N-terminal Truncation of Prion Protein Affects Both Formation and Conformation of Abnormal Protease-resistant Prion Protein Generatedin Vitro *

Transmissible spongiform encephalopathy diseases are characterized by conversion of the normal protease-sensitive host prion protein, PrP-sen, to an abnormal protease-resistant form, PrP-res. In the current study, deletions were introduced into the flexible tail of PrP-sen (23) to determine if this region was required for formation of PrP-res in a cell-free assay. PrP-res formation was significantly reduced by deletion of residues 34–94 relative to full-length hamster PrP. Deletion of another nineteen amino acids to residue 113 further reduced the amount of PrP-res formed. Furthermore, the presence of additional proteinase K cleavage sites indicated that deletion to residue 113 generated a protease-resistant product with an altered conformation. Conversion of PrP deletion mutants was also affected by post-translational modifications to PrP-sen. Conversion of unglycosylated PrP-sen appeared to alter both the amount and the conformation of protease-resistant PrP-res produced from N-terminally truncated PrP-sen. The N-terminal region also affected the ability of hamster PrP to block mouse PrP-res formation in scrapie-infected mouse neuroblastoma cells. Thus, regions within the flexible N-terminal tail of PrP influenced interactions required for both generating and disrupting PrP-res formation.

Transmissible spongiform encephalopathy (TSE) 1 diseases are a family of fatal neurodegenerative disorders, which affect both humans and animals. These diseases are characterized by the accumulation of an abnormal form of prion protein (PrP), which is associated with the pathogenic process and differs from normal PrP in its secondary structure and chemical properties. The susceptibility of PrP to digestion with proteinase K (PK) is generally used to distinguish the normally proteasesensitive PrP (PrP-sen) from the abnormal protease-resistant form (PrP-res) (1). The C-terminal globular domain of PrP-sen, residues 125-231, is composed of three ␣-helices and two ␤-strands connected by loops and turns (2,3) and has two potential N-linked glycosylation sites at residues 181 and 197 (4 -6). In contrast, the N-terminal portion of PrP-sen, residues 23-124, has the properties of a flexible random coil polypeptide (2,3,7) although some bends and turns could be associated with residues 90 -119 (8). Following conversion of PrP-sen to PrP-res, the ␤-sheet content of the molecule increases (9 -11), and residues 90 -231 become protease-resistant, whereas residues 23-89 remain susceptible to protease digestion (12,13).
It is apparent from several in vitro studies that the amyloidogenic region of PrP from residues 113-120 is important for the generation of PrP-res. First, PrP-sen with this region deleted is not converted to a protease-resistant form when expressed in scrapie-infected neuroblastoma cells (14). Furthermore, peptides composed of amino acid residues within this region inhibit conversion of PrP-sen to PrP-res in both cell-free conversion reactions (15,16) and scrapie-infected neuroblastoma cells (16).
The in vitro relevance of the N-terminal region of PrP-sen is supported in vivo by the altered disease susceptibility of PrP knockout mice expressing a transgene encoding various truncated PrP-sen molecules. In these studies, mice expressing PrP-sen with a deletion of residues 32-93 were susceptible to mouse scrapie, albeit with reduced levels of detectable PrP-res, altered clinical signs, and pathology (17). In contrast, mice expressing further truncated PrP-sen (residues 32-106) were not susceptible to infection (18). Thus, it appears that residues upstream of amino acid 93 influence PrP-res formation and scrapie pathogenesis in mice, whereas residues immediately downstream (94 -106) appear to be important for susceptibility to scrapie infection.
