Structure and polymorphism of the mouse prion protein gene

Proc. Natl. Acad. Sci. USA Vol. 91, pp. 6418-6422, July 1994

Genetics

Structure and polymorphism of the mouse prion protein gene (scrape Incubation time/physkcal m /yeast artical ch

does/intron detn/prin protn gene pro

)

DAVID WESTAWAY*, CAROL COOPER*, SHERRY TURNERt, MARIA DA COSTA*, GEORGE A. CARLSONt,

AND STANLEY B. PRUSINER*t§

Departments of *Neurology and of *Biochemistry and Biophysics, University of California, San Francisco, CA 94143; and tMcLaughlin Research Institute, Great Falls, MT 59405

Contributed by Stanley B. Prusiner, December 28, 1993

lines expressing a Prn-pb cosmid from I/LnJ mice had abbreviated scrapie incubation times of 75-115 days (14, 15). Although results obtained with Tg(Prn-pb) mice are compatible with a pseudodominant model for the action of Prn-pb (15), we cannot formally exclude a Prn-i/Sinc gene distinct from Prn-p. To define Prn-p-linked genes and to facilitate transgenic modeling of prion diseases, we have therefore constructed a physical map of the Prn-p chromosomal region.

Missense mutations ln the prion protein (PrP) ABSTRACT gene, overexpression Oldte ceilular isoform of PrP (PrPC), and infection with prions coining the scrapie isoform of PrP (NrP&) ad cause neurodegenerative disease. To understand better the physiology and expression of PC, we retrieved moue PrP gene (Prn-p) yeast artificial chromosome (YAC), cosmid, phage, and cDNA clones. Physic mapping positions Prm-p %3O@ kb from ecotropic virus integration site number 4 (Evi-4), compatible with failure to detect recombination between Prn-p and Evi-4 in genetic crosses. The Prn-pa allele encompasses three exons, with exons 1 and 2 eng the mRNA S' utnsted region. Exon 2 has no equivalent in the Syrian hamster and human PrP gene. The Ppb gene shae this intron/exon structure but harbors an =6-kb deletion within intron 2. While the Prn-pb open reading fame encodes two ano acid subt ns linked to prolonged scrapie incubation periods, a don f intron 2 sequences aiso charactries inbred strains such as Rf/S and MOLF/Ei with shorter incubation periods, making a relationship between intron 2 size and scrapie pathogenesis unlikely. The promoter regions of a and bP n-p alleles include consensus Spl and AP-1 sites, as well as other conserved motifs which may represent binding sites for as yet unidentified transcription factors.

MATERIALS AND METHODS Genomic Cloning. Recovery and characterization of yeast

artiflcial chromosomes (YACs) containing Prn-p. DNA pools

from a C57BL/6J library cloned in pYAC4 (16) for screening were made available through the courtesy of Shirley Tilghman and David Koos (Princeton University). PCR was used to screen for the presence of Prn-p exon 3, progressing through complex pools (1920 clones), simple pools (96 clones), and finally the rows and columns of a 96-well plate (17). Chromosome-sized yeast DNAs from positive clones were embedded in low-melting temperature agarose, digested with restriction endonucleases, and analyzed by pulsed-field gel electrophoresis (18, 19). After electrophoresis, DNA in the gel was fragmented by UV exposure and alkaline transferred to nylon filters as recommended by the manufacturer (Bio-Rad) and hybridized with radiolabeled probes: left and right YAC arm probes (2.6- and 1.7-kb fragments from a BamHI and Pvu II double digest of pBR322 and five Prn-pspecific fragments). These were, from 5' to 3', a 1.7-kb BamHI/EcoRI fragment excised from the cos6.I/LnJ-4 clone (5' flank), a 2.2-kb intron 1 fragment generated by amplifying cos6.I/LnJ-4 DNA with exon 1 and 2 primers, an exon 2-specific probe (20) previously referred to as 5'UT, a 2-kb insert from a hamster PrP gene cDNA clone as a probe for the open reading frame (ORF) (21), and a 7.6-kb 3' flanking region BamHI/BamHI fiagment excised from cos6.I/LnJ-4 (3' flank; see Fig. 1). A fragment PCR amplified from a DNA clone specific for Evi-4 (H. G. Bedigian, The Jackson Laboratory) was also tested for inclusion within the YACs. Phage cloning. High molecular weight genomic DNA was prepared from the brain of a 129Sv mouse (kindly provided by J. Latimer and R. Pedersen, University of California, San Francisco). This DNA, partially digested with Mbo I, was filled-in with dA and dG using the Klenow fragment of DNA polymerase I and then ligated with T4 DNA ligase (IBI) to Xho I-digested dC and dT filled-in bacteriophage A vector FIX II. In vitro packaging was as recommended by the manufacturer (Stratagene) and recombinant phage were

