Contractions and expansions in gene size and number - Human Gene Evolution - page 363

Yamagata et al ., 1996), whilst a second disease-associated haplotype has been
found in Africans (Krahe et al ., 1995). Similarly, haplotype analysis has pointed to
a single origin for Japanese and Caucasian Machado-Joseph disease chromosomes
(Takiyama et al ., 1995), a single origin for Japanese and Caucasian dentatorubral
and pallidoluysian atrophy chromosomes (Yanagisawa et al ., 1996), a single origin
for the Friedreich ataxia expansion (Cossée et al ., 1997), at least two origins for the
Huntington disease expanded repeat (Squitieri et al ., 1994) and a small number of
FRAXA progenitor chromosomes (Hirst et al ., 1994).
One of the best examples of the evolutionary emergence of an expanded triplet
repeat is that found in the SRP14 gene (15q22) encoding the 14 kDa Alu RNA-
binding protein (Chang et al ., 1995). The human protein is larger than that of
mouse and dog on account of an extra 28 residue alanine-rich C-terminal tail
which is translated from a 3

GCA-rich trinucleotide repeat. In the prosimian
Galago , the relevant sequence at the site of the human triplet repeat is GCA GCA,
whereas the mouse possesses the sequence CCA GCA. By contrast, the African
green monkey possesses 33 GCA repeats, whilst the owl monkey possesses 52 sug-
gesting that a CCA GCA substitution occurred in an ancestral prosimian

thereby creating two consecutive GCA codons which then facilitated GCA expan-
sion in higher primates. Interestingly, however, no intra-specific variability in
repeat size was detected in primates.

8.9.4 Evolution of repeat number in the genes underlying disorders of
triplet repeat expansion

The normal range for CAG repeat number in the HD gene is 8-35. Although
modal CAG repeat length is fairly similar between different human populations,
the degree of spread varies, being greatest among Africans and lowest among the
Japanese (Rubinsztein et al ., 1994; Watkins et al ., 1995). The breadth of this nor-
mal range suggests that natural selection is acting weakly if at all on HD alleles
below the disease threshold. However, the population distribution of CAG repeat
number in the HD gene exhibits an apparent asymmetry in that more alleles lie
above rather than below the modal length. Using computer simulations,
Rubinsztein et al . (1994) have shown that the distribution of HD alleles in human
populations is explicable in terms of a simple length-dependent mutational bias.
The observed distribution of alleles is thus explicable merely in terms of mutation
and genetic drift thereby obviating the need to invoke positive selection
(Rubinsztein et al ., 1994). It may be, however, that coding sequences can tolerate
CAG repeat-encoded polyglutamine tracts relatively well, thereby minimizing the
effect of negative selection (Green and Wang, 1994). More controversially, in the
case of myotonic dystrophy and Machado-Joseph disease, meiotic drive (also
termed segregation distortion ), the excess recovery of one of a pair of alleles in the
gametes of an heterozygous parent, has been proposed as being responsible for
maintaining the frequency in the population of chromosomes bearing triplet
repeats capable of expansion into the disease range (Chakraborty et al ., 1996;
Leeflang et al ., 1996; Takiyama et al ., 1997).
Djian et al . (1996) examined the disease-associated CAG repeats in the HD ,
MJD , AR , and SCA1 orthologues of various nonhuman primate species. For the
HD and MJD genes, CAG copy number was polymorphic in the nonhuman pri-