may affect normal function Sequence Information
Mad Cow Home ... Best Links ... Search this site

Human mutations are strongly concentrated

Human mutations are strongly concentrated

webmaster 16 Nov 98
Mutations giving rise to CJD have no satisfactory explanation yet. The figure shows their highly non-linear distribution within the primary sequence; much of the protein has few or no sites, while the stretch from 178 to 217 has many adjacent to structural features such as helix 2 and 3, the glycosylation sites, the disulfide cysteines, and the GPI attachment site.

Most of the changes are conservative in terms of amino acid substituted but non-conservative in terms of invariance of residue position, as can be seen by analyzing 102 prion sequences from 77 mammalian species:

P102L: proline in all species
P105L: proline in all species
A117V: alanine in all species
M129V: leucine seen in elk allele, methionine all other species
Y145-: tryptophan seen in all rodents, tyrosine all other species
N171S: serine is wildtype in great apes, except gorilla is asparagine
D178N: aspartate in all species
V180I: valine in all species
T183A: threonine  in all species (glycosylation site)
H187R: histidine in all species, possible arginine in cat
F198S: phenylalanine in all species, leucine in spider monkey
E200K: glutamate in all species
D202N: aspartate in all species
R208H: glutamine seen in sheep, goat alleles
V210I: valine in all species
Q212P: glutamine in all species
Q217R: glutamine in all species
E219K: glutamate in primates, glutamine all other species, possibly arginine in nyala
M232R: alanine in rodents and carnivores, valine in artiodactyls, methionine in primates, threonine in mouse, isoleucine in baboon
Note from this analysis that N171S is very likely to be neutral, contrary to one report. Codon 232 is very curious: it comes at a critical place immediately downstream of the GPI splice junction, yet is probably the tolerant to substitution of any residue in the entire post-cleavage residues are very highly conserved; however 3-4 residues upstream of the splice junction are also highly variable. This could just as well be a stragegy for controlling GPI cleavage (and release of prion into extra-cellular space) as for the extent of GPI attachment -- the relative proportions may vary by species or cell type.

R208H is worrisome to see in sheep and goat as alleles -- I wonder what the incidence of early scrapie is in this genotype with the new tests. M129L in elk is of interest in CWD. The more sequences that build up, the odder these seem.

129V is commonly said to be a neutral polymorhism but that is a poor fit to the evolutionary data. I looked briefly to see if, after adjusting for allele frequency, whether sporadic CJD was more common in V129V than M129M (though of course if some or all sporadic CJD is food-bourne it would be from M129M since all food except elk sausage has this and like-like would suppress statistics for V129V): but MM:VV was 2.7:1 {Windl] in sporadics whereas MM:VV alleles were 2.8:1. Growth hormone statistics favored V129V by a factor of 2.5; affected donor genotypes were not known and may have been too few to be a proportional mix.

Of course CJD isn't the selective pressure on residue 129 or anything else because historically it had no effect to speak of on reproductive success, disease having such late onset and lifespans being so short. If we ever get an enzymatic activity for normal prion protein, I will be curious to see if V129V compares unfavorably. The same holds for all the other non-synapomorphic positions, ie, these mutations likely have adverse effects on normal function as well as cause CJD. However, CJD mutations must have their effect in heterozygotes to be detectable whereas disruption of normal function might be partly masked or compensated by the normal allele. That is why the 3 rare homozygous cases of K200K would warrant such careful study if their condition had been detected before the onset of CJD (through early genetic screening of a known kindred).

The mutations not seen are just as informative as the ones observed. That eliminates bad thermodynamics per se. Safir et al. [Nature Medicine Oct 1998] have raised the issue of rate of clearance of bad conformer as perhaps more critical than rate of formation. In other words, it is not enough for a mutation to give rise to bad conformer, that bad conformer must also have exceedingly slow clearance by proteases. V210E is much more extreme than V200I though also a single base change -- perhaps it is not suble enough in its effects and forms lots of bad conformer that turns over too rapidly to give CJD. That is a quantifiable feature of the above list -- the mutations observed are on the bland side of mutationally accessible change. (A third are CpG hotspots which have a preordained direction of change but many potential hotspots are not producing observable mutation -- again, possibly not bland enough.)

P105L:cca S L T A Q R
A117V:gca T V P S E G
M129V:atg I V T L K R
N171S:aac S D K H Y T I
T183A:aca A I P S K R
H187R:cac Y R Q D N P L
F198S:ttc L S I V Y C
V210I:gtt I A L F D G
Q212P:cag R - H K E P L
Q217R:cag R - H K E P L
M232R:atg I V T L K R









Mad Cow Home ... Best Links ... CGi">Search this site