Biology Reference
In-Depth Information
lesions occur disproportionately at the most evolutionarily conserved positions
within the splice site and fall into three main categories: (i) mutations within 5
or 3
splice sites which reduce the amount of correctly processed mature RNA
and/or activate alternative ('cryptic') splice sites in the vicinity, (ii) mutations out-
with actual splice sites which create cryptic splice sites, and (iii) mutations in the
branch-point sequence (Krawczak
et al
., 1992). The vast majority of the patholog-
ical lesions affecting mRNA splicing so far reported have been single base-pair
substitutions within splice sites. This is not only because these are comparatively
frequent but also because they are both readily detectable and highly likely to
result in a severe clinical phenotype. Disease-associated mutations affecting 5
splice sites are approximately twice as frequent as mutations at 3
splice sites
(Krawczak
et al.
, 1992). This discrepancy coincides with a much higher level of
sequence conservation at 5
splice sites and is likely to reflect the strong require-
ment for U1 snRNA binding at 5
splice sites to promote alignment and cleavage.
Regarding the phenotypic consequences of pathological mutations affecting
mRNA splicing, the exclusion of one or more exons from the end-product (
exon
skipping
) is observed more frequently than
cryptic splice site utilization
(Krawczak
et
al.
, 1992). Some evidence exists, at least for 5
splice sites, that cryptic splice site
usage is favored under conditions where a number of potential sites are present in
the vicinity of the mutated splice site and where these potential splice sites exhibit
sufficient homology to the consensus sequence (Krawczak
et al.
, 1992). In most
such cases, the activating mutation improves the similarity between the cryptic
site and the splice site consensus sequence. At 3
splice sites, the amount of
mRNA product consequent to the utilization of the cryptic splice site appears to
be correlated with the level of similarity to the consensus sequences; at 5
splice
sites, the distance to the nearest wild-type splice site may also play a role
(Krawczak
et al.
, 1992).
Splicing mutations and gene inactivation.
The alteration of mRNA processing
as a result of mutations in splice sites also occurs during evolution. One dramatic
example is that of the once active human
L
-gulono-
-lactone oxidase (
GULOP
)
gene (chromosome 8p21.1) which was inactivated at least in part by the introduc-
tion of single base-pair substitutions at the invariant bases of splice sites
(Nishikimi
et al.
, 1994; see Chapter 6, section 6.2.4). Another similar example is
provided by the
CSHL1
gene, long regarded as a pseudogene of the human pla-
centally expressed growth hormone/chorionic somatomammotropin gene family
which is clustered on chromosome 17q22-q24. The
CSHL1
'pseudogene' was
originally thought to have been inactivated by the introduction of a G
A substi-
tution in the first position of the intron 2 donor splice site (Press
et al.
, 1994).
However,
in vitro
expression studies have shown that it may only have been par-
tially inactivated (Misra-Press
et al.
, 1994). Although five alternatively spliced
forms of
CSHL1
mRNA are produced from the 'pseudogene,' the majority of
these transcripts lack exon 2 which encodes the signal peptide necessary for nor-
mal secretion. Some low abundance
CSHL1
mRNAs are nevertheless produced
which possess the leader peptide and these could in principle encode novel hor-
mones of physiological significance.
