Biology Reference
In-Depth Information
7.5.1 Neighboring-nucleotide effects on the rate of germline single
base-pair substitutions in human genes
In terms of their relative frequency of occurrence, the most important category of
single base-pair substitution causing human genetic disease is represented by C
T
and G
A transitions within CpG dinucleotides; some 23% of all pathological sin-
gle base-pair substitutions found within the coding regions of human genes are of
this type (Krawczak et al ., 1998). Allowing for the confounding effects of codon
usage and differential clinical observation likelihoods through consideration of
relative mutabilities, this proportion translates into a mean transition rate for
either CG
CA that is five times higher than the base mutation rate.
Krawczak et al . (1998) demonstrated that the proportion of pathological
CG
TG or CG
(TG,CA) transitions is significantly higher for autosomal genes (25.0%) than
for X-linked genes (17.7%). These proportions are a direct reflection of the signif-
icantly lower frequency of CpG in the coding sequences of X-linked genes (2.9%)
as compared to autosomal genes (3.7%). Krawczak et al . (1998) speculated that the
lower CpG frequency in X-chromosomal genes may be a consequence of a gener-
ally increased level of DNA methylation resulting from the evolutionary recruit-
ment of this post-synthetic modification to play a role in X-inactivation
(Hornstra and Yang, 1994; Jamieson et al ., 1996).
For CpG dinucleotides to be hypermutable in the context either of genetic dis-
ease or evolution, they must be methylated in the germline (El-Maarri et al .,
1998). Since it cannot be excluded that the efficiency of both DNA methyltrans-
ferase action (Smith, 1994; Smith and Baker, 1997) and G: T mismatch repair
(Sibghat-Ullah and Day, 1993) may be influenced by sequence motifs flanking the
CpG dinucleotide, the question arises as to whether some CpGs may be intrinsi-
cally more mutable than others by virtue of their DNA sequence context.
Significant differences in the relative mutation rate of CpG dinucleotides depend-
ing upon their flanking nucleotides were indeed noted by Krawczak et al . (1998).
These results were consistent with those of Ollila et al . (1996) who noted a prefer-
ence for 5
purines flanking mutated CpG dinucleotides.
Comparison of the sequences flanking the human and chimpanzee
pyrimidines and 3
-globin
genes has shown that the CpG dinucleotide is hypermutable over evolutionary
time and subject to high frequency C
A transitions (Savatier et al .,
1985). Indeed, some 40% of the CpG dinucleotides present in either the human or
chimpanzee sequences were found to be affected by nucleotide sequence changes.
Similar conclusions have been drawn by Perrin-Pecontal et al . (1992). Comparison
of the CpG mutation rates exhibited by globin gene and pseudogene sequences
from human, chimpanzee and macaque yielded an estimate of the rate of 5mC
deamination of ~1 × 10 -16 (Cooper and Krawczak, 1989). The absence of any sig-
nificant difference between deamination rate estimates derived from gene and
pseudogene sequence data suggested that the action of selection has had a negli-
gible effect on the transition rate at CpG dinucleotides in primate
T or G
-globin genes.
Indeed, this constancy in the CpG deamination rate is consistent with a neutral-
ist view of gene evolution. Moreover, the successful use of evolutionary compar-
isons of DNA sequences to derive consistent values of the CpG deamination rate
has demonstrated the feasibility of using the CpG deamination rate as a 'molecu-
lar clock', at least over relatively short periods of evolutionary time.
 
 
Search WWH ::




Custom Search