PrP knockout mice in which only the open reading frame was disrupted (‘Zürich I’) remained healthy. However, more extensive deletions resulted in ataxia, Purkinje cell loss and ectopic expression in brain of Doppel (Dpl), encoded by the downstream gene, Prnd. A new PrP knockout line, ‘Zürich II’, with a 2.9 kb Prnp deletion, developed this phenotype at ∼10 months (50% morbidity). A single Prnp allele abolished the syndrome. Compound Zürich I/Zürich II heterozygotes had half the Dpl of Zürich II mice and developed symptoms 6 months later. Zürich II mice transgenic for a Prnd‐containing cosmid expressed Dpl at twice the level and became ataxic ∼5 months earlier. Thus, Dpl levels in brain and onset of the ataxic syndrome are inversely correlated.
PrP, the prion protein, plays a central role in the pathogenesis of transmissible spongiform encephalopathies such as scrapie or BSE. The normal form of PrP, designated PrPC, is encoded by a single‐copy gene (Basler et al., 1986) and is expressed in the brain of healthy and prion‐infected organisms to about the same extent (Chesebro et al., 1985; Oesch et al., 1985).
Several PrP knockout lines have been generated that differ in the design of the gene disruption (Figure 1). The first two lines described, hereafter called Prnp0/0 Zürich I and Prnp−/− Edinburgh, developed and reproduced normally (Büeler et al., 1992; Manson et al., 1994). In Prnp0/0 Zürich I mice, PrP codons 4–187 (72% of altogether 254 codons) were replaced by a neomycin (neo) cassette. Mice homozygous for the disrupted gene expressed transcripts containing the neo and the residual Prnp sequence in the brain, but no PrPC‐containing protein was detected. In the Prnp−/− Edinburgh mice the PrP gene was disrupted by the insertion of a neo cassette into a unique KpnI site following residue 93 of the PrP open reading frame (ORF). No PrP mRNA or PrP‐related protein was detected in the brain of homozygous Prnp−/− Edinburgh animals.
In a third PrP knockout line, here designated Prnp−/− Nagasaki, a 2.1 kb genomic DNA segment comprising 0.9 kb of intron 2, 10 bp of 5′ non‐coding region, the entire PrP ORF and 0.45 kb of the 3′ non‐coding region, was replaced with a neo cassette (Sakaguchi et al., 1996). These mice developed normally but exhibited severe ataxia and Purkinje cell loss in later life (Sakaguchi et al., 1996) as well as demyelination of peripheral nerves (Nishida et al., 1999). Because this phenotype was abolished by introduction of a PrP transgene (Nishida et al., 1999) it was concluded that both ataxia and peripheral nerve degeneration were due to the absence of PrP. A fourth line, Rcm0, was cited in Moore et al. (1999) and Silverman et al. (2000) as resembling the Nagasaki line with respect to the extensive PrP gene deletion and the ataxic syndrome.
Recently a gene, Prnd, was found 16 kb downstream of the murine PrP gene, which comprises an ORF encoding a 179‐residue protein (Moore et al., 1999). The predicted protein, which was named Doppel (Dpl), has ∼25% identity with the C‐terminal two‐thirds of PrP. Prnd‐derived mRNA is expressed from a promoter upstream of exon 1 of Prnd at relatively high levels in testis and heart but at very low levels in brain of wild‐type mice. However, in Nagasaki and Rcm0 mice, but not in Zürich I mice, Prnd‐specific RNAs were present at relatively high levels in brain. These transcripts originate at the Prnp promoter, run beyond the Prnd ORF and are processed by one or more splicing events that link the 3′ end of the second PrP exon directly or indirectly to the Dpl‐encoding exon (Moore et al., 1999; Li et al., 2000). Dpl has been identified as an N‐glycosylated, GPI‐linked protein in testes of normal, and in brain of Rcm0 PrP knockout, mice carrying the extensive PrP gene deletion (Silverman et al., 2000). The cerebellar syndrome was therefore attributed to the ectopic expression of Dpl in the brain (Moore et al., 1999; Li et al., 2000).
