Synthetic Miniprion PrP106

Elucidation of structure and biological properties of the prion protein scrapie (PrP(Sc)) is fundamental to an understanding of the mechanism of conformational transition of cellular (PrP(C)) into disease-specific isoforms and the pathogenesis of prion diseases. Unfortunately, the insolubility and heterogeneity of PrP(Sc) have limited these studies. The observation that a construct of 106 amino acids (termed PrP106 or miniprion), derived from mouse PrP and containing two deletions (Delta 23-88, Delta 141-176), becomes protease-resistant when expressed in scrapie-infected neuroblastoma cells and sustains prion replication when expressed in PrP(0/0) mice prompted us to generate a corresponding synthetic peptide (sPrP106) to be used for biochemical and cell culture studies. sPrP106 was obtained successfully with a straightforward procedure, which combines classical stepwise solid phase synthesis with a purification strategy based on transient labeling with a lipophilic chromatographic probe. sPrP106 readily adopted a beta-sheet structure, aggregated into branched filamentous structures without ultrastructural and tinctorial properties of amyloid, exhibited a proteinase K-resistant domain spanning residues 134-217, was highly toxic to primary neuronal cultures, and induced a remarkable increase in membrane microviscosity. These features are central properties of PrP(Sc) and make sPrP106 an excellent tool for investigating the molecular basis of the conformational conversion of PrP(C) into PrP(Sc) and prion disease pathogenesis.


INTRODUCTION
Prion-related encephalopaties are characterized by accumulation of pathogenic forms of the prion protein (PrP) 1 , termed PrP scrapie (PrP Sc ), in the central nervous system. Unlike the normal cellular isoform (PrP C ), PrP Sc has a high content of β-sheet secondary structure, is insoluble in non-denaturing detergents and partially resistant to protease digestion, and has the propensity to form amyloid fibrils. Elucidation of the structure of both PrP isoforms is fundamental to understand the molecular mechanism of the conformational transition from PrP C to PrP Sc . Unfortunately structural studies of the pathogenic form have been hampered by the insolubility, heterogeneity and complexity of PrP Sc preparations. Synthetic and recombinant fragments of PrP have been more useful for such investigations (1). A construct of 106 amino acids (designated as PrP106 or miniprion), derived from mouse PrP and containing two deletions (∆ 23-88, ∆ 141-176), was expressed in scrapie-infected neuroblastoma cells generating a protease resistant polypeptide, displaying better solubility 7 nm-absorbing peak was collected and characterized by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS). The pooled fractions containing the homogeneous labelled miniprion were lyophilzed. The probe was removed by a 2 h-treatment with 10% triethylamine (TEA) in acetonitrile/water (1:1). After lyophilization, the peptide was dissolved in concentrated TFA and precipitated with diethylether. The pellet was redissolved in acetonitrile/water (1:1) and relyophilized. Tipically this procedure enables to obtain about 50 mg of pure sPrP106, representing an overall yield of ∼5%.

Circular dichroism spectroscopy
Circular dichroism (CD) spectra were recorded with a Jasco J-710 CD spectrometer (Jasco, Easton, MD) scanning spectra at a speed of 20 nm/min, with a bandwidth of 2 nm and a step resolution of 0.2-1 nm using quartz cells with an optical path of 0.1 cm. Aliquots of lyophilized sPrP106 were dissolved in (i) deionized water, (ii) 1 mM sodium acetate, pH 5.5, (iii) saline, (iv) 1 mM sodium acetate containing liposomes or 20% TFE, (v) deionized water containing 20%, 40% or 60% TFE, at a concentration of 0.05-1 mg/ml as specified in Table 1.
Protein solutions were incubated at room temperature for different times (from 1 h to one week) before measurement. No substantial variations were observed at different incubation times. Background spectra were subtracted, and the data were converted to molar ellipticity.
Each spectrum shown is the result of the accumulation of 4-5 individual spectra. The secondary structure content was calculated using a program DICROPROT V2.5, which contains a least square method based on the Gauss-Jordan elimination (10), the self consistent method and the variable selection method (11
For the experiments liposomes were resuspended in 1 mM sodium acetate, pH 5.5. CD experiments were performed after addition of liposome suspensions to sPrP106 solution in the same buffer, to obtain a final peptide concentration of 0.1 mg/ml and a protein:lipid molar ratio of 1:25.