In the present study, a cell-free conversion assay was used to study the biochemical influence of the flexible N-terminal tail of PrP on the formation of protease-resistant PrP. This assay generates de novo PrP-res (19) and has been shown to mimic the species and strain-specific interactions that occur between PrP-sen and PrP-res to form protease-resistant PrP (20 -25). The conversion of a series of PrP-sen molecules with progressive deletions within the N-terminal tail was studied in the presence or absence of post-translational modifications such as glycosylation and the addition of the C-terminal glycophosphatidylinositol (GPI) anchor. Our results show that deletions of residues 34 -94 and 95-113 have distinct consequences for conversion of PrP-sen to a protease-resistant form and suggest an underlying role for the flexible N-terminal domain of PrPsen in scrapie pathogenesis. ORF (26) was digested with KpnI and NsiI, and a synthetic oligonucleotide polylinker encoding restriction enzyme sites KpnI, XhoI, EcoRI, NheI, AgeI, MluI, NsiI was inserted. The resulting plasmid was digested with AgeI and EcoRI, and a 0.2-kb EcoRI-AgeI fragment from the 5Ј end of mouse PrP (27) was inserted to create plasmid p23-7. The amino acid sequence of this portion of mouse PrP is identical to HaPrP. Deletions were then produced by annealing four to six overlapping oligonucleotide sequences of the upper and lower strand of HaPrP between the AgeI (amino acid residue 34) and NsiI (amino acid residue 138) sites to reconstitute clones deleted between residues 34 -94 (Ha⌬94), 34 -113 (Ha⌬113), 34 -120 (Ha⌬120), and 34 -124 (Ha⌬124) (sequence available on request). These annealed oligonucleotides were then inserted into p23-7. PrP deletion mutants were sequenced, subcloned into the pSFF retroviral expression plasmid, and transfected into the retroviral vector-packaging cell lines (⌿2 and PA317) as described previously (28,29). To create GPI-anchored deletion mutants, clones in Bluescript KS (ϩ) were cloned into pSFF at BamHI and XhoI sites. To create GPI-negative deletion mutants, the StuI to XhoI region of each clone in Bluescript was replaced with the StuI/XhoI fragment of a mouse PrP clone that lacked the GPI signal sequence in the 3Јuntranslated region and that shared sequence homology with hamster PrP. The resulting clones were subcloned into pSFF as described above. GPI-positive and GPI-negative wild type PrP (HaWT) clones have been described previously (19,29). The pSFF vector was used as a negative control for the immunoprecipitated protein used in the cell-free assays and for transduction of scrapie-infected mouse neuroblastoma (ScϩN2a) cells.
Cell-free Conversion Assay-Cell-free conversion of PrP-sen to PrPres was described previously (19). Briefly, 200 ng of PrP-res, purified from the brains of hamsters infected with the 263K strain as described previously (10), was pretreated for 1 h at 37°C in 2.5 mM guanidine hydrochloride and then incubated with 25,000 cpm of 35 S-labeled PrPsen in conversion buffer (0.75 M guanidine hydrochloride, 50 mM sodium citrate buffer, pH 6.0, 5 mM cetylpyridinium chloride, 1.2% sarkosyl) at 37°C for 48 h. After incubation, one-tenth of the reaction was precipitated in methanol, and the remaining nine-tenths was treated with 80 l of PK (12 g/ml) for 1 h at 37°C. These samples represented PK-and PKϩ samples, respectively. PK digestion was stopped by the addition of thyroglobulin and Pefabloc (Roche Molecular Biochemicals) and precipitated with 4 volumes of cold methanol. The resulting precipitate was resuspended in sample buffer and run on a 16% SDS-PAGE gel. The PKϪ and PKϩ samples were quantified by phosphor autoradiographic image analysis of the dried gel (Molecular Dynamics Storm Phospho-rImager system). PrP-res was considered to be all PK-resistant products not found in a control lane lacking PrP-res. The amount of PKresistant PrP generated in the cell-free assay was calculated as a percentage of the input PrP-sen.
Binding Analysis-The amount of 35 S-labeled PrP-sen bound to PrPres was determined under conditions of the cell-free assay. After incubation for 48 h, the conversion reaction was centrifuged at 14,000 ϫ g for 10 min at room temperature. The supernatant was removed, methanol-precipitated in the presence of a carrier protein (thyroglobulin), and saved as the unbound fraction. The pellet was washed in wash buffer (50 mM sodium citrate buffer, pH 6.0, 5 mM cetylpyridinium chloride, 1.2% sarkosyl) and centrifuged again. The wash was saved and methanol precipitated. The final pellet was resuspended in wash buffer and methanol-precipitated. The bound (pellet) and unbound (supernatant and wash) fractions were analyzed on a 16% SDS-PAGE gel, and the amount of 35 S-labeled PrP-sen recovered in each fraction was quantified using the Storm PhosphorImager system. To calculate the percentage of total PrP-sen bound, the amount of bound 35 S-labeled PrP-sen (all glycoforms) was divided by the total (bound and unbound) 35 S-labeled PrP-sen recovered. The amount of unglycosylated 35 S-labeled PrP-sen bound to PrP-res was also calculated. All values were normalized for background aggregation by subtracting the percentage of total or unlgycosylated 35 S-labeled PrP-sen found in the pellet in the absence of PrP-res.