Here, we describe the generation of a further PrP knockout line, hereafter called Prnp−/− Zürich II, in which the PrP‐encoding exon and its flanking regions were replaced by a loxP site. The homozygous mice, like their Prnp−/− Nagasaki counterparts, developed progressive ataxia and age‐dependent Purkinje cell loss starting at 5–6 months, with half the mice affected by ∼10 months, and showed ectopic expression of two Prnd‐derived mRNAs and Dpl in brain. A single wild‐type PrP allele fully corrected the deleterious phenotype in agreement with previous reports (Sakaguchi et al., 1996; Nishida et al., 1999). The F1 offspring of a cross between the Zürich I and Zürich II lines showed partial complementation, in that they remained healthy 6 months longer than the Zürich II mice, despite the absence of a PrP ORF. The level of Prnd‐derived mRNAs and Dpl in brain was half that in Zürich II mice.
Introduction into Zürich II mice of a cosmid comprising both the Prnd gene and a Prnp locus lacking the PrP ORF resulted in accelerated development of the ataxic phenotype and was associated with increased levels of Prnd‐derived transcripts and Dpl in brain. These results suggest that the allele with deletions extending beyond the PrP ORF is pathogenic in a dose‐dependent fashion and that this pathogenicity is due to ectopic expression in brain of Dpl (Moore et al., 1999; Li et al., 2000) rather than to the absence of PrP. N‐proximally truncated PrP, which resembles Dpl, also causes ataxia that, as in the case of ectopic Dpl expression, is counteracted by full‐length PrP (Shmerling et al., 1998).
Mice carrying a deletion of the PrP ORF and its flanking sequences develop a cerebellar syndrome
We generated a PrP‐deficient mouse line by gene targeting in embryonic stem cells, as described in Materials and methods and in Figure 2A. In this mouse the wild‐type PrP allele was replaced by the Prnp lox1 allele (to be referred to hereafter as Prnp− Zürich II allele), in which 0.26 kb of intron 2, 10 bp of the 5′ flanking sequence, the entire ORF, the 3′ non‐coding region of exon 3 and 0.6 kb of 3′ adjacent sequence were replaced by a 34 bp loxP sequence.
Mice carrying the Prnp− Zürich II allele were bred to homozygosity; hemizygous Zürich II (Prnp−/+) and homozygous Zürich II (Prnp−/−) mice were born in the ratio expected from Mendelian inheritance. In brains of Prnp−/− Zürich II mice no PrP‐specific mRNA was found (Figure 2D), nor could PrP be detected by immunoblotting (Figure 2E) or immunohistochemistry (Figure 4B).
Prnp−/− Zürich II mice were normal until ∼6 months of age, when they began to develop intention tremor and trembling gait (Figure 3A); this phenotype, which was taken as criterion for diagnosis of the cerebellar syndrome, was present in 50% of the mice by 10 months, and in 100% by 13 months (n = 63) (Figure 3B). Progression of the disease led to hypotony of the hind limbs and impaired equilibrium on a moving surface (Crawley and Paylor, 1997) as compared with wild‐type animals (P.Valenti and A.Cozzio, data not shown). The hindlegs seemed more affected than the forelegs. Histological analysis revealed a severe age‐dependent loss of Purkinje cells in Prnp−/− Zürich II mice (Figure 4A), which was most prominent in the vermis regions I–VIII (Figure 5), the simple lobule, crus I and II ansiform lobules, paramedian lobule and copula pyramis, while vermis regions IX and X were less affected (data not shown). The average cell loss in vermis I–VIII was ∼40 and 80% at 30 and 63 weeks, respectively. Concomitant with the gaps in the monolayer of Purkinje cells, the thickness of the molecular layer in these areas was reduced, probably due to the loss of the dendritic trees of the Purkinje cells. Importantly, the granular cell layer as the major afferent source of the Purkinje cells showed no alterations that could be held responsible for secondary damage of their target cells.
Mice homozygous for the Prnp lox2 allele (in which the PrP reading frame and surrounding regions were flanked by loxP sequences) were derived from the same founder as the Prnp−/− Zürich II mice, as explained in Materials and methods. These animals also remained healthy for at least 57 weeks and showed no Purkinje cell loss (data not shown).
Northern analysis of 5 μg of poly(A)+ mRNA from Zürich II mouse brains showed two Prnd‐specific RNAs of ∼4.3 and 2.7 kb, with about the same intensity as the bands from the same amount of poly(A)+ mRNA from testis, which were slightly smaller, 3.9 and 2.2 kb (Figure 6A). The molecular weights are similar to those determined by Li et al. (2000), 3.4 and 2.2 kb, but considerably higher than those reported by Moore et al. (1999), 2.7 kb and 1.7 kb for both brain and testis, perhaps because of discrepancies in the molecular weight determinations. No Prnd‐specific bands were detected in wild‐type brain. Prnp0/0 Zürich I brain showed a single weak band at ∼3.4 kb, with ∼1/20th the intensity of the sum of the two bands in Zürich II brain. Its origin was not investigated further, but the transcripts may reflect synthesis from the Prnd promoter (Moore et al., 1999).