Proteinase-K digestion
sPrP106 was digested with PK in the presence or absence of detergents. In a first set of experiments, the protein (0.2 mg/ml) was incubated in 10 mM Tris·HCl, pH 7.4/100 mM NaCl/0.5% NP-40/0.5% sodium deoxycholate (DOC) at 37°C up to 1h. The PK:protein ratios employed were 1:320, 1:160, 1:32, 1:16, 1:3.2. The reaction was stopped by adding phenylmethanesulfonyl fluoride to a final concentration of 5 mM. Samples were analyzed by SDS-PAGE and visualized by Coomassie staining or by Western blotting using monoclonal antibody 3F4 and polyclonal antibody R340. Alternatively, the protein (1 mg/ml) was incubated at 37°C in 1 mM sodium acetate, pH 5.5, with an enzyme to protein ratio of 1:50.
During PK-digestion, aliquots were taken every 30 min up to 5 h. After addition of an equal volume of 1% TFA, the samples were analysed by MALDI MS.

MALDI mass spectrometry
Fragments or full-length sPrP106 were analyzed by a Bruker Biflex TM or a Reflex III TM MALDI mass spectrometer. Few µl of the sample were mixed with an equal volume of a saturated solution of sinapinic acid (Sigma-Aldrich) in acetonitrile/0,1% TFA 1:3 (v:v) and 1 µl of the mixture was deposited on the MALDI target.

Electron microscopy
Aliquots of lyophilized PrP106 were suspended in aqueous solutions with different pH and ionic strength at a final concentration of 0.1 and 0.3 mM. The solvents included (i) deionized water, pH 5.0, (ii) 10 mM Tris·HCl, pH 5.5, (iii) 1 mM sodium acetate, pH 5.5, and (iv) 1 mM sodium acetate, 120 mM NaCl, pH 5.5. In some experiments, the peptide was subjected to reduction and partial denaturation in 10 mM Tris·HCl, 100 mM DTT, 2 M guanidinium hydrochloride (Gdn·HCl), pH 5.5, for 24 hours, followed by refolding by dialysis against deionized water before analysis. After incubation at 37°C for 1h, 48 h and 7 days, 10 µl of each sample were air-dried on gelatine-coated slides, stained with Congo red

Neurotoxicity
Primary cultures of rat cortical neurons were prepared as previously described (13).
Briefly, brains were removed from fetal rats on embryonic day 17. Cortical cells were dissociated and plated at the density of 5x10 5 cells per well. The cells were cultured in basal medium Eagle (BME-Hanks' salt, GIBCO) supplemented with 10% fetal calf serum (FCS, GIBCO) and 2 mM glutamine. sPrP106 was dissolved in deionized water and added to the medium, for the acute treatment (final concentration 5 and 10 µM), on day 5 of culture and the cell viability was determined 24 hours later. For the chronic treatment SPrP 106 was added to the medium on days 1, 3, 5 of culture, reaching final concentrations of 0.5, 1, 5 and 10 µM in the well, the cell viability was determined at day 7. Cell viability was assessed quantitatively by the MTT method (14) with an automated micro plate reader (Perkin-Elmer lambda reader).

Synthesis, purification and characterization of PrP106
PrP106 was synthesized by Fmoc-based SSPS with some modifications. It was necessary to introduce several double coupling cycles to enhance the coupling efficiency in correspondence of the difficult sequences identified by conductivity measurements. In addition, three deprotection cycles where added after coupling of each amino acid. This step was found to greatly improve the yield of the synthesis.
A special strategy was adopted for the purification of sPrP106 from its truncated forms, which were co-eluting or eluting very closely to the full-length protein (Fig. 1A). The target protein was labelled with a specific and transient labelling, a lipophilic Fmoc-based chromatographic probe. The probe reacted efficiently with the free amino terminus of the protein as checked by ninhydrin reaction. It should be recalled that the N-terminus of sPrP106 is the only available site for the reaction, since the N-termini of truncated peptides are acetylated in the capping step after each coupling cycle. On RP-HPLC the derivatized synthetic PrP106 (dsPrP106) eluted considerably later than the closely related impurities (Fig.   1B) and was easily monitored at 300 nm (Fig. 1C). The purified dsPrP106 was characterized by MALDI MS (Fig. 2A). The probe was removed by basic hydrolysis with TEA and discharged simply by extraction with diethylether, thus avoiding an additional chromatographic step. Finally the pure protein was characterized by MALDI MS (Fig. 2B).
The observed molecular weights of dsPrP106 and sPrP106 were 11998.6 and 11572.6 Da, respectively, which are in good agreement with the calculated molecular weights (12001.7 and 11570.1 Da).

