Dominant-negative Effects of the N-terminal Half of Prion Protein on Neurotoxicity of Prion Protein-like Protein/Doppel in Mice*

Prion protein-like protein/doppel is neurotoxic, causing ataxia and Purkinje cell degeneration in mice, whereas prion protein antagonizes doppel-induced neurodegeneration. Doppel is homologous to the C-terminal half of prion protein but lacks the amino acid sequences corresponding to the N-terminal half of prion protein. We show here that transgenic mice expressing a fusion protein consisting of the N-terminal half, corresponding to residues 1-124, of prion protein and doppel in neurons failed to develop any neurological signs for up to 730 days in a background devoid of prion protein. In addition, the fusion protein prolonged the onset of ataxia in mice expressing exogenous doppel. These results suggested that the N-terminal part of prion protein has a neuroprotective potential acting both cis and trans on doppel. We also show that prion protein lacking the pre-octapeptide repeat (Δ25-50) or octapeptide repeat (Δ51-90) region alone could not impair the antagonistic function against doppel.

Prion protein-like protein/doppel is neurotoxic, causing ataxia and Purkinje cell degeneration in mice, whereas prion protein antagonizes doppel-induced neurodegeneration. Doppel is homologous to the C-terminal half of prion protein but lacks the amino acid sequences corresponding to the N-terminal half of prion protein. We show here that transgenic mice expressing a fusion protein consisting of the N-terminal half, corresponding to residues 1-124, of prion protein and doppel in neurons failed to develop any neurological signs for up to 730 days in a background devoid of prion protein. In addition, the fusion protein prolonged the onset of ataxia in mice expressing exogenous doppel. These results suggested that the N-terminal part of prion protein has a neuroprotective potential acting both cis and trans on doppel. We also show that prion protein lacking the pre-octapeptide repeat (⌬25-50) or octapeptide repeat (⌬51-90) region alone could not impair the antagonistic function against doppel.
The normal prion protein (PrP C ) 2 is a glycosylphosphatidylinositol (GPI)-anchored membrane glycoprotein expressed most abundantly in the central nervous system, particularly in neurons, and to a lesser extent in non-neuronal tissues, including the heart, lung, spleen, and kidney (1,2). It is well known that conformational conversion of PrP C into the abnormally folded amyloidogenic isoform, PrP Sc , plays a pivotal role in the pathogenesis of transmissible spongiform encephalopathies or prion diseases, including Creutzfeldt-Jakob disease in humans and bovine spongiform encephalopathy in cattle (1,3). How-ever, the physiological function of PrP C remains largely unknown.
We and others identified a novel gene, Prnd, that encodes a GPI-anchored PrP-like protein, termed Doppel (Dpl), 16 kb downstream of the murine PrP gene Prnp (4,5). Dpl is expressed in the testis, heart, kidney, and spleen of wild-type mice but not in the brain where PrP C is actively expressed. Intriguingly, some lines of mice devoid of PrP C (Prnp 0/0 ), including Ngsk, Rcm0, and Zrch II, ectopically expressed Dpl in their brains, particularly in neurons, because of an unusual intergenic splicing between Prnp and Prnd, developed ataxia, and Purkinje cell degeneration (5,6). However, others, such as Zrch I and Npu, neither ectopically expressed Dpl nor exhibited ataxia and Purkinje cell degeneration (4,5). It was finally confirmed that Dpl is neurotoxic, and PrP C antagonizes the neurotoxicity of Dpl by a demonstration that transgenically expressed Dpl caused ataxia and Purkinje cell degeneration in nonataxic Zrch I Prnp 0/0 mice but not in wild-type mice (7)(8)(9). However, the exact mechanism of the antagonistic interaction of PrP C and Dpl remains unknown.
Dpl shares 23% identity in amino acid composition with PrP (4, 5) and bears conformational similarity to the C-terminal globular domain of PrP C , both comprising three ␣-strands and two short ␤-strands (10). However, Dpl lacks the amino acid sequences corresponding to the N-terminal half of PrP C (4,5). Interestingly, it was shown that PrP with truncated N-terminal residues 32-121 or 32-134, termed PrP⌬32-121 or PrP⌬32-134, respectively, exhibited neurotoxicity similarly to that of Dpl, causing ataxia and cerebellar neurodegeneration in nonataxic Zrch I Prnp 0/0 mice but not in wild-type mice (11,12). Therefore, it might be possible that the neurotoxicity of Dpl is attributable to lack of the corresponding N-terminal part of PrP C . However, this remains to be elucidated.