Western analysis using a polyclonal antibody against recombinant Dpl (Figure 6B) reveals a strong Dpl band in brain of Zürich II mice, almost as intense as in testes of wild‐type mice, but no detectable Dpl in Zürich I or wild‐type brain (less than ∼5% of that in Zürich II brain).
F1 offspring of a Zürich I × Zürich II cross have a mitigated ataxic phenotype
Five breeding pairs (Prnp0/0 Zürich I × Prnp−/− Zürich II) yielded 33 F1 offspring with a compound heterozygote Prnp0/− Zürich I/II genotype; 13 of these were evaluated over the long term (Figure 3B). They remained healthy until 12 months of age and only then began exhibiting the same cerebellar symptoms as those seen in Prnp−/− Zürich II mice at 5–6 months. After 17 months, all F1 mice showed symptoms (Figure 3B). Histological analysis revealed that at 6 weeks both homozygous Prnp−/− Zürich II and compound heterozygous Prnp0/− Zürich I/II mice showed a normal laminar arrangement (Figure 5). At 30 weeks of age, the compound heterozygotes showed no statistically significant cell loss (average of vermis I–VIII, −15%), and at 63 weeks, when Prnp−/− Zürich II mice had lost 80% of their Purkinje cells, the heterozygotes showed an average loss of only 30%. At 95 weeks, ∼70% of the Purkinje cells were lost (Figures 4A and 5), with the exception of vermis IX–X where the loss was only ∼25% (data not shown). Offspring of intercrosses of compound heterozygous Prnp0/− Zürich I/II mice showed co‐segregation of the cerebellar phenotype with the Zürich II allele. Among 28 offspring, six mice were homozygous Prnp0/0 Zürich I, nine were homozygous Prnp−/− Zürich II and 13 animals were compound heterozygous Zürich I/II. At 30 weeks of age, all seven homozygous Zürich II mice observed showed the cerebellar phenotype; two mice were killed for histological examination and showed ∼40% Purkinje cell loss (data not shown). The six Zürich I mice remained healthy for up to at least 48 weeks.
A cosmid expressing Prnd‐derived mRNAs but not PrP mRNA in brain of Zürich II mice causes accelerated appearance of the ataxic syndrome
The DNA segment contained in Prnpb cosmid cos6.I/LnJ‐4 extends from 19 kb upstream to 19 kb downstream of the PrP ORF and comprises the Prnd locus (Westaway et al., 1994; Moore et al., 1999). In connection with a different project, the PrP ORF was replaced with the 900 bp tet‐dependent transactivator ORF (Gossen and Bujard, 1992; Furth et al., 1994; Schultze et al., 1996). Nuclear injection yielded founders transgenic for this modified cosmid, Cos‐tTA, (A.Cozzio, D.Rossi and C.Weissmann, unpublished data). In order to determine whether an allele containing the flanking regions of the third PrP exon but not the PrP ORF could abrogate the ataxic syndrome, the cosmid transgene was bred into Zürich II mice. Unexpectedly, mice homozygous for the Zürich II (Prnp−) allele and containing Cos‐tTA showed a much accelerated course of the disease: onset of ataxia was at ∼19 weeks (50% morbidity) (Figure 3B) and Purkinje cell loss in vermis I–VIII was 40 and 70% at 6 and 30 weeks, respectively (Figure 5). Expression of tTA at a 4‐fold higher level in mice transgenic for a similar vector [mouse line Prnp‐tTA/F959 (Tremblay et al., 1998)], which, however, does not contain the Prnd locus, has no deleterious effects up to at least 10 months of age (A.Servadio and D.Rossi, unpublished results). For reasons that are not immediately clear, brains of Zürich II mice containing the Prnd cosmid showed about twice the level of chimeric Prnd‐derived mRNAs (Figure 6A) and Dpl protein (Figure 6B) compared with Zürich II brain. Maybe the replacement of the PrP ORF by a prokaryotic sequence impairs splicing to the acceptor site of the third exon (Lavigueur et al., 1993; Zandberg et al., 1995; Chiara et al., 1996; Dominski and Kole, 1996). Thus, there is again an inverse correlation between the time to appearance of ataxic symptoms and the level of Prnd‐derived mRNAs and Dpl.