Transduction and Analysis of Scrapie-infected Mouse Neurolastoma Cells Expressing Deletion Mutants of Hamster PrP-Expression of hamster PrP molecules in scrapie-infected mouse neuroblastoma (ScϩN2a) cells by transduction with the retroviral expression vector pSFF and detection of PrP-res expression in these cells has been described previously (29,(32)(33)(34). Transduction efficiency was measured by immunofluorescence using the 18.8 monoclonal antibody (35), which reacts with the retroviral gag gene expressed by the pSFF vector used in this study. Mouse PrP-res and foreign PrP-res derived from HaWT and Ha⌬94 were detected using the rabbit anti-peptide antibody R30 (31). Expression of foreign HaWT and Ha⌬94, but not mouse PrP, was detected with the hamster-specific monoclonal antibody 3F4. Foreign PrP-res derived from Ha⌬113, Ha⌬120, and Ha⌬124 was detected using the rabbit anti-peptide antibody R20 and was distinguished from endogenous mouse PrP-res by the size shift associated with the deletion mutants. The Western blot was developed with the enhanced chemiluminescense reagent system as described by the manufacturer (Amersham Pharmacia Biotech).
Statistical Analysis-Statistical comparisons between groups were performed using one-way analysis of variance and the Newman-Keuls Multiple Comparison test using the PRISM TM software package.

Deletion of Amino Acid Residues within the Flexible N-terminal Tail of PrP Reduces Cell-free Conversion of Hamster PrPsen-
The role of the flexible N-terminal tail of PrP-sen in conversion of protease-sensitive PrP to the protease-resistant PrP was determined using a biochemical cell-free conversion assay. The conversion of hamster PrP with progressive deletions within the N-terminal tail was compared with conversion of HaWT ( Fig. 1). All PrP-sen deletion mutants were converted to a protease-resistant form under the conditions used (Fig. 1b). However, conversion of PrP lacking residues 34 -94, Ha⌬94, was reduced by ϳ35% relative to HaWT (Fig. 1c), and deletion of an additional nineteen residues (residues 94 -113), Ha⌬113, reduced conversion by a further 25%. Conversion was not significantly affected by further deletions within the N-terminal tail to residue 120 (Ha⌬120) or 124 (Ha⌬124). Thus, it appears that residues 34 -94 and 34 -113 influence conversion. This is in agreement with the proposed importance of residues in the vicinity of residue 94 in the conversion of PrP-sen to PrP-res (18).
Following PK digestion of the conversion products, multiple protease-resistant PrP molecules were detected between 12 and 33 kDa. The distribution of these products was affected by deletion of amino acids within the N-terminal tail (Fig. 1b). The conversion products of HaWT and Ha⌬94 were similar in molecular mass, giving two groups of bands at 19 -22 and 25-33 kDa. This was consistent with PK digestion to residue 89 of full-length PrP and little to no digestion of Ha⌬94, which has previously been shown for N-terminally truncated PrP (17). In contrast, smaller products of 12 and 14 kDa as well as a 19 -22-kDa form were associated with conversions of Ha⌬113, Ha⌬120, and Ha⌬124. The 12-and 14-kDa conversion products appeared to result from PK cleavage at one or two new sites within the PrP polypeptide. Since the 12-and 14-kDa products were only a minor component of conversions of HaWT and Ha⌬94, the new proteolytic sites appear to be more accessible in PrP-res derived from PrP-sen with at least residues 34 -113 deleted from the N-terminal tail.
Conversion of HaWT and Ha⌬94 in the cell-free assay resulted in a prominent protease-resistant product between 25 and 33 kDa. The prominence of this conversion product is associated with conversion of a 60-kDa isoform of PrP that has been described previously to represent a dimerized form of PrP (19,36). We were concerned that the reduced conversion of Ha⌬113, Ha⌬120, and Ha⌬124 relative to HaWT and Ha⌬94 may have been due to the reduced ability of these truncated molecules to efficiently convert the dimer isoform. However, conversion of HaWT, Ha⌬94, and Ha⌬113 without the dimer isoform resulted in a quantitative pattern of conversion consistent with that reported for conversion in the presence of the dimer (data not shown).
