizes the β2-α2 loop of PrP (residues 168-174) was not able to detect MoPrP 169F , because of the point mutation in its recognition site. Brain homogenates from RML-infected wt Prnp +/+ mice were used as controls. 10 μg of total protein was treated with 25 μg/ml PK for 30 min at 37˚C and loaded onto the gel. Non-digested samples from RML-infected wt and Tg(MoPrP 169F );Prnp -/were used as controls. Bands were detected with the different anti-PrP antibodies as indicated in the Figure  Passage of brain homogenate from 263K inoculated Tg((MoPrP 169F );Prnp -/mouse (g) (previously shown in Figs 2B and S3B) into hamster PrP expressing Tg81 mice led to death and accumulation of PK resistant PrP. In contrast Tg(MoPrP 169F );Prnp -/mice showed marked reduction of PK resistant material. In Tg (MoPrPF 169F );Prnp +/mice the amount of PK resistant material was also reduced compared to the passaged Tg81 mice, but showed some interindividual variability. 20 μg of total protein per lane was treated or not with 25 μg/ml PK for 30 min at 37˚C. Bands were detected with the anti PrP antibody POM1 (200 ng/ml). (TIF) S9 Fig. Fractionation-and PK resistance analysis of brain homogenates from prion infected mice. (A) Western blot analyses of total PrP from differentially fractionated brain homogenate samples of wt Prnp +/+ mice inoculated with noninfectious brain homogenate or RML and RML inoculated Tg(MoPrP 169F );Prnp -/and Tg(MoPrP 169F );Prnp +/mice to confirm data from Fig 5. A technical replicate of the data presented in Fig 5 is shown with all fractions (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) loaded to a SDS-PAGE from a 7-28% OptiPrep gradient. (B) Effect of proteinase K on PrP aggregates migrating in low and high density fractions. Western blot analysis of pooled low (fractions 1-4) and high (fractions 12-15) density fractions from the ultracentrifugation experiment show PK (2 μg/ml) sensitive PrP in the low-density fractions. In high density fractions, PK resistance is maintained in wt Prnp +/+ , whereas reduced resistance is observed in Tg (MoPrP 169F ) mice. Fractions were methanol precipitated and adjusted to 190 ng of total protein per lane. PrP was detected using the anti-PrP antibody POM1 (200 ng/ml). (TIF) (TIF) S1 File. Excel sheet containing data that led to Fig 1B, Fig 2, Fig 3F, Fig 4B and 4C, Fig 5B,  S2 Fig, S3 Fig and S4 Fig. (XLSX) S1 Table. Clinical assessment and scoring of mice inoculated with prions. The mice were observed every other day after prion inoculation for clinical signs including gait, grooming, activity, rough hair coat, limb paresis and ataxia. Once the mice showed the first sign of scrapie (grade 1), they were monitored every day and wet food was supplied in the cage. When the mice reached score grade 2 that hampered the mice reaching the water bottle, they were euthanized by CO 2 inhalation. (DOCX)