Abrogation of the effects of the Zürich II allele by a single Prnp wild‐type allele or a PrP‐encoding transgene cluster
Hemizygous Prnp−/+ Zürich II mice exhibited no cerebellar symptoms up to at least 100 weeks (Figure 3B). Because, as described above, a single Zürich II allele causes disease at ∼68 weeks (50% morbidity), we conclude that a single Prnp+ allele can abrogate the deleterious effects of the Prnp− Zürich II allele for practically the lifespan of the mouse.
In a further experiment, Prnp−/− Zürich II mice were crossed with Tga20 mice. These mice carry multiple PrP transgenes devoid of the large intron on a Zürich I PrP knockout background and express PrP in most areas of the brain with the distinct exception of Purkinje cells (Fischer et al., 1996). The F1 progeny were back‐crossed with Zürich II mice to eliminate the Zürich I allele. None of the resulting offspring presented clinical symptoms by 90 weeks of age (Figure 3B). There was no loss of Purkinje cells at 63 weeks and 40% loss at 90 weeks (Figure 5), showing that expression of PrP strongly mitigated the cerebellar phenotype even when it was not expressed in Purkinje cells at a discernible level (Fischer et al., 1996) (Figure 4B).
Five independent PrP knockout mouse lines have been reported. Three of these show cerebellar symptoms and loss of Purkinje cells on ageing, namely the Prnp−/− Nagasaki mice (Sakaguchi et al., 1996), the Rcm0 mice of Moore et al. (Moore et al., 1999; Silverman et al., 2000) and the Prnp−/− Zürich II mice (this paper), while the Prnp−/− Edinburgh (Manson et al., 1994) and the Prnp0/0 Zürich I mice (Büeler et al., 1992) do not. The strategies used to abolish PrP differed in an important respect: in the lines remaining healthy, PrP expression was abrogated either by placing an insert within the PrP coding region (the Edinburgh mice) or by replacing the coding region between codons 3 and 188 by a neo cassette. In contrast, the Nagasaki, Rcm0 and Zürich II lines were generated by deleting not only the ORF, but also 5′ flanking sequences extending into the second intron and 3′ non‐coding sequences. As shown in Figure 1, the three lines have in common the loss of 270 bp upstream of the PrP reading frame and of 450 bp downstream. Whilst the deleted sequences in the Nagasaki mice were replaced by a neo cassette, which, at least in some cases, causes an abnormal phenotype (Fiering et al., 1995), those in the Zürich II mice were replaced by a 34 bp loxP sequence, which is not known to cause deleterious effects.
Although the cerebellar phenotype resulting from extended deletions in the PrP gene can be rescued by a wild‐type PrP allele, it is clearly not caused by the absence of PrP, because the Zürich I and the Edinburgh mice remain healthy despite their lack of PrP (Weissmann, 1996). In the Edinburgh mice no PrP‐specific mRNA was detected, so that significant levels of any fusion protein containing PrP sequences are unlikely. In the Zürich I mice the neo cassette was inserted between the third and the 188th codon of the PrP sequence; although its coding and 3′ non‐coding sequence were in‐frame with the 67 residual PrP codons, there were two termination codons in between, which would preclude read‐through. The relevant DNA segment from the Zürich I mice currently in use was resequenced and the presence of the termination codons confirmed (data not shown). Therefore, the presence of a PrP fragment or a fusion product is not responsible for maintaining the normal phenotype in either of the two lines.
The fact that the knockout lines showing the cerebellar phenotype lack sequences flanking the PrP ORF suggested that critical information was partly or entirely located in these regions. The pathological phenotype could come about either by loss of function, if these flanking regions controlled the formation or encoded part or all of some essential protein or RNA, or by gain of function, if the extended deletion resulted in the production of a deleterious product. The finding that introduction of a wild‐type Prnp allele, either by breeding or as a transgene, abrogated the ataxic phenotype could be accommodated by either explanation: because the Prnp allele contains the flanking sequences, it could supply the missing function conjectured by the loss‐of‐function hypothesis. Alternatively, within the framework of the gain‐of‐function hypothesis, PrP might overcome the pathogenic effect of a postulated deleterious product.