We previously showed that the N-terminal residues 23-88 of PrP C are involved in the antagonistic function of PrP C against the Dpl neurotoxicity by demonstrating that PrP lacking the residues 23-88 completely lost the ability to rescue Ngsk Prnp 0/0 mice from Dpl-induced Purkinje cell degeneration (13). Residues 23-88 include most of the PrP-specific octapeptide repeat (OR) region, which includes residues 51-90. Recent lines of evidence from cell culture experiments show that the OR may be involved in the neuroprotective function of PrP C (14 -16). However, the biological relevance of OR in the neuroprotective function of PrP C against Dpl is not yet understood in vivo.
In this study, we generated transgenic (tg) mice, tg(PrP⌬OR) and tg(PrPN-Dpl), expressing PrP lacking OR and Dpl fused with the N-terminal half of PrP C , respectively. We also produced tg(PrP⌬preOR) mice expressing PrP without the pre-OR region. By intercrossing these tg mice with mice transgenically overexpressing Dpl in neurons on the genetic background of nonataxic Zrch I Prnp 0/0 , we investigated whether or not these mutant molecules could antagonize Dpl neurotoxicity, rescuing mice from ataxia and Purkinje cell degeneration.

EXPERIMENTAL PROCEDURES
Construction of Transgenes-A DNA fragment corresponding to the N-terminal residues 1-124 of PrP was first amplified by PCR with primer a (5Ј-cccaagcttctcgagatggcgaaccttggc-3Ј, the underlined sequence corresponds to the HindIII and XhoI sites, and the boldface sequence represents a start codon) and primer f (5Ј-cttgatgaaggctccaaggccccccactac-3Ј, the underlined sequence corresponds to DNA encompassing residues 58 -62 of Dpl, and the italic sequence corresponds to DNA encompassing residues 120 -124 of PrP) using PrP cDNA as a template. The resulting DNA fragment, containing the DNA sequence corresponding to residues 58 -62 of Dpl at the 3Ј site, was then utilized as a 5Ј primer to amplify another DNA fragment corresponding to residues 58 -179 of Dpl together with primer i (5Ј-cccaagcttctcgagttacttcacaatgaa-3Ј, the underlined sequence corresponds to the HindIII and XhoI sites, and the boldface sequence represents a stop codon) using Dpl cDNA as a template, resulting in amplification of a DNA fragment for the fusion protein PrPN-Dpl consisting of residues 1-124 of PrP and residues 58 -179 of Dpl. After DNA sequence confirmation of the amplified fragment, it was inserted into a unique SalI site of the Syrian hamster PrP cosmid vector, CosSHa.tet (InPro Biotechnology, Inc. South San Francisco, CA), to construct the PrPN-Dpl transgene.
A DNA fragment corresponding to the N-terminal residues 1-24 of PrP was first amplified by PCR with primers a and c (5Ј-ggtgccaccctgaggctttttgcagaggcc-3Ј, the underlined and italic sequences correspond to DNAs encompassing residues 51-55 and 20 -24 of PrP, respectively) using PrP cDNA as a template. The resulting DNA fragment containing the DNA sequence corresponding to residues 51-55 of PrP at the 3Ј site was then utilized as a 5Ј primer to amplify another DNA fragment corresponding to residues 51-254 of PrP together with primer g (5Ј-cccaagcttctcgagtcatcccacgatcag-3Ј, the underlined sequence corresponds to the HindIII and XhoI sites, and the boldface sequence represents a stop codon) using PrP cDNA as a template, resulting in amplification of a DNA fragment for the deletion protein PrP⌬preOR consisting of residues 1-24 and 51-254 of PrP. After DNA sequence confirmation of the amplified fragment, it was inserted into a unique SalI site of the Syrian hamster PrP cosmid vector, CosSHa.tet (InPro Biotechnology, Inc.), to construct the PrP⌬preOR transgene.