Glycosylation Affects the Generation of Protease-resistant PrP from PrP-sen with Deletions within the N-terminal Tail-Glycosylation may affect the solubility and structure of PrP-sen and could therefore influence its conversion to PrP-res. The effect of glycosylation on cell-free conversion of N-terminal deletion mutants was investigated with unglycosylated PrP derived from tunicamycin-treated cells (Fig. 2). In contrast with the effect of deletions within the N-terminal tail of glycosylated PrP-sen (Fig. 1c), the amount of protease-resistant PrP derived from unglycosylated PrP-sen was unchanged by deletions within the N-terminal tail (Fig. 2c). However, the distribution of the conversion product was very different (Fig. 2b). The main protease-resistant product of HaWT conversion had a molecular weight of ϳ21 kDa, which was again consistent with digestion of 89 amino acids from the N terminus of the 30-kDa unglycosylated protein. Protease-resistant Ha⌬94 had a similar molecular weight to protease-resistant HaWT. The lower molecular weight conversion products previously described for PrP with deletions Ha⌬113, Ha⌬120, and Ha⌬124 were also present in the protease-resistant products of HaWT and Ha⌬94. This suggested that the putative internal PK digestion sites may be more accessible in the unglycosylated than in the glycosylated PrP conversion product. Ha⌬113 and Ha⌬120 were converted to a protease-resistant form that appeared to have the same molecular mass as the input PrP-sen, ϳ16 kDa, whereas the majority of this fully protease-resistant PrP was absent from conversions using Ha⌬124. The lower molecular mass bands at 12 and 14 kDa were also detected in conversions of Ha⌬113, Ha⌬120, and Ha⌬124.
The low molecular mass bands at 12 and 14 kDa may represent an alternative conformation of PrP-res that is more sus-ceptible to PK digestion. To assess the influence of the Nterminal tail on the formation of the fully protease-resistant form, we quantified the amount of PrP-res with a molecular mass greater than 15 kDa (Fig. 2d). By this method of analysis, PrP-res formation was decreased following deletion of residues 34 -94 and 34 -113. This result was consistent with the effect of deletion of these residues from glycosylated PrP-sen (Fig. 1c). Cell-free Conversion of N-terminal Deletion Mutants of PrPsen Lacking the GPI Anchor-PrP-sen is located on the external surface of the cell membrane, where it is anchored by a GPI moiety added to the C terminus of the protein at the time of translation. The GPI anchor is made hydrophobic by the presence of fatty acid chains, which may influence PrP solubility and conversion in an aqueous solution. To investigate the effect of deletions within the flexible N-terminal region on conversion of PrP in the absence of this modification, constructs were made that lacked the signal for the GPI anchor. GPI-negative PrP-sen was not heavily glycosylated and was found predominantly in a nonglycosylated form (19) (Fig. 3a). However, unlike conversion of unglycosylated GPI-anchored PrP (Fig. 2c), in the absence of the GPI anchor, deletion of residues 34 -94 (Ha⌬94) resulted in a 36% decrease in conversion relative to HaWT (Fig. 3c). Progressive deletion to residues 113, 120, and 124 had a slight, but not significant, cumulative effect on conversion with the mean percentage conversion reduced relative to HaWT by 46, 49, and 63%, respectively. The pattern of protease-resistant PrP bands ranged in size from 7 to 21 kDa and was similar to that observed for unglycosylated PrP (Figs.  2b and 3b). Therefore, in addition to the effect of deglycosylation, removal of the GPI anchor also influenced conversion of N-terminal deletion mutants of PrP.