The discovery of Prnd, the gene encoding Dpl, and its expression in the brains of Zürich II, Nagasaki and Rcm0 mice, suggested that the PrP knockout alleles in these animals give rise to a deleterious product. Analysis of brain‐derived cDNAs indicated that in wild‐type mice Dpl mRNA is very weakly expressed, mainly from a promoter upstream of exon 1 of Prnd, whilst the strong expression in Nagasaki and Rcm0 mice is due to chimeric RNAs that originate at the Prnp promoter, run all the way across and past the Prnd ORF and are processed by one or more splicing events that link the 3′ end of the second PrP exon directly or indirectly to the Dpl‐encoding exon (Moore et al., 1999; Li et al., 2000). This intergenic splicing, which was also detected by PCR at very low levels in wild‐type mice, is greatly enhanced in the ataxic mice because the splice acceptor site upstream of the PrP‐encoding exon (Figure 1) is deleted, thus diverting the splice to a downstream acceptor site. Prnd‐specific mRNA was expressed undiminished in brain of Nagasaki mice ‘cured’ by the introduction of a PrP‐expressing transgene (Nishida et al., 1999).
The hypothesis that expression of Dpl in brain is responsible for the ataxic syndrome is supported not only by the fact that in Zürich I knockout mice containing a single Zürich II allele onset of ataxia and Purkinje cell degeneration is retarded, but also by the finding that Dpl expression in the brain at twice the level of that in Zürich II mice, found in mice transgenic for a cosmid devoid of the PrP ORF but containing Prnd, accelerates appearance of the symptoms.
Why should overexpression of Dpl cause ataxia and concurrent overexpression of PrP restore normal function? Shmerling et al. (1998) found that introduction into Zürich I Prnp0/0 mice of an amino‐proximally truncated transgene encoding PrP devoid of the octa repeats and the conserved 112–126 region (PrPΔ32–134) leads to ataxia and degeneration of the cerebellar granule cell layer within weeks of birth. Moreover, introduction of a single wild‐type PrP allele prevented the disease. They proposed that PrP interacts with a ligand to elicit an essential signal and that a conjectured PrP‐like molecule with lower binding affinity can fulfil the same function in the absence of PrP. According to this hypothesis, in PrP knockout mice the truncated PrP could interact with the ligand, displacing the PrP‐like molecule, without, however, eliciting the survival signal. If PrP has the higher affinity for the ligand, it would displace its truncated counterpart and restore function. Because Dpl resembles the truncated PrP, it might cause disease by the same mechanism (Moore et al., 1999; Silverman et al., 2000). However, because the promoter used to express the truncated PrP is active in granule cells but not in Purkinje cells, while the wild‐type PrP promoter directing the ectopic expression of Dpl is active in Purkinje cells (Li et al., 2000), the cellular targets may be different. Indeed, targeting the truncated PrP to Purkinje cells causes ataxia and degeneration of Purkinje cells (E.Flechsig, R.Leimeroth and C.Weissmann, unpublished data).
The wild‐type Prnp allele or a Prnp‐containing cosmid gives rise to PrP expression in Purkinje cells that may counteract the conjectured deleterious effect of Dpl. Interestingly, the Tga20 transgene also has this beneficial effect, although there is no detectable expression in Purkinje cells. Perhaps PrP expressed on the synapses of neighbouring cells can supply the protective effect or can be physically transferred, as shown for other GPI‐linked proteins (Kooyman et al., 1995; Anderson et al., 1996; Brunschwig et al., 1999; McHugh et al., 1999). Interestingly also, the ataxia provoked by expression of truncated PrP in Purkinje cells can be abrogated by the Tga20 allele, perhaps by the same mechanism (E.Flechsig, R.Leimeroth, I.Hegyi and C.Weissmann, unpublished data).
It would thus seem that chronic ablation of PrP per se has only quite modest effects such as alterations in circadian activity rhythms and sleep patterns (Tobler et al., 1996) and demyelination in the peripheral nervous system with old age (Nishida et al., 1999). The hypothesis that expression of Dpl in Purkinje cells causes cell death and ataxia and that the deleterious effects can be counteracted by concomitant expression of PrP is attractive and strongly supported by our results.
Finally, it is worth re‐emphasizing the pitfalls that may beset the interpretation of knockout experiments. In the instances described, deletions of the PrP‐encoding exon gave rise to a severe phenotype that could be reversed by introduction of an intact PrP‐expressing transgene, the classical experiment correlating a phenotype with ablation of a gene. Nonetheless, in this case the conclusion was misleading, because the ataxic phenotype did not result from the deletion of the PrP ORF but from the incidental up‐regulation of a deleterious gene product whose pathogenicity was offset by wild‐type PrP.