A DNA fragment corresponding to the N-terminal residues 1-50 of PrP was first amplified by PCR with primers a and d (5Ј-atgggtaccccctcctgggtaacggttgcc-3Ј, the underlined and italic sequences correspond to DNAs encompassing residues 91-95 and 46 -50 of PrP, respectively) using PrP cDNA as a template. The resulting DNA fragment containing the DNA sequence corresponding to residues 91-95 of PrP at the 3Ј site was then utilized as a 5Ј primer to amplify another DNA fragment corresponding to residues 91-254 of PrP together with primer g using PrP cDNA as a template, resulting in amplification of a DNA fragment for the deletion protein PrP⌬OR consisting of residues 1-50 and 91-254 of PrP. After DNA sequence confirmation of the amplified fragment, it was inserted into a unique SalI site of the Syrian hamster PrP cosmid vector, CosSHa.tet (InPro Biotechnology, Inc.), to construct the PrP⌬OR transgene.
Generation of Transgenic Mice-The plasmid-derived sequences were removed from each of the transgene constructs, and the resulting DNAs were injected into the zygotes of C57BL/6 mice to generate tg mice as described elsewhere (17,18).
Expression Vectors for Wild-type PrP C , PrP⌬preOR, PrP⌬OR, and PrP⌬23-88-The DNA fragments encoding wild-type mouse PrP C and PrP⌬23-88 were amplified by PCR with a sense primer (5Ј-tcggatccagtcatcatggcgaaccttggc-3Ј; the underlined sequence corresponds to a BamHI site; the boldface sequence corresponds to a start codon) and an antisense primer (5Ј-cctctagacctcatcccacgatcaggaaga-3Ј; the underlined sequence corresponds to an XbaI site; the boldface sequence corresponds to a stop codon) using mouse genomic DNA extracted from wild-type mice and tg(PrP⌬23-88) mice (13). After confirmation of the DNA sequences, each DNA fragment was digested by BamHI and XbaI and introduced into a pcDNA3.1 vector (Invitrogen) to generate pcDNA3.1-moPrP and pcDNA3.1-PrP⌬23-88. pcDNA3.1-PrP⌬preOR and -PrP⌬OR were constructed by digestion of each of the already cloned PCR products with HindIII and subsequent insertion of the digested fragments into a HindIII site of pcDNA3.1 vector.
Diagnosis of Ataxia-The behavior of mice was inspected at least every 3 days evaluating difficulties for walking straight or trembling in their hindquarters on initiation of movement and during walking. When mice showed such abnormal behaviors, they were subjected to a second inspection at least 3 days later. At this time, if the same or exacerbated symptoms were obvious, mice were diagnosed with ataxia, and the date of the first recognition of the abnormal behaviors was registered as the onset of the ataxia. If the symptoms were trivial or difficult to diagnose as ataxia by an investigator, another investigator also inspected the mice to confirm the symptoms. In this case, mice were not diagnosed as ataxia until the two investigators independently confirmed the symptoms.
Western Blotting-Homogenates (10%, w/v) were prepared in a lysis buffer containing 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 0.5% Triton X-100, 0.5% sodium deoxycholate, 1 mM EDTA, and protease inhibitor mixture (Nakalai Tesque Co., Kyoto, Japan) and centrifuged at low speed. Protein concentrations of the resulting supernatant were determined using the BCA protein assay kit (Pierce). Total proteins were electrophoresed through a 12% SDS-polyacrylamide gel and electrically transferred to an Immobilon-P polyvinylidene difluoride membrane (Millipore Corp.). The membrane was immersed in 5% nonfat dry milk containing TBST (0.1% Tween 20, 100 mM NaCl, 10 mM Tris-HCl, pH 7.6) for 1 h at room temperature and incubated with M20 goat polyclonal antibodies (Santa Cruz Biotechnology, Santa Cruz, CA), SAF32 mouse monoclonal antibody (SPI-BIO, Montigny le Bretonneux, France), or FL176 rabbit polyclonal antibodies against human Dpl (Santa Cruz Biotechnology) for 2 h at room temperature in 1% nonfat dry milk containing TBST. The membrane was washed once in TBST for 15 min and three times for 5 min. Signals were visualized using horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) and the ECL system (Amersham Biosciences).
PNGase F Digestion-PNGase F digestion was performed according to the manufacturer's protocol (New England Biolabs, Inc., Ipswich, MA). Briefly, mouse brain homogenates were denatured by boiling for 10 min in the presence of 0.5% SDS and 1% mercaptoethanol and then treated with PNGase F (500 units/liter) in 1% Nonidet P-40 and 0.05 M sodium phosphate, pH 7.5, for 60 min at 37°C.