The N-terminal Region of Glycosylated PrP-sen Modifies the Binding of PrP to PrP-res-Conversion of PrP-sen to PrP-res is a two-stage process, which begins with binding of PrP-sen to PrP-res, followed by conversion of the bound PrP-sen to a protease-resistant form (21,37,38). To investigate whether deletions within the N-terminal tail affected the solubility of PrP-sen and its ability to bind to PrP-res, [ 35 S]PrP-sen was incubated in the absence or presence of PrP-res under the cell-free assay conditions. The reaction was centrifuged, and the amount of PrP-sen found in the pellet fraction was quantified (Fig. 4). In the absence of PrP-res (Fig. 4a, solid bars), more than 20% of [ 35 S]HaWT was detected in the pellet fraction. Deletion of residues 34 -94 reduced this to ϳ10%, whereas less than 5% of PrP with residues 34 -113 or greater deleted was pelleted in the absence of PrP-res. Thus, residues within the N-terminal region increased the formation of pelletable PrP-sen self-aggregates. Furthermore, this phenomenon was unaffected by the glycosylation state of the PrP-sen molecule (Fig. 4a, open bars). When these studies were repeated in the presence of PrP-res a significant amount of each deletion mutant was associated with the PrP-res pellet (Fig. 4b, closed  bars). Although there was a marginally significant reduction in binding of Ha⌬120 and Ha⌬124 relative to HaWT and Ha⌬94 (p Ͻ 0.05), this was not sufficient to explain the reduced conversion of these constructs in the cell-free assay (Fig. 1c). Analysis of the unglycosylated component of each molecule indicated that binding of the N-terminally truncated PrP molecules was increased in the absence of glycosylation. In contrast, the ability of the HaWT to bind PrP-res was unaltered in its unglycosylated state (Fig. 4b, open bars). We are uncertain why the binding of HaWT is unaffected by its glycosylation state, but it may reflect the high amount of self-aggregation associated with this molecule. Overall, truncation of the N-terminal tail affected the solubility of PrP-sen, and glycosylation appeared to alter the ability of PrP-sen to bind to PrP-res.
Expression in the cell-free assay. PrP-sen expressed without the GPI anchor is predominantly unglycosylated and is not produced as a dimer. The molecular mass of GPI-negative PrP-sen (a) and protease-resistant PrP (b) were calculated from molecular mass markers and run at a different molecular mass from GPI-anchored PrP because of the loss of the GPI anchor. Percentage conversion relative to HaWT (mean Ϯ S.E.; n ϭ 5-6) was calculated from the pixel volume of protease-resistant PrP found between 7 and 21 kDa relative to input PrP-sen (c). Under these conditions, the mean percentage conversion Ϯ S.E. of GPI-negative HaWT was 47.3 Ϯ 3.6%.
PrP-sen in its conversion to PrP-res in a cell-free assay. There is currently no assay available to study conversion of hamster PrP-sen mutants in live cells. However, HaWT has been shown to block PrP-res formation in scrapie-infected mouse neuroblastoma cells (32). We therefore investigated the function of the N-terminal region by expressing the deletion mutants in ScϩN2a cells chronically infected with mouse scrapie (Chandler/RML strain). Under conditions in which complete interference was observed in cells transduced with HaWT, only partial interference was mediated by Ha⌬94, and no interference with PrP-res formation was observed in cells transduced with Ha⌬113, Ha⌬120, or Ha⌬124, relative to the pSFF vector alone (Fig. 5). This was unlikely to have been a result of altered cellular processing of the truncated PrP-sen molecules, because all of the mutated molecules were detected on the surface of transduced ScϩN2a cells (data not shown). Therefore, hamster PrP residues 34 -94 and 34 -113 appeared to be important for interference with the generation of mouse PrP-res in these scrapie-infected mouse cells.
The expression of a foreign PrP-sen species may block PrPres formation by binding PrP-res and blocking further conversion. Therefore, to determine if the reduced interference by the N-terminal deletion mutants was due to a decreased ability to bind mouse-derived PrP-res, we repeated the binding experiments using PrP-res from the Chandler/RML strain of mouse scrapie. The Chandler/RML strain of scrapie was used to persistently infect the ScϩN2a cells used for the interference experiments described above. Under the conditions of the cellfree assay, all HaPrP deletion mutants could bind mouse PrPres (Chandler/RML) at equivalent levels (Fig. 6). Therefore, the inability of the deletion mutants to interfere with PrP-res formation was not due to a reduced ability to bind mouse PrP-res. DISCUSSION Deletion of residues 34 -94 and 34 -113 from within the N-terminal tail of PrP-sen influenced the quantity and confor-mation of PrP-res generated in a cell-free assay. It is unclear how these residues may be influencing PrP conversion. However, the primary amino acid sequence within the 34 -94 deletion is composed of five octapeptide repeats that may influence intramolecular and intermolecular interactions between PrP molecules and/or other cellular factors that are required for efficient conversion (39,40). Extra copies of the octapeptide repeat are associated with heritable TSE disease in both humans (41) and transgenic mice (42,43) and have been shown to induce PrP aggregation and alter PrP processing in tissue culture cells (39,44,45). In the current study, a considerable amount of HaWT PrP was pelleted by centrifugation in the absence of PrP-res, whereas deletion of the octapeptide repeat region in deletion mutant Ha⌬94 significantly reduced the amount of this PrP-sen self-aggregation (Fig. 4a). The reduced ability of PrP to form self-aggregates was associated with a parallel decline in conversion efficiency. The role of these residues might be to enhance PrP polymerization, which might be beneficial for conversion to a protease-resistant form (40). Deletion of additional residues 95-113 (Ha⌬113) caused a further decrease in PrP polymerization and a concurrent decrease in cell-free conversion. Therefore, residues within the N-terminal tail required for PrP-PrP interactions and efficient conversion include both the octapeptide repeat region and amino acids between residue 95-113. Interestingly, insertions and point mutations within this region (residues 95-113) are known to lead to genetic TSE diseases in humans (41) and have been shown to confer biochemical properties reminiscent of PrP-res on the mutated PrP-sen molecule (46).