Materials and methods
Generation of PrP knockout mice by homologous recombination and Cre‐mediated site‐directed recombination
For the deletion of the PrP ORF‐containing exon and its flanking regions, a targeting vector PrP lox3 (Figure 2A) was constructed that contained a 7.1 kb genomic segment extending from the XhoI cleavage site 1.4 kb upstream of Prnp exon 3 to the BamHI site 3.7 kb downstream of exon 3. For the selection of homologous recombinants, a loxP‐flanked cassette that contained the herpes simplex virus type I thymidine kinase (HSV‐tk) gene (Mansour et al., 1988) and the neomycin resistance gene (neo) controlled by the phosphoglycerol kinase promoter (PGK) (Soriano et al., 1991) was inserted at the NcoI site 0.26 kb upstream of exon 3. A further loxP site was introduced at the XbaI site 600 bp downstream of exon 3. A diphtheria toxin chain A (dipA) cassette for selection against non‐homologous recombinants was placed at the 5′ end (Palmiter et al., 1987). The targeting construct was cloned in pBS‐KS(−).
E14.1 ES cells, a subline of E14TG2a cells (Hooper et al., 1987) derived from 129/OlaHsd mice (Harlan UK Ltd, UK) were obtained from K.Rajewsky and cultured on irradiated mouse embryonic fibroblasts in ES medium (Dulbecco's modified Eagle's medium supplemented with 20% fetal calf serum, 0.1 mM β‐mercaptoethanol, 1 mM sodium pyruvate, 100 U penicillin/ml, 100 μg streptomycin/ml, 0.5 μg leukaemia inhibitory factor (LIF)/ml).
PrP lox3 (10 μg) was linearized with NotI and electroporated into ∼3 × 107 ES cells in 800 μl of phosphate‐buffered saline (PBS) using a Bio‐Rad gene pulser (240 V, 500 μF); G418 (0.4 mg/ml) was added after 24 h. ES cell clones were screened in pools of eight by PCR using primers ΔUp3 (5′‐CAAGGGCCCCTCCTCCTAGAA) and TK (5′‐ACGCGCAGCCTGGTCGAA). Positive clones were confirmed by Southern blot analysis. A homologous recombination event identified by PCR from within the selection marker located in the 5′ region was detected in 82 of 1046 (7.8%) G418‐resistant clones. However, concurrent insertion of the third loxP site (3′ of exon 3) was a rare event found in only five of the 82 clones (6%). To eliminate the loxP‐flanked selection cassette, 10 μg of Cre‐expressing pBS185 (Sauer and Henderson, 1990) were electroporated into 107 ES cells in 800 μl of PBS as above and the cells were plated on to T75 feeder‐cell‐coated flasks. After 72 h the cells were replated at different densities (103–106 cells per plate) on to 90‐mm feeder plates and immediately exposed to 0.2 μM 1‐(2‐deoxy‐2‐fluoro‐β‐d‐arabinofuranosyl)‐5‐iodouracil (Bristol‐Myers Squibb, Stamford, CT). Of the surviving 1–4% of the ES cell colonies, half had retained the loxP‐flanked PrP exon 3 (Prnp lox2 genome configuration) and half had undergone recombination between the first and third loxP site, giving rise to the Prnp lox1 knockout allele (Figure 2A). The Prnp lox2 allele was screened for by PCR using primers P5 (5′‐GCTTTCTTCAAGTCCTTGCTCCTGCTGTAG) and Nco (5′‐TATTGCATGGTTGTTACGCC), and the Prnp lox1 allele with primers P5 and 3′loxP3 (5′‐GGGCAGGAGAAACAGTTTGG). Deletions were confirmed by Southern analysis. Positive ES cell clones with normal karyotypes were injected into C57BL/6 blastocysts.
One of the Prnp lox2‐containing clones gave rise to a highly chimeric animal that produced an F1 generation carrying either the Prnp lox1 or the Prnp lox2 allele in addition to the wild‐type allele (Figure 2C), indicating that the founder was triple chimeric for the two mutated and the wild‐type allele in its germline. Both lines were bred to homozygosity; both mutant alleles showed Mendelian inheritance.