In Situ Hybridization-In situ hybridization was performed as described elsewhere (8). Briefly, mouse brains were fixed in 4% paraformaldehyde, embedded in paraffin, and sliced to 5 m thickness. The tissue sections were then deparaffinized, digested with 10 mg/ml proteinase K for 10 min at 37°C, and soaked in 0.25% acetic anhydride, 0.1 mM triethanolamine hydrochloride, pH 8.0, 0.9% NaCl for 10 min. After this, the sections were hybridized with PrP cRNA probes labeled with digoxigenin-UTP (Roche Diagnostics) in buffer (50% formamide, 10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 0.6 M NaCl, 0.5 mg/ml yeast tRNA, 0.25% SDS, 5ϫ Denhardt's solution) at 50°C for 16 h, and followed by several washes in 4ϫ SSC and immersion in 50% formamide, 2ϫ SSC at 50°C for 30 min. The probe used for PrP was derived from the PCR product corre-sponding to PrP residues 26 -187. The hybridized sections were then digested with 20 g/ml RNase A at 37°C for 30 min and finally washed in 0.2ϫ SSC at 50°C for 20 min. Signals were detected by enzyme-linked immunosorbent assay using alkaline phosphatase-conjugated anti-digoxigenin Fab fragments (1:500, Roche Diagnostics) and nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate.
Immunohistochemistry-Deparaffinized sections were placed in 3% H 2 O 2 in methanol for 30 min at room temperature to abolish endogenous peroxidase activity. The tissue sections were incubated overnight at 4°C with anti-spot 35 (calbindin) polyclonal antibodies, IBL-N rabbit antibodies against the N-terminal peptide of PrP (Immuno Biological Laboratories, Gunma, Japan), and ICSM-18 monoclonal antibody recognizing residues 146 -159 of murine PrP. To detect immunoreactivities, we used the EnVisionϩ system in accordance with the manufacturer's recommendations (Dako, Glostrup, Denmark). The antibody-bound peroxidase was detected with 0.04% diaminobenzidine (Sigma).

RESULTS
Generation and Characterization of tg(PrPN-Dpl), tg(PrP⌬preOR), and tg(PrP⌬OR) Mice-The amino acid alignment of PrP and Dpl depicts the homology between the C-terminal regions of the two proteins, corresponding to the residues 125-254 of PrP and 58 -179 of Dpl, both of which form a neurotoxic globular structure with three ␣-helices and two ␤-strands (Fig. 1A). Therefore, to evaluate the effects of the N-terminal region of PrP on Dpl in cis, the PrP N-terminal residues 1-124 were fused to the Dpl residues 58 -179 to make the PrPN-Dpl transgene (Fig. 1A). The PrP⌬preOR and PrP⌬OR, PrP deletion mutants lacking the N-terminal residues 25-50 and 51-90, respectively, were also constructed to examine the involvement of each region in protection from the Dplinduced neurodegeneration (Fig. 1A). We introduced each corresponding DNA into the Syrian hamster PrP cosmid vector, CosSHa.tet (20), allowing each of the mutant proteins to be expressed under the control of the hamster PrP promoter (Fig.  1B). These transgenes were then microinjected into fertilized eggs of C57BL/6 mice, yielding four founders from the PrPN-Dpl transgene and two from each of the PrP⌬preOR and PrP⌬OR transgenes. All of these founders successfully transferred the transgenes into their offspring. These tg mice were successively intercrossed with nonataxic Zrch I Prnp 0/0 mice to eliminate endogenous PrP C .