sPrP106 adopts a stable -sheet-rich secondary structure
The secondary structure of sPrP106 in different environments was investigated by CD spectroscopy in the far UV. Measurements were performed at protein concentrations of 0.05, 0.1, 0.5 and 1 mg/ml. The protein adopted a stable conformation with high percentage of βsheet (around 40%, calculated with different deconvolution programs), independently from its concentration (Fig. 3A), ionic strength (Fig. 3B) and time of incubation (Table 1). All spectra showed the characteristic β-sheet single minimum between 210 and 220 nm. The secondary structure was not substantially affected by the addition of 20% TFE or liposomes to peptide solutions in 1 mM sodium acetate, pH 5.5 (Fig. 3B). A similar result was observed when 20% TFE was added to sPrP106 dissolved in deionized water at a concentration of 1 mg/ml (Table   1). Conversely, at low protein concentration (0.1 mg/ml), the addition of 20%, 40% or 60% TFE to deionized water resulted in progressive increase in α-helical structure (Fig. 3C, Table   1); with high percentage of TFE (40%-60%), the 208-nm band clearly showed a larger intensity than the 222-nm band, a typical feature of a protein where α helices and β sheets are located in separated domains and do not intermix along the polypeptide chain (20).
To verify whether the structural characteristics and stability of sPrP106 were the result of inter-molecular disulfide bridge formation during the purification process, the protein was denaturated with 20 mM sodium acetate pH 5.5/8 M Gdn·HCl/100 mM DTT and refolded following dialysis against 20 mM sodium acetate, pH 5.5. The CD spectrum of refolded sPrP106 was marked by a high β-sheet content, although a broader minimum band was observed in the 210-220 nm region ( fig. 3D).
In addition to CD determinations, the Double Prediction Method (18) was used for secondary structure prediction. The algorithm indicated that the region of the protein comprising residues 138-217 has propensity to acquire β-sheet structure (Fig. 4).

sPrP106 contains a PK resistant C-terminal domain
Experiments of PK-resistance were performed in different conditions. In a first set of experiments sPrP106 was digested in the presence of detergents (NP-40 and DOC) at increasing PK:protein ratio from 1:320 to 1:3.2. The digestion products were analyzed by SDS-PAGE and visualized by Coomassie staining or Western blotting using two different antibodies. Figure 5A shows that an approximately 5-6 KDa-resistant core, constituted probably by more than one component, withstood digestion even at a PK:protein ratio of 1:3.2 after 60 min incubation. The PK-resistant fragments were not detected by Western blotting using monoclonal antibody 3F4 (Fig. 5B), whose epitope is located on the N-terminal segment of the protein. Conversely, the antibody R340, which recognizes epitopes on the Cterminal portion (beyond amino acid 134) of mouse PrP, immunoreacted with the PKresistant C-terminal fragments of sPrP106 (Fig. 5C).
The PK-resistant fragments were characterized by MALDI MS. To be able to directly analyze the samples on MALDI MS, the digestion was carried out in 1 mM sodium acetate, pH 5.5, in the absence of detergents. The information obtained from Western blot analysis, the purity and the relatively small size of sPrP106, and the accuracy of the mass measurements allowed us to unambiguously identify the sequence of PK-resistant fragments.
Already after 30 min digestion, sPrP106 lost its N-terminal portion generating peptides starting at amino acid residue Ser134 and having a ragged C terminus corresponding to residues 224, 220, 218, 217, and 215 (Table 2) in agreement with published results obtained with rPrP106 (4). The most abundant fragment was 134-224 after 30 min incubation (data not shown) and 134-217 after 2 up to 5 h of PK digestion (Fig. 6).