Mice transgenic for the transactivator cosmid Cos‐tTA
Cosmid cos6.I/LnJ‐4 (Westaway et al., 1994) comprises the Prnp and the Prnd locus. The Cos‐tTA vector is derived from it by replacing the PrP ORF with the tetracycline‐controlled transactivator (tTA) ORF (Gossen and Bujard, 1992). A 0.9 kb PCR product of the tTA ORF was prepared using pUD‐Combi (kindly provided by H.Bluethmann) as template and the primers ATG (5′‐CATGCCATGGCTAGATTAGATAAAAGTA) (NcoI site italicized) and MfeI (5′‐GCGCAATTGTTACTTGTCATCGTCGTCC) (MfeI site italicized). A fragment extending from the XbaI site within PrP intron 2 to the initiator ATG of the PrP ORF was prepared by PCR with the template NZWBamHI (Büeler et al., 1992) and the primers XbaI (5′‐GTGAGCCTGCCAAAACTCTAGATGTTTCCTG) (XbaI site italicized) and AflIII (5′‐TTCGACATGTTGACTGATCTGC) (AflIII site italicized). The amplicon harbouring the tTA ORF was joined by its 5′‐NcoI end to the NcoI‐compatible 3′‐AflIII end of the PrP amplicon and cloned into XbaI‐ and SmaI‐restricted pTM1 (Moss et al., 1990) in a three‐way ligation The NcoI–MfeI fragment of this intermediate plasmid, comprising 260 bp of intron 2, 10 bp of 5′ non‐coding region and the entire tTA ORF, was introduced into NcoI–MfeI‐cleaved NZWBamHI, replacing the PrP ORF and 80 nucleotides of 3′ non‐coding region by the tTA ORF, to yield NZW‐tTA. The 3.7 kb BamHI–MfeI fragment of NZW‐tTA was inserted in a three‐way ligation into an intermediate construct containing the 18 kb PmeI–SalI fragment from the PrP cosmid cos6.I/LnJ‐4. The tTA‐containing PmeI–SfiI fragment from this construct was ligated in a two‐way ligation to the PmeI–SfiI backbone fragment of the original cosmid to yield the final vector Cos‐tTA (Figure 2B). Cos‐tTA was digested by NotI and electrophoresed through a 0.5% agarose gel (Seakem GTG; FMC Bioproducts, Denmark). The tTA/Prnp insert was electro‐eluted using BIOTRAP (Schleicher & Schuell), ethanol precipitated, resuspended in injection buffer (10 mM Tris–HCl pH 7.5, 5 mM NaCl, 0.1 mM EDTA) and filtered through a Micropore filter (Millipore Ultrafree MC 0.45 μm).
Injections into pronuclei of fertilized oocytes from Prnp0/0 Zürich I mice were performed by conventional methods. Founders were identified by Southern blot analysis of tail DNA and crossed with Prnp−/− Zürich II animals. Offspring positive for the cosmid were identified by PCR using primers TA‐2 (5′‐GCACCATACTCACTTTTGCCCTTTAGA) and 3′Mfe (5′‐CAGGGGTATTAGCCTATGGGGGACACAG) (Figure 2B) and mated with Prnp−/− Zürich II mice to eliminate the Prnp0/0 Zürich I allele.
Southern blot analysis
About 10 μg of genomic DNA were cleaved with appropriate restriction enzymes, electrophoresed through 1% agarose gels, depurinated and transferred to Hybond N+ membranes. After prehybridization, hybridization and washing (Church and Gilbert, 1984), membranes were exposed to X‐ray film or scanned with a PhosphorImager (Molecular Dynamics, USA).
The various Prnp alleles were detected with an α‐32P‐labelled 650 bp XbaI–NcoI fragment (X/N probe, Figure 2A) covering a segment of intron 2 upstream of the loxP site. XbaI digestion produces fragments of 3.5 kb for Prnp+, 4.0 kb for Prnp0 (=Prnp0 Zürich I), 2.3 kb for Prnp lox1 (=Prnp− Zürich II) and 5.2 kb for Prnp lox2. The Cos‐tTA transgene was detected in XbaI‐digested genomic DNA, using as probe the tTA ORF (cleaved out of pUD‐Combi, a gift from H.Bluethmann, with NcoI and BamHI) or a 0.9 kb PCR product of the tTA ORF used for the cloning of Cos‐tTA (Figure 2B).