We also performed immunohistochemical analysis of cerebella from these tg mice with the Zrch I Prnp 0/0 background using two different antibodies, rabbit polyclonal IBL-N and mouse monoclonal ICSM-18 antibodies, which are directed against residues 24 -37 and 146 -159 of murine PrP, respectively. Both antibodies showed no immunoreactivities in the cerebella of Zrch I Prnp 0/0 mice (Fig. 3, E-H). In contrast, the molecular and granule cell layers of normal C57BL/6 mice were clearly stained with both antibodies (Fig. 3, A-D). However, there seemed to be no immunoreactivity in the Purkinje cell layer (Fig. 3, A-D). These staining patterns of PrP C in the cerebellum of normal mice were consistent with previous reports (21)(22)(23). PrP⌬preOR mutant protein was expressed in the cerebella of tg(PrP⌬preOR)/Prnp 0/0 mice indistinguishably from PrP C in C57BL/6 mice, detectable in the molecular and granule cell layers but not in the Purkinje cell layer (Fig. 3, K and L). PrP⌬OR and PrPN-Dpl mutant proteins were also expressed in the molecular and granule cell layers of tg(PrP⌬OR)/Prnp 0/0 and tg(PrPN-Dpl)/Prnp 0/0 mice, respectively (Fig. 3, M-R). However, the mutant proteins were more abundant in the granule cell layer than in the molecular layer (Fig. 3, M-R). Moreover, in tg(PrP⌬OR)/Prnp 0/0 mice, the Purkinje cell layer was   devoid of the signal, but the basolateral area surrounding some but not all Purkinje cells was strongly stained (Fig. 3, M and N and Fig. 7). In tg(PrPN-Dpl)/Prnp 0/0 mice, the cell bodies of Purkinje cells appeared positive, and some cells scattered in the granule cell layer were strongly stained in the cell bodies (Fig. 3,  Q and R). These cells are currently unidentified. Moreover, cortical neurons of tg(PrPN-Dpl)/Prnp 0/0 mice but not wild-type and tg(PrP⌬OR)/Prnp 0/0 mice were positively stained in the cell bodies by IBL-N antibodies (data not shown).
PrPN-Dpl Delays Onset of Dpl-induced Ataxia and Purkinje Cell Degeneration in Mice-No tg(PrPN-Dpl)/Prnp 0/0 mice showed any abnormal symptoms, including ataxia, up to 730 days after birth, at the time of writing (Fig. 4A). Purkinje cells were also unaffected in these mice (data not shown). The signals visualized by anti-Dpl antibodies on a Western blot of brain homogenates from tg mice was about 35% that of Ngsk Prnp 0/0 mice, ectopically expressing Dpl in neurons under the control of the PrP promoter (Fig. 2C). These results indicate that, unlike wild-type Dpl, the fusion protein PrPN-Dpl might be nontoxic to Purkinje cells even in the absence of PrP C , although we could not completely rule out the possibility that the lack of neurotoxicity of PrPN-Dpl is because of its lower expression.

PrP⌬23-88 Is Expressed in the Cerebellum of Mice and on the Surface of Cultured Cells
Similarly to Wild-type PrP C -We previously showed that PrP⌬23-88 was incompetent to rescue Ngsk Prnp 0/0 mice from the Dpl-induced Purkinje cell degeneration, indicating that the region comprising the residues 23-88 is important for PrP C to be protective against Dpl (13). To further gain insights into the role of the residues 23-88 in the neuroprotective function of PrP C , we investigated cytological expression of PrP⌬23-88 in the cerebellum of mice. The cerebella from tg(PrP⌬23-88) mice on the Ngsk Prnp 0/0 background as well as from Zrch I Prnp 0/0 and tg(PrP⌬OR)/Prnp 0/0 mice were subjected to immunohistochemistry using IBL-N and ICSM-18 antibodies. Consistent with the results shown in Fig. 3, no signals could be detected in Zrch I Prnp 0/0 mice, and tg(PrP⌬OR)/Prnp 0/0 mice showed abundant expression of PrP⌬OR in the molecular and granule cells layers but not in the Purkinje cell layer (Fig. 7). PrP⌬23-88 was detected in the molecular and granule cell layers but not in the Purkinje cell layer (Fig. 7), similarly to PrP⌬OR (Fig. 7) and wild-type PrP C (Fig. 3, A-D). We also investigated the cell surface expression of PrP⌬23-88 using cultured cells in comparison with that of wild-type PrP C and two other neuroprotective mutants, PrP⌬preOR and PrP⌬OR. COS-7 monkey kidney cells were transiently transfected with each expression vector and then subjected to flow cytometry analysis using SAF61 monoclonal antibodies against PrP-(142-160) residues. PrP⌬23-88 was detected on the cell surface of COS-7 cells similarly to that of wild-type PrP C , PrP⌬preOR, and PrP⌬OR (Fig. 8). These results indicate that lack of the residues 23-88 neither alter cell types for PrP⌬23-88 to be expressed in the cerebellum of mice nor impair the cell surface expression of PrP⌬23-88.

