Variations in the PrP gene sequences of apes
DNA from six species of apes representing all four genera was amplified by PCR. As shown in Figure 1, of the five orang-utan specimens, two had one less octarepeat structure on one of the alleles. The two orang-utans with one less octarepeat seem to be unrelated, as one was born in captivity and the other in the wild. Nucleotide sequencing of the entire ORFs of the six ape species resulted in similarities in the nucleotide and amino acid level of 97.8 to 99.7% and 98 to 99.6%, respectively, when compared with the human PrP (Table 1, see Figure 2). The gorilla PrP gene was most similar to the human gene, followed by the chimpanzee and gibbon genes. While the gibbon and siamang species (Hylobates lar and H. syndactylus) exhibited much more variation on the nucleotide level (15 substitutions versus six for the chimpanzee), these 15 nucleotide changes resulted in only two amino acid substitutions. The orang-utan PrP gene is most distant from the human gene with 17 nucleotide exchanges and five amino acid substitutions. Allele-specific sequencing of the two variant orang-utan PrP alleles (Figure 1) was performed after excising the lower band from agarose gels; sequencing of these bands revealed that the deletion of one octarepeat had occurred at octarepeat 2 or 3. This difference from the classical phylogeny was confirmed by sequence-based phylogenetic trees. The orang-utans are more distant than the gibbon from the human, even when the octarepeat region was excluded from the analysis.
The PrP genes of 12 species of old-world monkeys, representing five genera, were sequenced (Table 1). The degree of amino acid identity compared with the human PrP varied between 92.9 and 97.2%. The three genera of Papio (baboon and mandrill), Macaca (five species) and Cercopithecus (African green monkey; three species) were almost identical at the amino acid level with only minor differences on the nucleotide level (Table 1 and Figure 2). This contrasts with results from other genes where greater variability was found between and within these species (Goodman et al., 1990; Melnick et al., 1993). The genus of African green monkey contained the deletion of one octarepeat structure on both alleles (Figure 1). This deletion occurred between the third and fifth octarepeat. Almost all variations found in this group (except the missing octarepeat and positions 220 and 234) occurred between amino acid positions 97 and 168 (Figure 2) and are located almost exclusively outside the four predicted a-helical regions of PrP C (named H-1 to H-4: Figure 3). Characteristic variations were seen at positions 97 (S : N), 100 (N : H) and 108 (N : S). The substitution at codon 100 is unique in that other residues are present in new-world monkeys, domestic animals and rodent species. There are no Figure 1. PCR analysis of PrP genes from non-human primate species with variations in the octarepeat region. The internal primers HM-1 and HM-20 were used to emphasize the variations. On top the species are indicated, on the left a molecular size marker is given (100 base-pair ladder). Two out of 5 studied orang-utans (OU 1 to 5) revealed a missing octarepeat on one allele. AGM, spider and squirrel monkeys showed the deletion or insertion of 1 octarepeat on both alleles. Colobus monkey and gibbon ape represent normal PrP alleles. amino acid variations in the N-terminal segment of the protein, codons 1 to 96 (Figures 2 and 3). Variations in the PrP gene sequences of new-world monkeys
For the seven new-world monkey species studied, 34 to 39 nucleotide exchanges were found resulting in seven to ten amino acid substitutions compared to humans (Table 1). Within this group, the PrP gene of the spider monkey showed the greatest sequence identity with human PrP when the deletions were not considered. It is noteworthy that fewer exchanges were present around the first putative helical region than in old-world monkeys, with the important exception of position 112 (M : V). Group specific variations were seen at codon 170, where the marmoset, tamarin and spider monkeys have an S : N substitution that is identical with that found in three subspecies of hamsters. At codon 182, the squirrel monkey has an I : V substitution that lies within the third putative a-helical region; at codon 205, the Aotes monkey has an M : I substitution that lies within the fourth putative a-helical region and is the same as that found in sheep.
Variations in octarepeat structures
Variations in the number of octarepeats are of considerable interest since addition of two, four, five, six, seven, eight or nine octarepeats in humans results in an inherited prion disease (Goldfarb et al., 1991; Owen et al., 1989, 1992; Palmer et al., 1993; Poulter et al., 1992). One additional octarepeat has been reported in some cattle but it does not seem to alter the susceptibility of these animals to BSE (Goldmann et al., 1991; Prusiner et al., 1993). We found an additional octarepeat for squirrel monkeys; it is present on both alleles and is located between the second and third octarepeat. Deletion of one octarepeat in humans occurs with a frequency of about 0.5% and is not associated with disease (Collinge et al., 1989; Goldfarb et al., 1991; Melnick et al., 1993; Owen et al., 1991; Puckett et al., 1991; Vnencak-Jones & Phillips, 1992). For the genera of African green monkey and spider monkeys we detected the identical deletion of one octarepeat located between the third and fifth octarepeat; this deletion occurs on both alleles. From two of five analyzed orang-utans there was a deletion for one octarepeat on one allele (Figure 1), which is located between the first and fourth octarepeat.