The cellular isoform of the prion protein, PrPC, is a glycolipidated sialoglycoprotein. Although its physiological function remains obscure, derivatives of this molecule play a central role in the pathogenesis of prion diseases in humans and animals (1). A posttranslational variant of PrPC, PrPSc, is the major and possibly sole component of infectious pathogens, prions, which cause scrapie, iatrogenic Creutzfeld-Jakob disease (CJD), and other central nervous system disorders (2). Mutations in the human gene encoding PrPc (PRNP) segregate with disease loci on chromosome 20 specifying Gerstmann-Straussler-Scheinker disease and familial CJD (3). More recently, elevated PrP gene dosage and expression has been associated with a lethal neuromyopathy (4). Studies of experimental scrapie in mice have played a key role in deciphering the structure and replication of prions. The scrapie resistance of PrP gene-ablated mice (Prn-pOfo) provides additional evidence (5, 6) in support of the proposal that PrPsc is an essential and major component of the infectious prion particle (1, 7, 8). In other experiments, crosses between New Zealand White (NZW) and I/LnJ mice revealed that a scrapie incubation time gene, Prn-i, was genetically linked to Prn-p (9, 10). I/LnJ mice carry the b haplotype of Prn-p, which encodes PrP-B molecules that differfrom those in NZW (Prn-pa) mice at codons 108 and 189 (11), strongly suggesting that Prn-p is synonymous with Prn-i and the scrapie incubation time gene Sinc defined earlier in VM mice (12, 13). Unexpectedly, transgenic (Tg) (Prn-pb)

Abbreviations: PrP, prion protein; PrPC, cellular isoform of prion protein; ORF, open reading frame; YAC, yeast artificial chromosome. ITo whom reprint requests should be addressed at: Department of Neurology, HSE-781, University of California, San Francisco, CA 94143.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

6418

Genetics: Westaway et al.

Proc. Nati. Acad. Sci. USA 91 (1994)

plated on the Escherichia coli strain NM621, transferred to nylon filters (Micron Separations, Westborough, MA) and screened with Prn-p probes by standard procedures. Amplification of Prm-p cDNAs. Single-sided specificity amplification. Aliquots of I/LnJ and NZW mouse brain cDNA were tailed with dG residues, diluted, and amplified using Anchor and Anchor-C primers (22) and two nested Prn-p primers (TCAGAACTGAACCATTTCAACCGAGCTGAAGCATTCTGCC and GCTTCTTGCAGAGGCCAACATCA). After subcloning, five colonies corresponding to NZW- and I/LnJ-derived inserts were sequenced. DNA Sequencing. Sequences were determined using Sequenase version 2.0 (United States Biochemical) and deazanucleotides. Bluescript subclones were sequenced in doublestranded form using vector- and insert-specific primers. For some problematic regions, Taq polymerase was used instead of Sequenase. Supercoiled double-stranded plasmid templates were sequenced using methods described by Hattori and Sakaki (23) and Chen and Seeburg (24).

RESULTS AND DISCUSSION A Physical Map of the Prn-p Region. PCR-based screening of a C57BL/6J mouse YAC library (16) yielded two clones containing the ORF of Prn-pa. Complete and partial digestions with rare-cutting restriction endonucleases and pulsedfield gel electrophoresis were used to analyze the YACs. Four probes within a 40-kb interval including Prn-p (Fig. 1C), as well as probes for the left or right vector arms, were used to orient Prn-p within the YACs and to define a region of overlap of -70 kb between PYB6/2.GC1 (Cl, lower) and PYB6/2.GC2 (C2, upper; Fig. 1A). Not I and Pme I sites within C1 (='330 kb) and C2 (=310 kb) are shown. The Prn-p gene was located very close to the left arm in Cl; the 5' flanking marker was not included within this YAC (Fig. A

r

4cY2

LN%

60

R BE

------ --'2.GC2 PYR6

4O

PY86:2.GC1

_j

11

P

.: tI

I-r?.

o

P-0-it-n-

F

.!.

B

Nc

a n,

Po-r>

I!,!,'

,:

n .7

C

2

Pine

cos6.1

:c;s6.1 LnJ-'3 r.l-4

Fio. 1. Prn-p molecular clones. (A) Two overlapping YAC clones from a Prn-pa (C57BL6) mouse. Vector sequences are represented by stippling; L and R indicate left and right vector arms. Open boxes (not to scale) indicate genes. Pme, Pme I; Not, Not I. Fragment sizes are listed in kb. (B) Contig of four phage clones spanning the Prn-pa gene. Italicized numbers indicate exons. (C) Two cosmid clones spanning the Prn-pb gene. Scale is the same as in B. Coordinates of probe fragments used to map the YAC clones are indicated by shaded boxes ("probes").