DISCUSSION
Accumulating evidence indicates a neuroprotective role for PrP C . For instance, Prnp 0/0 mice are highly sensitive to ischemic or traumatic brain damage, developing more severe pathological changes than in wild-type mice (24 -27). In contrast, Dpl, the first identified structural homologue of the C-terminal domain of PrP C , is neurotoxic causing ataxia and Purkinje cell degeneration in mice (7)(8)(9). Interestingly, PrP C functionally antagonizes the neurotoxicity of Dpl, preventing the neurode-   a Expression levels were compared with those of PrP in wild-type mice using Western blotting. b The times were expressed as mean Ϯ S.E. days after birth. c These 24 mice were produced by breeding of tg(Dpl32)/Prnp 0/0 mice with tg(PrPN-Dpl)/Prnp 0/0 mice, tg(PrP⌬preOR)/Prnp 0/0 mice, and tg(PrP⌬OR)/Prnp 0/0 mice. d These times were slightly different from those previously reported (8) probably due to more strict diagnostic criteria for ataxia. e These 15 mice were produced by breeding tg(Dpl32) mice with Zrch I Prnp 0/0 mice. generation (7)(8)(9). However, the mechanism of the antagonistic interaction between PrP C and Dpl or the truncated PrPs remains to be elucidated. trans and cis Neuroprotection by the N-terminal Domain of PrP C against Dpl in Mice-In this study, we showed that PrPN-Dpl (the N-terminal residues 1-124 of PrP C fused with the residues 58 -179 of Dpl) was itself nontoxic and could mitigate the neurotoxicity of wild-type Dpl in Zrch I Prnp 0/0 mice, prolonging the times to the onset of ataxia and Purkinje cell degenera-tion. Residues 58 -179 of Dpl are homologous to residues 125-254 of PrP (10), which encompasses the neurotoxic PrP⌬32-134 peptide. Drisaldi et al. (16) showed that Dpl lacking the N-terminal residues 29 -49 or 50 -90 was still neurotoxic to primary granule cells from Zrch I Prnp 0/0 mice. It is therefore very likely that Dpl-(58 -179) is neurotoxic, similarly to the wild-type Dpl in mice devoid of PrP C . Thus, these results indicate that the N-terminal region of PrP might have neuroprotective potential acting both cis and trans on Dpl in mice. Interestingly, Rossi et al. (28) showed that Zrch II Prnp 0/0 mice, which develop ataxia and Purkinje cell degeneration because of the ectopic expression of Dpl in Purkinje cells, could be rescued by breeding with tga20 mice expressing PrP C abundantly in the molecular and granule cells but not in Purkinje cells. This suggests that PrP C expressed by neighboring cells, such as molecular and granule cells, is able to counteract the neurotoxicity of Dpl that is expressed on Purkinje cells and that the trans neuroprotection of PrP C might involve intercellular counteraction against Dpl.
OR Is Dispensable for Neuroprotective Function of PrP C against Dpl in Mice-In this study, we also showed that PrP⌬OR, PrP lacking the OR alone, rescued mice from the ataxia and Purkinje cell degeneration induced by Dpl. This clearly indicates that the OR is unnecessary for PrP C to antagonize the neurotoxicity of Dpl in mice. Interestingly, Shmerling et al. (11) described that the OR is also unnecessary for PrP C to antagonize the neurotoxicity of truncated PrPs. They showed that granule cell death induced by PrP⌬32-134 could be abrogated by PrP⌬32-93, which lacks the entire OR and about 2/3 of the pre-OR in mice (11). In contrast, in primary cultures of granule cells from Zrch I Prnp 0/0 mice, apoptotic cell death induced by transient overexpression of Dpl could be successfully rescued by wild-type PrP C but not by PrP lacking the OR (16). Dpl was preferentially toxic to Purkinje cells and not to granule cells in mice (8,28,29). Therefore, Dpl toxicity may vary in primary cultured granule cells and mouse models. However, why PrP lacking the OR has differential activity against Dpl in primary cultured granule cells and mice is unknown. Kuwahara et al. (31) showed that hippocampal neuronal cell lines established from Prnp 0/0 mice easily succumbed to apoptosis after serum withdrawal. Furthermore, expression of the anti-apoptotic molecule Bcl-2 could rescue cell lines from apoptosis (31). Bounhar et al. (14) also showed that PrP C prevented human primary neurons from Bax-induced apoptosis. This suggests that the neuroprotective function of PrP C might involve anti-apoptotic activities. Interestingly, PrP lacking OR failed to rescue the cells from serum withdrawal-and Bax-induced apoptosis, indicating that the OR plays an important role in the anti-apoptotic function of PrP C (14,32). Furthermore, our present results showing that PrP⌬OR antagonized Dpl in mice clearly indicates that neuroprotection by PrP C against Dpl is not associated with OR-mediated anti-apoptotic activities.
