what-when-how
In Depth Tutorials and Information
TABLE 10.5
Analysis of the Effects of the -1G>A sequence Alterations using the splicing site finder Online Analysis Tool
Sequence
Variant
OI
Type
Splicing
Score
Splicing
Score
Wild-Type Sequence
Mutant Sequence
c.958-1G>A
I
aacggggcttcccctctctcctgc
ag
GGTGCTCGT
95.13
aacggggcttcccctctctcctgca
ag
GTGCTCGT
84.37
c.859-1G>A
IV
ctcactcattattctctgatctac
ag
GGTGAGCCT
90.88
ctcactcattattctctgatctaca
ag
GTGAGCCT
78.24
c.1669-1G>C
III
gctgctttcgtgcctcccatcctt
ag
GGTCCCGCC
83.4
gctgctttcgtgcctcccatcctt
ac
GGTCCCGCC
-
a frameshift which will lead to NMD of the resulting
mRNA. The available data are somewhat limited and
are confined to
COL1A1
. In that gene “-1G>A” sequence
variants are found immediately upstream of exons
5, 13, 15, 19, 26, 32, 35, 37, 38, 39, 40, 43 and 48, each of
which begins with a G. Seventeen of the 21 documented
patients have OI type I with the other four being type
IV. Two of these latter cases are the result of sequence
variation at the splice junction immediately before exon
13 (c.859-1G>A).
21
It is perhaps instructive to analyze
the predicted effects on splicing using
in silico
methods
and compare the results with a splice-site variant adja-
cent to exon 15 (c.958-1G>A) resulting in two indepen-
dent cases of OI type I.
21
Using the Splicing Site Finder
online analysis tool,
51
the effects of the -1G>A sequence
alterations can be analyzed (
Table 10.5
). Splice donor
and acceptor sites which conform precisely to their con-
sensus sequences are assigned a score of 100 and those
with divergent sequences are assigned lower scores. The
c.958-1G>A variant resulting in OI type I in two unre-
lated individuals
21
is assigned a score of 84.37 compared
with the wild-type sequence score of 95.13. By compari-
son, the c.859-1G>A variant resulting in OI type IV in
two unrelated individuals
21
is assigned a score of 78.24
compared with the wild-type sequence score of 90.88.
In both cases, the wild-type sequences are spliced suf-
ficiently well to produce functional α1-chains. Although
the sequence variations in each case result in the splic-
ing score being reduced by a similar magnitude, it is
presumed that the lower mutant score of 78.24 for the
c.859-1G>A variant perturbs splicing such that the adja-
cent exon is skipped sufficiently frequently that a slightly
more severe OI type IV phenotype results. For the sake
of comparison, the
COL1A1
c.1669-1G>C variant, which
results in OI type III (unpublished), completely elimi-
nates the splice acceptor site and is assumed to result in
skipping of the adjacent exon 25. Unfortunately there
are no mRNA analysis data for any of the three variants
presented in
Table 10.5
and the cause of the phenotypic
variation remains purely speculative.
To further complicate the picture, unrelated individu-
als harboring the same -1G>A splice site variant do not
always present with the same severity of OI. The c.1768-
1G>A variant produces OI type I in two individuals
52,53
but OI type IV in a third,
21
and the c.2128-1G>A variant
produces OI type I in three individuals
21
but OI type IV
in a fourth (unpublished). This suggests that phenotype
predictions may be confounded by additional variants
which might modify the splicing process. These might
be local
cis
-acting variants such as single-nucleotide
polymorphisms (SNPs)
54
which have a direct effect
on splicing, or SNPs in regulatory regions influencing
expression levels
55,56
of the normal or mutant alleles.
Another possible modifier of phenotype severity is
trans
-acting variation in tissue-specific splicing factors.
57
Not all splice-site sequence variants exert their effects
by way of the mechanisms described above. The
COL1A1
c.642+1G>A variant has been shown to disrupt splic-
ing through the activation of cryptic splice donor sites
in intron 7 and exon 8 resulting in five different mutant
mRNAs
58
and a phenotype of OI type IV. However, the
identical sequence variant in an unrelated patient results
in OI type III.
21
Two patients have been described with
the nearby c.642+5G>A sequence variant
59,60
producing
OI types I and IV. In the case of the OI type IV patient,
59
analyses demonstrate activation of the intron-7 cryptic
splice site and retention of part of intron 7 in the resulting
mRNA as described for the c.642+1G>A variant.
58
These
results again emphasize the difficulty in attempting to
derive genotype
/
phenotype correlations.
The vast majority of reported sequence variants
resulting in the disruption of normal splicing are
located close to exon ends in the flanking donor and
acceptor sites. Although this is probably broadly con-
sistent with the true distribution of such variants,
there is an element of bias in that most OI-causing
sequence variants are determined nowadays by PCR-
based analyses confined to the exons and immediately
flanking intron sequences, rather than by analysis of
mRNA.
61
Consequently, deep-intronic sequence vari-
ants will be missed unless analysis of fibroblast mRNA
is also performed. This is exemplified by the
COL1A1
variant c.2451+77C>T in intron 35 which creates a new
donor splice site resulting in retention of the first 75 bp
of the intron which are translated into 25 amino acids,
maintaining the reading frame and resulting in OI
type II
62
(Rowe, personal communication). Additional
OI-causing deep-intronic variants almost certainly
remain undiscovered because of the current practice in
many diagnostic laboratories of only sequencing PCR-
amplified genomic DNA in the immediate vicinity of
the exons.