Clustering of non-human primate variations in the PrP gene
Twenty-seven variant positions were detected when the translated sequences of the PrP genes of 25 non-human primates were compared with the human gene (Figures 2 and 3): 24 of these were single amino acid substitutions. In general, these exchanged residues represent conservative substi-tutions. Of note, the N-terminal signal peptide, the Asn-linked consensus sites for glycosylation, the two Cys residues that form a disulfide bond and the GPI anchor attachment site are invariant among the primates. Most of the variations among primates are located outside the four putative a-helical regions (Figure 3); the frequency with which amino acid substitutions occur is plotted on the ordinate. The residues that vary among the PrP genes of primates differ from the mutations that are linked to or segregate with inherited prion diseases (Figure 4A; and see Palmer & Collinge, 1993; Prusiner &DeArmond, 1994). These variations in the primate PrP genes also do not occur at sites known to be polymorphic in humans, i.e. codons 129 and 219 (Kitamoto & Tateishi, 1994; Owen et al., 1990; Palmer et al., 1991). Methionine was found at codon 129 in all 78 of the non-human primate samples.
Variations in primate PrP genes were compared with the deduced amino acid sequences of PrP in ungulates and other domestic animals (Figure 4B). Ala133 is an invariant residue in primates. This corresponds to codon 136 in sheep, where an A : V polymorphism has been identified (Goldmann et al., 1990). In some breeds of sheep, susceptibility to scrapie seems to be governed by the amino acid encoded at this polymorphic site while in others it does not (Laplanche et al., 1993a,b). At codon 168, Glu is encoded in humans but Gln is found in all other primates. In sheep, the corresponding codon is 171 where the most frequently encoded amino acid is Gln; however this residue is polymorphic in sheep and Arg is often found in place of Gln. In Suffolk sheep, homozygosity for Gln at codon 171 was found to render these animals susceptible to scrapie (West-away et al., 1994). Amino acid substitutions at codons 108 and 189 of the murine PrP gene (Prn-p) have been found to be genetically linked to control of the scrapie incubation time (Figure 4C; and see Carlson et al., 1986; Westaway et al., 1987). The murine PrP allele designated a encodes Phe at residue 108 and Thr at 189, while the b allele encodes Ser at 108 and Val at 189. Among all the primates, at the corresponding positions Met (position 109) and Thr (position 190) were found. PrP Genes in Primates 368
Phylogenetic relationships deduced from PrP sequences
The sequence similarities for the ORF of the PrP gene within non-human primates differ from the previously reported molecular phylogeny of pri-mates (Brownet al., 1982; Goodman et al., 1989, 1990). Molecules exhibit (1) a high degree of sequence similarity between humans and the apes (Homi-noidae: Brown et al., 1982; Goodman et al., 1989; Ruvolo et al., 1991), (2) a substantial variability within the old-world monkeys (Melnick et al., 1993) and (3) that new-world monkeys are most distant from humans. The variations in the PrP molecules of primates violate these expectations. We compared phylogenetic trees generated by nucleotide analysis of the PrP genes with trees for mtDNA and b-globin genes (Figure 5 and see Brown et al., 1982; Goodman et al., 1990). The classical branching pattern within Hominoidea, represented in Figure 5B, is gibbon (Hylobatidae), orang-utan (Ponginae), and Homininae (human-chimpanzee-gorilla clade), with the chim-panzee as the closest relative to humans (Brownet al., 1982; Goodman et al., 1989, 1990). In our studies, gorilla PrP was most closely related to human PrP while chimpanzee and gibbon PrPs were next; whereas, orang-utan was the most divergent of the ape PrP molecules. To our knowledge, this phylogenetic pattern has not been reported for any other protein. The old-world monkeys showed an unexpectedly high degree of sequence similarity with many species possessing identical PrP
Distribution of variant amino acids
Approximately one third of the residues in PrP were found to vary among mammals (Table 2). The distribution of variant amino acids was examined for different regions of PrP. Within the primates, most of the variation in the PrP sequence (18%) is within the N-terminal region (codons 1 to 90) while the variation in the C-terminal region (codons 221 to 253) was considerably lower (3%). In contrast, variations in the N and C-terminal regions for all species were 38% and 36%, respectively. The variation in the central region of PrP (codons 91 to 220) was 12% for residues in primates and 22% for all species. Combining the data for all species, 52% of the residues varied within the N-terminal (codons 1 to 22) and the C-terminal (codons 232 to 253) peptides that are cleaved off as mature PrP C is formed. The same peptides in primates varied at only 7% of the residues.
Transitions and replacement substitutions
The nucleotide exchanges in the PrP gene account for about 85% of the transitions; this is lower than the number of transitions (92%) reported for mtDNA (Brown et al., 1982), but higher than the 70% found for b-globin cluster genes (Goodman et al., 1989). The ratio between silent and replacement substitutions was about 3:1. Within the apes, this ratio varied from 2:1 to 7.5:1. Within old-world monkeys, the ratio ranged from 3.2:1 to 3.9:1; for new-world monkeys, it varied from 3.8:1 to 4.9:1. These ratios are low.