The anti-apoptotic activity of PrP C may also be associated with anti-oxidative responses (32,33). Binding of PrP C to copper may be important for the anti-oxidative function of PrP C by either chelating copper or by activating anti-oxidant enzymes, such as Cu,Zn-superoxide dismutase, via transfer of the bound copper to the enzymes, or both (34 -36). Six conserved histi- dine residues have been identified as copper-binding sites in human PrP C , with four in the OR and two at positions 96 and 111 (37). As PrP⌬OR blocked Dpl-mediated neurotoxicity, ORmediated copper binding might not be involved in the neuroprotection of PrP C against Dpl. In addition, our previous result that PrP⌬23-88, in which two other histidine residues are preserved, failed to rescue mice from ataxia and Purkinje cell degeneration, indicate that copper binding at these sites might not be relevant to the antagonistic function of PrP C against Dpl. Taken together, these suggest that the copper binding-mediated function of PrP C , including anti-oxidative activity, is not associated with its neuroprotective function against Dpl. However, we cannot rule out copper binding to all histidine residues simultaneously for PrP C to have anti-oxidative function.
N-terminal Residues and the Neuroprotective Function of PrP C against Dpl in Mice-In this study, we also showed that PrP⌬preOR, PrP lacking residues 25-50, prevented Dpl-induced ataxia and Purkinje cell degeneration in mice as efficiently as PrP⌬OR. This indicates that N-terminal residues 25-50 are not required for PrP C to antagonize Dpl in mice. The two deletions, ⌬25-50 and ⌬51-90, almost entirely cover the region deleted in PrP⌬23-88, which failed to rescue mice from the neurotoxicity of Dpl (13). PrP⌬23-88 is a chimeric protein of mouse and hamster PrPs, containing two methionines at 108 and 111 in mouse PrP instead of leucine and valine. No such substitutions were present in PrP⌬preOR and PrP⌬OR. However, we previously showed that Ngsk Prnp 0/0 mice were successfully rescued from ataxia and Purkinje cell degeneration by full-length chimeric PrP with these methionine substitutions (13), clearly indicating that the incompetence of PrP⌬23-88 to antagonize Dpl is because of lack of residues 23-88 and not to the amino acid substitutions. We also showed here that PrP⌬23-88, PrP⌬preOR, and PrP⌬OR were similarly expressed in the cerebellum of mice, consistent with these mutant molecules being expressed under the control of the same hamster PrP promoter/enhancer. Moreover, in this study, we used tg(Dpl32)/Prnp 0/0 mice for the rescue experiments instead of Ngsk Prnp 0/0 mice because tg mice develop ataxia and Purkinje cell degeneration on the Zrch I Prnp 0/0 background much earlier than Ngsk Prnp 0/0 mice because of higher expression of Dpl in their brains (8). Dpl was expressed in tg(Dpl32) mice from the neuron-specific enolase promoter and in Ngsk Prnp 0/0 mice from the residual PrP promoter (4,8). However, Dpl was similarly expressed in neurons of tg(Dpl32) mice and Ngsk Prnp 0/0 mice with the highest expression in Purkinje cells and hippocampal neurons (4,8). Therefore, Dpl is toxic to Purkinje cells in the same way in both tg (Dpl32)/Prnp 0/0 mice and Ngsk Prnp 0/0 mice. Taken together, these results indicate that PrP⌬preOR and PrP⌬OR but not PrP⌬23-88 can antagonize Dpl neurotoxicity in mice.
PrP⌬preOR and PrP⌬OR but not PrP⌬23-88 have the N-terminal two amino acids (residues 23 and 24) conserved adjacent to the junction with the signal peptide. Thus, the two amino acids may be important for the neuroprotection of PrP C against Dpl. This may be consistent with the observation that PrP⌬32-93 protected against the truncated PrPs (11). Interestingly, in PrP⌬preOR the two amino acids are followed by residues starting from 51, generating a new N-terminal sequence (KKPQGGTWG), which is very similar to the N-terminal 9 residues (KKRPKPGGW) of wild-type PrP C and PrP⌬32-93. Six out of 9 of these amino acids are identical. Therefore, this new N-terminal sequence might mimic the function of wildtype PrP C . In PrP⌬OR, the N-terminal sequence is intact. Thus, these N-terminal residues might be important for the neuroprotection of PrP C against Dpl. However, it is possible that the antagonistic function of PrP C against Dpl is impaired only by a large deletion of the N-terminal domain with or without the N-terminal residues, as observed in PrP⌬23-88.
