The principle of kinship polymorphisms can be applied to human prion protein to predict non-CJD causing polymorphisms. The idea here is that normal prion sequences in phylogenetically adjacent species provide candidates for neutral changes in the target species. The variations are presumed nearly-neutral because they became established in a host population from an earlier consensus sequence via Kimura drift across Hardy-Weinberg equilibrium heterozygosity. Differing auxillary genes can weaken applicability.
As more CJD control genes are sequenced, these variations are predicted to be preferentially observed. A few human alleles are that don't seem to cause CJD in and of themselves but may modulate disease course or a species barrier. The analysis assumes that nascent pathology is not masked by the shorter life spans in other primates.
Beginning at the N-terminus using gorilla, orangutan, human, and chimp, we find six candidates for neutral polymorphisms. The table below shows these relative to known structural features of prion protein: one occurs in the signal peptide, two flank helix H1, and three occupy the variable loop between beta stand B2 and helix H3. (No allowance has been made for coordinated changes attributable to residue-residue interaction in the 3D structure.) With less certainty, this analysis could be extended to old world monkeys, and a similar analysis could be performed centered on cow and artiodactyl nearest-neighbors.
Consensus Change Position Domain DNA change Species affected glu asn 18 signal 1: G to A* orangutan ser asn 143 flanks H1 2: G to A orangutan his tyr 155 flanks H1 1: C to U orangutan met val 166 post S2 loop 1: A to G orangutan gln glu 168 post S2 loop 1: C to A human asn ser 171 post S2 loop 2: A to G chimp*Inferred from genetic code and assumed single base change. Here position 1 of glu codon changed from G to A. The mouse structure of Wuthrich et al is used to infer primate domains. The proposition states that a human with a serine at position 171, instead of a serine would not develop CJD, etc.
Context of nearly-neutral variants within human prion domain human amino acid sequence signal peptide MANLGCWMLVLFVATWSDLGLC / KKRPK pre-repeat PGGWNTGGSRYPGQGSPGGNRYP repeat region PQGGGGWGQ.PHGGGWGQ.PHGGGWGQ.PHGGGWGQ.PHGGGWGQ invariant core GGGTHSQWNKPSKPKTNMKHMAGAAAAGAVVGGLGG B1 Met129 H1 B2 YMLGSAMSRPIIHFGSDYEDRYYRENMHRYPNQVYYR Loop H2 Cys 1 Asn 1 PMDEYSNQNNFVHDCVNITIKQHTVTTTT Asn 2 H3 Cys 2 KGENFTETDVKMMERVVEQMCITQYERESQAYYQ GPI peptide RGS / SMVLFSSPPVILLISFLIFLIVG
Changes in DNA sequence need be analyzed separately from protein sequence, in a context of conserved hairpins, hot spots, replication fork slippage, and lost-and-found Okazaki fragments. Mutation in the prion gene seems driven by structural issues in prion DNA: at sites where contemproray mutations are occuring, we find the same substitutions, despite the genetic code offering alternative conservative single base change options and empirical transition/transversion expectations in humans.
While we now have three "hot spots" along the prion gene for CJD-causing mutations, this does not necessarily mean that the amino acids at these positions play exquisitely sensitive roles in the 3D structure or in triggering pathology. Enhanced occurrence could simply be a by-product of special chemistry and conformation within the sequence. A previously compiled list of deleterious mutations in human prion gene collates explanations for hot-spot frequency and aligns CJD-causing mutations in humans to seemingly harmless polymorphisms in sheep and cow.
How many familial CJD cases will it take to attain 'saturation' of the genetic map? That is, what changes in what codons can pathologically affect prion structure-function. This is hard to answer from screens. More rapid progress, at some cost in interpretability, could result from site-specific, directed mutagenesis of a chimeric transgenic mouse. By constructing specific point mutations by the thousands (as done for lac repressor) and inserting them individually back into the host, we could probe systematically which changes cause TSE, as a probe of prion domain structure-function.