6419

1C); exons 1 and 2 and the Pme I site within the promoter region were present. A probe for the Evi-4 hybridized to Cl; this is the closest Prn-p-linked marker identified to date. No Evi4-Prn-p recombinants were observed in two large test crosses (10, 25), providing strong evidence that C1 was derived entirely from a contiguous segment of chromosome 2. Sequences upstream from Prn-p on C2 have not yet been assigned to chromosome 2 and it is formally possible that this clone is a chimera. A Not I site and a Sal I site were located within the Prn-pa gene in both YAC clones. Interestingly, both restriction sites were known to be absent from the corresponding region of the Prn-pb cosmid I/LnJ-4 (14). Analysis offour overlapping Prn-pa phage clones confirmed the existence of these polymorphisms and permitted more refined physical mapping (Fig. 1 B and C; see also Fig. 3). As the longest Prn-p cDNA clone described previously is shorter than that predicted by primer-extension analyses (11, 26), we also isolated fulllength 5' cDNAs using a single-sided specificity PCR strategy (22). Sequence data from these clones in turn allowed the positioning of intron-exon boundaries with respect to the restriction maps. These data are collated in Figs. 1 and 2. Pn-p Gene Sructure and RNAs. Unlike the human and Syrian hamster PrP genes, the 5' untranslated region of the Prn-p mRNA is encoded by two exons. Thus, the uninterrupted Prn-p coding region lies within exon 3, not exon 2 as described previously (11). No distinguishing polymorphisms were found between exons 1 and 2 of the a and b alleles. While the function of exon 2 is unclear we note that the 5' end of this 98-bp sequence is more conserved with respect to the sheep homologue (underlined nucleotides in Fig. 2B). Multiple stop codons, in-phase with the PrP ORF, are a feature of the mRNA 5' untranslated regions of both mouse and sheep genes (27). Splice donor and acceptor sites flanking exon 2 were found to differ from a consensus by 3/13 and 2/8 mismatches, respectively (28), suggesting that splicing to include these sequences might not be obligatory. However, in PCR analyses of neonatal or adult mouse brain cDNA, amplified DNA fragments corresponding to exon 2 skipped mRNAs could not be detected (data not shown). Thus, most or all Prn-p mRNAs include the exon 2 sequence described here. To clarify the significance of a long ORF present on the antisense strand of most mammalian PrP genes (29), we sought corresponding mRNAs. In Northern blot experiments, a control sense-strand-specific exon 3 RNA probe detected an =2.4-kb Prn-p mRNA as well as a less abundant species with an apparent size of 3.9 kb. Analyses with an intron 1-specific probe indicates that the latter RNA represents a partially spliced mRNA precursor. Under analogous conditions an antisense strand probe yielded a weak, diffuse hybridization signal corresponding to RNAs ranging from 7.5 to 1.0kb (data not shown). Our data are compatible with other analyses (11, 30-32) indicating that RNAs encoding the antisense ORF are rare or absent from the central nervous system. While Moser et al. (33) detected a 4.5-kb antisense PrP NA in mouse and hamster brain, this transcript was also present in Prn-po/o mice. We suggest that these antisense PrP RNAs are not derived from the Prn-p locus but represent artefactual cross-hybridization, possibly mediated by G+C-rich sequences within the N terminus of the PrP ORF (34). The Size of Pra-p Inton 2 Varies Between Mouse Stran. Detailed mapping of molecular clones of a and b Prn-p alleles revealed a substantial difference in the size of intron 2 in a region encompassing the aforementioned Not I and Sal I sites. To exclude the possibility of cloning artefacts, an intron 2-derived hybridization probe was used to challenge BanHIdigested genomic DNA (Fig. 3 Left). DNA from 129Sv and NZW mice of the Prn-pa haplotype revealed a 6.7-kb frag-

6420

Genetics: Westaway et al.

A

Proc. Natl. Acad. Sci. USA 91 (1994)

20

40

60 *

a

100

80 a

120

140

GATCAGTCTTCCATTAAAGATGATTTTTATAGTTGCTGAGCGTCGTCAGGGAGTGCTGACACTGGGGGCG Motif 3

Motif 4 180

160

200

* GTTTAAACAGATACAAGCATTTAAGCCAGTCCGGAGCGGTfALPATTCCCCCCACCCCCCACCCCCCC G *

*

a

*

CTTCCTCGCTCCGTTAGGTAACCTTTCATTTTCTCAACTACCCATTATGTAACGGGAGCATTGGGTACTG Motif 1 Motif 2

AP-1

280 * CGAGAGACGCGGCGCGGCCATTGGTGAGCATCACGCCCC CCCCTCGCCCAGCCTAGCTCCCGCCTGCC qc 220

a

240

260

Spl

300

320

340

* PC A

Spl

A

380

360 a

AA

A

A

420 *

400

GGTGGCAGGTAAGCGGGCTGCTGAAGCCAGGCCTTGGCGAGCACTCAGCCTTCCGTCGTCAAGCTCGGC :T exon 1->