Northern blot analysis
Total RNA was prepared using the RNeasy kit (Qiagen) following the manufacturer's instructions. Poly(A)+ mRNA was purified using the Oligotex mRNA kit (Qiagen). RNA was electrophoresed through a 1% agarose gel and transferred to Hybond‐N+ (Amersham). Hybridizations were carried out according to standard techniques (Sambrook et al., 1989) using an α‐32P‐labelled probe (1–2 × 106 c.p.m./ml). Probe Δ was used for PrP mRNA (Büeler et al., 1992) (Figure 2A), cloned Dpl ORF for Dpl mRNA (Figure 2B) and the tTA ORF fragment described above for tTA transcripts. Filters were stripped by boiling in 0.5% SDS, 20 mM Tris–HCl pH 7.5 for 10 min, tested for residual signal and rehybridized with a glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) probe (Fort et al., 1985).
Western blot analysis
Brain homogenates (10% w/v) were prepared in PBS containing 0.5% Nonidet P40, 0.5% sodium deoxycholate by passing brains successively through 18‐ and 22‐gauge needles. Homogenates were centrifuged at 1500 g for 10 min and supernatants were adjusted to 8 mg/ml total protein. Samples (60 μg for PrP and 100 μg or as indicated for Dpl) were electrophoresed through 16% SDS–polyacrylamide gels (NOVEX™, San Diego, CA) and transferred to PVDF membranes (Immobilon‐P, Millipore, USA). PrP was detected with monoclonal antibody 6H4, which recognizes residues 143–151 of PrP (Korth et al., 1997) (1:10.000, Prionics AG, Switzerland). Dpl was probed with polyclonal anti‐Dpl rabbit antiserum 2234 (1:5000; raised against recombinant Dpl), pre‐adsorbed for 1 h at 37°C with 10% wild‐type brain homogenate. Blots were incubated with horseradish peroxidase‐conjugated anti‐mouse IgG1 antibodies (1:5000, Zymed, San Francisco, CA) for PrP or with horseradish peroxidase‐conjugated swine anti‐rabbit immunoglobulins (1:5000, Dako, Glostrup, Denmark) for Dpl. Blots were developed using the enhanced chemiluminescence kit SuperSignal West Pico (Pierce, Rockford, IL) and exposed to BIOMAX MR‐1 film (Kodak, Rochester, MN). Appropriate film exposures were scanned with a laser densitometer (Molecular Dynamics).
For immunohistochemical staining of PrP, sections were heated three times for 5 min in a microwave oven in 10 mM sodium citrate pH 6, and incubated with polyclonal anti‐PrP rabbit antiserum R340 (1:50; Brandner et al., 1996). Visualization was with the Tyramide Signal Amplification kit (NEN, Life Science, Boston, MA)
Histopathology and morphometric analysis
Mice were deeply anaesthetized and preterminally perfused with 1 ml of 1 mM EDTA in PBS and subsequently with 3% paraformaldehyde/1% glutaraldehyde in PBS. Brains were paraffin‐embedded and cut into 2 μm sections. Sections were stained with haematoxylin–eosin (HE) and for glial fibrillary acidic protein (GFAP; polyclonal antibody, 1:300 from Dako). Biotinylated secondary antibodies (goat anti‐rabbit; Dako) were used at a dilution of 1:200 and visualization was with avidin–biotin complex (Dako) according to the manufacturer's instructions. For quantification of Purkinje cells in the cerebellum, sections were stained with an anti‐calbindin monoclonal antibody (1:5000; Sigma Chemical Company, St Louis, MO) and visualized by a biotinylated rabbit anti‐mouse antibody (1:200; Dako) and avidin–biotin complex. Purkinje cells were counted at 200× magnification using a counting grid with square side length 100 μm. Ten adjoining squares were counted in each area indicated.
We thank Doron Shmerling for providing the precursor construct for PrP lox3, Axel Behrens for Dpl antiserum, Antonio Servadio for brains of Prnp‐tTA/F959 mice, Horst Bluethman for pUD‐Combi, Joseph Ecsoedi for synthetic oligonucleotides, Norbert Wey and Ray Young for photographic artwork, Marek Fischer for valuable advice and Gerlinde Stark for technical support. This study was supported by the Canton of Zürich, by grants from the Schweizerischer Nationalfonds to A.A. and C.W., from the Medical Research Council to John Collinge and C.W., from the European Union to A.A., from the Human Frontier Science Program to C.W. and from the Helmut Horten Foundation to A.C.
- Copyright © 2001 European Molecular Biology Organization