Interestingly, PrP with only the central residues 105-125 or 94 -134 deleted was reported to be neurotoxic, causing cerebellar degeneration or demyelination in mice, respectively (38,39). These results suggest that these central residues are essential for PrP C to be neuroprotective. However, PrP⌬23-88 contains these central residues but has no protective activity against Dpl (13). Therefore, the central residues alone might not be enough for PrP C neuroprotectivity, and other region(s), present among the N-terminal residues 23-88, may also be necessary for neuroprotection. These region(s) might be located in the N-terminal 2 or 9 residues. However, unrelated region(s) to the N-terminal 2 or 9 residues may also be necessary.
Possible Mechanisms for N-terminal Region Neuroprotectivity of PrP C against Dpl-There are reports showing that the N-terminal domain is involved in the subcellular trafficking of PrP C (40 -44). In this study, we found that PrP⌬23-88, PrP⌬preOR, and PrP⌬OR were expressed in the molecular and granule cell layers of the cerebellum and on the cell surface of COS-7 monkey kidney cells similarly to that in wild-type PrP C . This indicates that the cellular expression and cell surface transport of these mutant molecules may be unchanged. It is therefore unlikely that the cell surface localization of PrP⌬23-88 is different from that of PrP⌬preOR and PrP⌬OR because of the large deletion of the N-terminal domain, thus impairing the neuroprotective function of PrP C . The N-terminal part is also involved in efficiency of PrP C endocytosis. PrP⌬23-90 and PrP⌬48 -93, which lacks the OR region, were shown not to be efficiently internalized in mouse neuroblastoma N2a cells (44), indicating that lack of the OR alone might affect the internalization of PrP C . However, we showed here that PrP⌬OR was neuroprotective against Dpl in mice, indicating that the internalization may not be relevant to the neuro-protective activity of PrP C . Recently, Santuccione et al. (45) showed that PrP C activates p59 Fyn to enhance neurite outgrowth via recruitment of the neuronal cell adhesion molecule to lipid rafts, indicating that the proper localization at lipid rafts could be important for PrP C function. Interestingly, PrP⌬23-90 but not PrP lacking the OR region was not properly targeted to lipid rafts (44). Thus, PrP⌬23-88 but not PrP⌬OR and PrP⌬preOR may not properly localize at lipid rafts either because of lack of the N-terminal 2 or 9 residues or because of large scale deletion of the N-terminal domain with or without the N-terminal residues, resulting in unsuccessful rescue of mice from Dpl neurotoxicity.
Alternatively, the N-terminal region may be involved in the neuroprotective function of PrP C by eliciting a neuroprotective signal through an associated molecule, as in the models proposed so far (11,30,38,39,46). Among them, Weissmann and Aguzzi (46) proposed that PrP C binds to an as yet unidentified molecule and elicits a Purkinje cell survival signal through the N-terminal domain. Dpl can bind to the molecule but cannot generate the signal because of lack of the N-terminal domain, resulting in Purkinje cell degeneration. However, PrP C competes with Dpl for the molecule, thereby preventing Dpl-induced Purkinje cell degeneration. The results showing that PrPN-Dpl, PrP⌬preOR, and PrP⌬OR but not PrP⌬23-88 antagonize the neurotoxicity of Dpl suggests that the former three molecules bind the molecule and produce the survival signal through the N-terminal domain of PrP C , preventing neurodegeneration. This may be because they have a part of or the whole N-terminal domain. It might be also possible that Dpl itself may bind to its own unidentified cognate molecule to elicit a neurotoxic signal and PrP C , PrPN-Dpl, PrP⌬preOR, and PrP⌬OR but not PrP⌬23-88 may compete for the molecule via a part of or the whole N-terminal domain, thereby preventing Dpl-mediated neurotoxicity. However, these models can be verified only if the hypothetical molecules are identified.
In this study, we showed that the N-terminal domain mediates the neuroprotective function of PrP C against Dpl in trans and cis and that the OR region and residues 25-50 (pre-OR) are dispensable for the neuroprotective function of PrP C . However, to understand the exact molecular mechanism how the N-terminal domain is involved in the neuroprotective function of PrP C , further studies are required.