Worldwide re-occurrence is seen of the _same_ lethal FFI first codon position GAC to AAC transition resulting in asp to asn change at codon 178, just before the second alpha helix [Neurology 47:1313 Nov1996]. This same change has now been seen in lineages from Italy, Ireland, and Japan, mainly in a met/met background at codon 129. Multiple lineages are also known at codon 102, where spontaneous deamidation of a methylated C may account for the high frequency of localized mutation.
The high frequency of insertions and deletions in the octapeptide repeat region may induce CJD by a different mechanism. Extra octapeptide repeats don't necessarily act through perturbation of prion protein structure; indeed, the region is cleaved off in chickens. The repeats may function primarily in mRNA; CJD resulting from a dosage effect (increased translational efficiency) or improper in situ localization of synthesis. In other neurological repeat disorders repeat amplification is attributed to a looped-out hairpin on the lagging strand and replication read-through (Mol Cell Bio 16:6619 1996).
Where have changes occurred in prions over phylogenetic time? The graphic below consolidates sequences from 52 species and displays changes relative to proposed prion protein domains. The results are displayed quantitatively in four ways: (1) by simply counting the number of distinct amino acid substitutions at a given position, (2) by counting the number of species using each substitution, (3) [soon] by empirical liklihood of substitution in the sense of Dayhoff, and (4) [soon] by a theoretical measure of substitution conservatism.
Note that the signal peptide region and the C-terminal loop preceding the GPI anchor are evidently least constrained by selective pressure. The data affirm expectations from the NMR structure. However, the fragment studied did not include the highly invariant pre-repeat region and the toxic core, which are more stable per unit residue than any of the secondary structure regions.
The species used and substitutions found are listed elsewhere along with [soon] plain-text files giving all known prion gene and protein sequences, published and unpublished. Such a file is convenient for online sequence comparisons. Also vary useful is a [soon] database of gapped and aligned placental mammal sequences and a universal residue numbering scheme. This allows creation of optimized fuzzy BLAST homology probes which are needed to detect distant homologies (and so normal prion function and/or ligands). Plaintext prion DNA coding sequences are posted.
We have been told for years that sporadic CJD comprise 85% of all cases; indeed, a whole speculative literature has grown up around this subject (spontaneous conversion to rogue conformer, somatic mutation, etc.). The question now arises whether sporadic CJD is not being squeezed out of the picture by improved genetic studies and conventional dietary / iatrogenic transmission.
Collinge et al. have found prion promoter polymorphisms in CJD victims. Further, an amino acid change at a new codon, R208H, has just been added [Neurology 47:1305 Nov1996] to the growing list of prion gene mutations seemingly sufficient to cause familial CJD. The change is an A to G transition at the second position within the codon, causing an arginine to histidine shift in the middle of the final 18 residue long alpha helix.
There are two reasons for mis-classification of sporadic CJD. First, it is easy to miss a family history (that can be inferred at codon 208 from a younger sibling): early death from other causes, inadequate querying, short generational history, and so on. Second, it seems that earlier prion gene gel techniques have been inadequate. With DGGE (denaturing gradient gel electrophoresis), an 8M urea-formamide gradient amplifies the melting point differences of mis-matched hybridization pairs arising from single base changes and deletions and gives better detection of genetic changes.
A remarkable finding with ALS could have applicability to sporadic CJD: again making it genetic after all, even though in ALS the coding region of the gene is normal. A tissue-specific (motor cortex and spinal cord) error occurs in mRNA intron editing, yielding an erroneous protein sequence. Either some non-ORF mutation is giving cellular RNA-processing machinery the wrong directions for removing the introns, or a mutation has occurred in the processing machinery itself
The prion gene also has introns, but these are upstream of the coding region. Experimentally, prion protein itself would not normally be sequenced in a case of sporadic CJD because it is faster and safer to work on DNA. Usually, just the ORF DNA is sequenced. Introns are long tracts of DNA with little selective pressure, so irrelevent polymorphisms will make it hard to isolate causative events. Each such case of sporadic CJD might have its own separate genetic basis. There are a variety of reasons why these might not show up as familial (penetrance only in conjunction with other factors, etc.).