In our experiments, deletion of residues 95-113 influenced not only the efficiency of PrP conversion in the cell-free assay but also the nature of the PrP-res produced. There was no apparent shift in molecular weight of the largest bands derived from PK digestion of PrP-res generated from unglycosylated Ha⌬94, Ha⌬113, and Ha⌬120 (Fig. 2b). Therefore, in the presence of PrP-res, these constructs could adopt a fully proteaseresistant conformation, which probably reflects the absence of the primary digestion site at residue 89. However, PK digestion of Ha⌬113, Ha⌬120, and Ha⌬124 also resulted in the formation of lower molecular mass bands between 7 and 14 kDa. These bands are likely to be the result of protease digestion at multiple sites downstream of the primary digestion site, since antibody mapping indicated that they were a result of progres- HaWT or HaPrP N-terminal deletion mutants. Transduction efficiency was 60 -80% by indirect immunofluorescence using an antibody to the Gag protein of pSFF, and expression of the HaWT or deletion mutants was confirmed by radioimmunoprecipitation (data not shown). The immunoblot of transduced cell lysates after PK treatment was probed with the R30 anti-peptide antibody. No PK-resistant PrP was detected with the hamster-specific antibody 3F4, and no truncated conversion products were detected with the C-terminal antibody R20 (data not shown). Duplicate lanes have been spliced from the phosphor image. sive PK digestion from the N terminus. 2 In PrP-res derived from HaWT, residues 95-113 might protect these secondary cleavage sites from protease digestion. Alternatively, deletion of these residues may alter the conformation of the final protease-resistant product such that these secondary sites are more readily exposed to protease digestion. The significance of these altered forms of protease-resistant PrP in neurodegeneration and clinical disease remains to be determined. However, their formation in the cell-free conversion assay was template-dependent, since they were not detected in the absence of a PrP-res seed. Furthermore, similarly sized low molecular weight protease-resistant PrP forms have been detected in inherited TSE diseases of humans and in brain homogenates of scrapie-infected rodents (13,(47)(48)(49)(50), suggesting that these proteins and/or the PrP-res from which they are generated may play a role in TSE pathogenesis in vivo.
The conversion efficiency of unglycosylated PrP-sen was not decreased by deletion of residues within the N-terminal region (Fig. 2c). This was in contrast to the effect of these same deletions in the glycosylated protein (Fig. 1c). Interestingly, the binding of the N-terminally truncated PrP-sen to PrP-res was also improved in the absence of glycosylation. However, conversion of unglycosylated PrP appeared to result in a larger proportion of the lower molecular weight species in the protease-resistant PrP. Quantification of the percentage conversion excluding these lower molecular weight species gave a similar pattern for conversion of the PrP deletion mutants (Fig.  2d) as was seen previously for glycosylated PrP (Fig. 1c). Glycosylation may therefore modulate interactions between PrPsen and PrP-res and promote conversion to a more proteaseresistant conformation. Furthermore, the presence of a longer N-terminal region appears to overcome this glycosylation effect, since HaWT and Ha⌬94 were predominantly converted to higher molecular weight protease-resistant forms regardless of glycosylation state. Interactions of the flexible N-terminal tail of PrP-sen with antibodies (51) or other parts of PrP-sen itself (7) have been reported to alter the conformation of the glycosylated C-terminal domain of PrP. Therefore, residues 34 -113, which affected the conversion of glycosylated PrP-sen, might act by modulating the conformation of the C-terminal domain, thus enabling efficient interactions between PrP-sen and PrP-res.