sPrP106 assembles in branched filamentous structures
The nature of aggregates generated by sPrP106 was determined by electron microscopy (EM) after negative and positive staining and by polarized light microscopy after incubation with Congo red. The peptide was soluble in deionized water, 10 mM Tris·HCl, pH Following reduction, denaturation and refolding, peptide assemblies were more structured and better defined (Fig. 7C). They consisted of slightly curved, often branched filaments with a diameter ranging from 8 to 22 nm (19.7 ± 3.7) and a length ranging from 0.8 to 4 µm. The filaments had a rosary-like profile without a clear twisting or crossover. Spherical and ringshaped structures, likely representing the cross section of single filaments, were also observed. Conversely, finely granular, amorphous material was almost absent. Following incubation in sodium acetate containing 120 mM NaCl, the peptide aggregates showed intermediate features, in that they consisted of very abundant, thin, short, often branched filaments with a diameter ranging from 5 to 12 nm (9.4 ± 2.6) and a length of 0.5 to 2 µm, mixed with small amount of finely granular, osmiophilic material (Fig. 7B).

sPrP106 is neurotoxic to primary rat cortical neurons
To evaluate the effects of sPrP106 on cell viability, primary rat cortical neurons were exposed for 24 hours (5 and 10 µM) or for 7 days (0.5-10 µM) to the synthetic protein. The 24-hours exposure did not affect the neuronal viability (data not shown), whereas a statistically significant neurotoxic effect was detected after the chronic treatment at the higher protein concentrations (Fig. 8). The cell viability was reduced of about 40% and 60%, compared to the controls, with the exposure to sPrP106 at 5 and 10 µM respectively (Fig. 8).

sPrP106 increases membrane microviscosity
The ability of sPrP106 to affect membrane microviscosity was analyzed on suspensions of primary rat cortical neurons following incubation with DPH fluorescent probe. Table 3 reports the FP value before and after addition of increasing concentrations of sPrP106. A striking, dose-dependent rigidifying effect on nerve cell membranes was detected at 5, 10 and 20 µM protein concentration, with 56%, 64% and 105% increase in basal FP value, respectively. This effect was higher than the one observed with the syntetic peptide concentration. sPrP106 was also tested on artificial membranes of different composition and charge, using 10 µM protein concentration. The rigidifying effect was consistent with neutral liposomes prepared with PC:Chol (1:1) and PC:SP (1:1) and negatively charged PG:Chol (1:1) liposomes, with an increase in FP of 31%, 45% and 24% of basal value, respectively.
The membrane protein topology prediction method (19), TMHMM, gave a high score for the insertion of sPrP106 in the membrane through a transmembrane helix spanning residues 111-133, with the N-terminal extracellular domain corresponding to residues 89-110 and the C-terminal cytoplasmatic domain comprising residues 134-230 (Fig. 4).