A further post-translational modification of PrP-sen involved expression of the molecule without the signal sequence for the GPI anchor. This modification resulted in a molecule that was mostly unglycosylated, which may have contributed to the improved conversion efficiency of Ha⌬113, Ha⌬120, and Ha⌬124. However, efficient conversion of GPI-negative PrP was still dependent on residues 34 -94, which supported our earlier hypothesis that intermolecular interactions between residues in the PrP octapeptide repeat region may be necessary for efficient conversion of PrP-sen. However, for unglycosylated PrP containing a GPI anchor, residues 34 -94 were not required for efficient conversion (Fig. 2c), suggesting that in the absence of glycosylation, the hydrophobic fatty acids of the GPI anchor of PrP-sen might promote the PrP-PrP interactions required for conversion. Thus, in the absence of a GPI anchor, the N-terminal region may be required to mediate interactions between PrP-sen molecules to promote efficient conversion to a protease-resistant form.
Residues 34 -93 are not required for scrapie susceptibility, since transgenic mice expressing PrP-sen lacking these residues remain susceptible to scrapie infection, albeit with delayed kinetics, reduced levels of detectable PrP-res, and an altered pattern of neurodegeneration (17). The decrease in PrP-res produced from cell-free conversion of Ha⌬94 could correspond with the decreased PrP-res formation and delayed onset of disease observed in these transgenic mice. However, the cause of the altered pattern of neurodegeneration is unknown. Transgenic mice expressing further truncated PrP-sen lacking residues 34 -106 are reportedly not susceptible to scrapie infection (18). This is in contrast with the current study, in which PrP-sen truncated between residues 34 and 113 could be converted to a protease-resistant form. This discrepancy may reflect a simple difference between sequences of PrP-sen that are essential for conversion of mouse PrP versus the hamster PrP conversion presented here. Certainly, different amino acid residues have been shown to influence speciesspecific conversion of mouse and hamster PrP (52,53). However, a more attractive explanation is that residues between positions 94 and 106 are required for initial infection with a TSE agent but are not essential for the subsequent formation of PrP-res. We have shown that the protease-resistant products associated with conversion of N-terminally deleted PrP-sen lack a large portion of the N-terminal region. Thus, it may also be that the antibodies used to detect PrP-res in the transgenic mouse model were specific to regions of PrP not present in the lower molecular weight protease-resistant products described here (17).
Expression of a foreign PrP species in mouse scrapie-infected neuroblastoma cells can block the generation of mouse PrP-res (32). In the present studies, interference with mouse PrP-res formation in ScϩN2a cells was reduced by deletion of residues 34 -94 from hamster PrP-sen and completely eliminated by deletion of residues 34 -113. There was a strong correlation between the effect of deletion of these regions of the PrP molecule on interference in ScϩN2a cells and conversion in the cell-free assay (Fig. 1). It is therefore possible that residues 34 -113 play a similar role in both processes. Previous cell-free conversion studies suggested that interference with PrP-res generation by the expression of a foreign PrP species was primarily caused by inhibition of the acquisition of protease resistance rather than by preventing binding of the homologous PrP (54). Consistent with this conclusion, the ability of the hamster-derived PrP-sen to bind to mouse PrP-res was unaffected by the deletions within the N-terminal tail. This suggests that residues in the flexible N-terminal tail of PrP-sen may prevent the homologous PrP species from acquiring protease resistance by interfering with a step of conversion process subsequent to binding. Interestingly, in the system described here, molecules that have been shown to interfere with PrP-res generation in ScϩN2a cells have contained one or both residues of the 3F4 epitope (32,52), and deletion of this epitope in the Ha⌬113 mutant ablated the ability of HaPrP to induce interference. The noninterfering PrP-sen molecules may also be less able to interact with other components of the conversion process, such as glycosaminoglycans. Zulianello et al. (55) proposed that PrP residues 23-34 may bind to an auxiliary molecule required for conversion. However, as residues 23-34 were present in both interfering and noninterfering deletion mutants described in the present paper, these residues do not appear to account for the interference observed here.