DISCUSSION
Synthetic peptides derived from discrete PrP regions have been extensively used to identify protein domains that could be involved in the conformational transition of PrP C to PrP Sc and in disease pathogenesis (1). In particular, the peptide PrP106-126 was found to possess many of the PrP Sc properties, including the propensity to adopt a β-sheet conformation and form amyloid fibrils, the partial resistance to PK digestion, and the ability to induce nerve cell degeneration and glial cell activation in vitro (1,13,21).
A major concern in using small PrP peptides is that they can adopt conformations, which may differ from their actual structure in the entire protein. Moreover, although such peptides are very useful for cell culture studies, they are not capable to induce disease when administered to animals. Supattapone et al. (3) showed that the minimal sequence that sustains prion replication corresponds to a redacted version of PrP lacking residues 23-88 and 141-176. This peptide termed PrP106 or miniprion is currently regarded as an excellent model to unravel both the conversion of PrP C to PrP Sc and prion pathogenesis.
Our decision to generate synthetic rather than recombinant PrP106 was based on the assumption that accurate mechanistic studies can be carried out only with a pure protein devoid of any cellular contaminant. This idea is substantiated by the knowledge that several proteins with high affinity for PrP C or PrP Sc may play an important role in the conversion reaction and disease propagation (5). Ball et al. have recently reported the synthesis of PrP polypeptides by highly optimized Fmoc/t-Boc protocols and chemical ligation (7). Our procedure to generate relatively large amount of sPrP106 was straightforward and simple. Its strength resides on a special purification strategy based on the use of a lipophilic chromatographic probe (8), which enables us to purify the full-length protein from the side products by a single chromatographic step. This fact minimizes the losses and compensates for the low yield of the chemical synthesis of a long polypeptide by classical Fmoc-based SSPS. Following this strategy, the overall yield of PrP106 was ∼5%, which is satisfying for an easy-to-aggregate PrP polypeptide. sPrP106 showed interesting biochemical features. First, the protein had a high propensity to adopt a secondary structure with high β-sheet content. The amount of β-sheet was similar in different conditions, at variance with PrP106-126 that shows an environmentaldependent conformational polymorphism. Furthermore, the behaviour of the synthetic protein was slightly different from that of its recombinant counterpart, rPrP106 (4). In fact, sPrP106 exhibited a well-ordered structure already at low concentration (0.1 mg/ml, i.e. 8.6 µM) in 1 mM sodium acetate pH 5.5, while rPrP106 was found to undergo a concentration-dependent conformational transition from an unfolded monomeric state to an ordered β-sheet multimeric assembly in the range of 30-100 µM (4). Second, sPrP106 contains a C-terminal domain, which is notably insensitive to PK degradation. Even at stringent conditions of PK-to-protein ratio and prolonged digestion times, a protease-resistant core mainly composed of a fragment spanning residues 134-217 was consistently detected. The results from CD determination and PK-digestion suggest that the tertiary structure of sPrP106 may consist of a flexible Nterminal domain and a C-terminal β-sheet domain starting approximately at Ser134. This hypothetical structure is supported by the algorithm of the Double Prediction Method (18).
The macromolecular assemblies generated by sPrP106 in different conditions were analyzed by EM. Protein suspensions in aqueous solutions contained a few spherical particles with a diameter of 20-30 nm, short filamentous structures of irregular thickness ranging from 4 to 10 nm, and amorphous granular material. The heterogeneity of sPrP106 aggregates was not longer observed when the protein was subjected to reversible denaturation followed by refolding. Under these conditions, the preparations essentially contained branched filaments with a relatively regular, rosary-like profile, and a diameter ranging from 8 to 22 nm. The filamentous aggregates were not birefringent under polarized light after Congo red staining. It remains possible that further manipulations, such as reducing conditions, low pH, partially denaturating environments (0.5-1 M Gdn·HCl) could favour the formation of amyloid-like fibrils (22,23). Nonetheless, the formation of ordered but not amyloid-like structures might be compatible with the peculiar parallel β-helix fold, which is the feature of PrP Sc 106 crystals as recently shown by electron crystallography studies (24).
Evidence suggests that the neuropathological changes observed in prion diseases are due, at least in part, to accumulation of PrP Sc . This view is supported by the observation that the protease resistant core of PrP Sc has a variety of pathogenic effects in vitro, including neurotoxicity and ability to interact with plasma membrane yielding increased microviscosity.
Similar effects have been observed with short synthetic peptides, in particular with PrP106-126. Our study shows that sPrP106 not only maintains these properties but it is more effective than PrP106-126. In fact, the concentrations required to obtain the same extent of nerve cell death and a similar increase in membrane microviscosity were five times lower with sPrP106 than with PrP106-126 (13,16). At variance with PrP106-126, the interaction of sPrP106 with multilamellar vescicles was influenced by their net charge and composition (12). This is likely related to a distinctive interaction of sPrP106 with membranes rather than to the net charge of the polypeptides, since sPrP106 and PrP106-126 are both positive. sPrP106 contains the membrane-spanning domain (roughly residues 113-135) of PrP (25,26). Interestingly, the application of a membrane protein topology prediction method (19), TMHMM, indicated that sPrP106 possesses the characteristics to insert into the membrane through a transmembrane helix spanning residues 111-133.
In conclusion, our study shows that sPrP106 (i) readily adopts a stable β-sheet conformation, (ii) contains a protease-resistant core, (iii) assembles in relatively regular filamentous structures, (iv) is highly neurotoxic and (v) interacts with membranes inducing a remarkable increase in microviscosity. These properties make sPrP106 a promising model to investigate the PrP C >PrP Sc conversion process and disease pathogenesis.    for the monoclonal antibody 3F4. The PK-resistant fragment is defined by bold and underlined letters. In correspondence of PrP106 sequence the secondary structure prediction is shown, where "t" stands for β-turn, "c" for random coil, "h" for α-helix and "e" for extended strand. For the secondary structure prediction the Double Prediction Method was used (18).
Results from the application of a membrane protein topology prediction method (19), TMHMM, is also reported. The segment designated with "m" is a transmembrane helix, while "o" and "i" segments are extracellular and cytoplasmatic domains, respectively.