Limits...
In silico and in vivo splicing analysis of MLH1 and MSH2 missense mutations shows exon- and tissue-specific effects.

Lastella P, Surdo NC, Resta N, Guanti G, Stella A - BMC Genomics (2006)

Bottom Line: However, apart from the mutations in the donor and acceptor sites, the effects on splicing of other sequence variations are difficult to predict.However, computer predictions do not always correlate with in vivo splicing defects.Our results suggest that the available algorithms, while potentially helpful in identifying splicing modulators especially when they are located in weakly defined exons, do not always correspond to an obvious modification of the splicing pattern.

View Article: PubMed Central - HTML - PubMed

Affiliation: Section of Medical Genetics, Department of Biomedicine in Childhood, University of Bari, Italy. geneticamedica@medgene.uniba.it

ABSTRACT

Background: Abnormalities of pre-mRNA splicing are increasingly recognized as an important mechanism through which gene mutations cause disease. However, apart from the mutations in the donor and acceptor sites, the effects on splicing of other sequence variations are difficult to predict. Loosely defined exonic and intronic sequences have been shown to affect splicing efficiency by means of silencing and enhancement mechanisms. Thus, nucleotide substitutions in these sequences can induce aberrant splicing. Web-based resources have recently been developed to facilitate the identification of nucleotide changes that could alter splicing. However, computer predictions do not always correlate with in vivo splicing defects. The issue of unclassified variants in cancer predisposing genes is very important both for the correct ascertainment of cancer risk and for the understanding of the basic mechanisms of cancer gene function and regulation. Therefore we aimed to verify how predictions that can be drawn from in silico analysis correlate with results obtained in an in vivo splicing assay.

Results: We analysed 99 hMLH1 and hMSH2 missense mutations with six different algorithms. Transfection of three different cell lines with 20 missense mutations, showed that a minority of them lead to defective splicing. Moreover, we observed that some exons and some mutations show cell-specific differences in the frequency of exon inclusion.

Conclusion: Our results suggest that the available algorithms, while potentially helpful in identifying splicing modulators especially when they are located in weakly defined exons, do not always correspond to an obvious modification of the splicing pattern. Thus caution must be used in assessing the pathogenicity of a missense or silent mutation with prediction programs. The variations observed in the splicing proficiency in three different cell lines suggest that nucleotide changes may dictate alternative splice site selection in a tissue-specific manner contributing to the widely observed phenotypic variability in inherited cancers.

Show MeSH

Related in: MedlinePlus

Effects of the hMLH 1 exon 17 mutations on ESE sequences identified by the different algorithms. The complete sequence of exon 17 is shown (exonic sequence in capital bold). Numbering is relative to the nucleotide position in the ORF. The four exon 17 mutations are shown, wild type sequence underlined. The consequences of the four mutations on the predicted motif scores identified by ESEfinder, RescueESE and PESX are shown below.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC1590028&req=5

Figure 3: Effects of the hMLH 1 exon 17 mutations on ESE sequences identified by the different algorithms. The complete sequence of exon 17 is shown (exonic sequence in capital bold). Numbering is relative to the nucleotide position in the ORF. The four exon 17 mutations are shown, wild type sequence underlined. The consequences of the four mutations on the predicted motif scores identified by ESEfinder, RescueESE and PESX are shown below.

Mentions: Figure 2A,B shows the results of the RT-PCR experiments on Cos-7 cells transfected with the 20 different mutations analyzed and their respective normal controls consisting of the corresponding non mutated exon. As a positive control for the splicing assay we used the C6354T mutation in exon 51 of the FBN1 gene, which has already been reported to cause exon skipping both in vivo and in vitro [20]. Overall, the results of this analysis showed that even if all the mutations fall in ESE sites predicted from either ESEfinder, or RescueESE or PESX, less than half of them led to splicing alterations. In particular, aberrant splicing should be expected whenever a mutation abrogates one or more ESE sites without creating novel sequence motifs recognized by other SR proteins. However, even when the mutations were clustered in a small exonic region, they demonstrated a splicing proficiency not always corresponding to the one expected on the basis of the algorithms predictions. Paradigmatic results were obtained from the splicing assay of the four different hMLH1 mutations T1958G, C1961T, A1963G and G1976C, that all lie in exon 17. The first two lead to concurrent creation of novel ESE sites, in addition to those already present in the wild type sequence, and disruption of one ESE motif for the SRp55 and one for the SF2/ASF splicing factor, respectively (see figure 3 and table 1). In fact, these two mutations did cause only slight changes to the ratio of exon inclusion compared to the normal exon (fig. 2). The A1963G mutation abrogated the same two ESE motifs, as well as one of the two ESEs predicted by RescueESE, but A1963G while not creating any novel ESE sequence for ESEfinder and RescueESE, did generate an ESE sequence according to PESX (fig. 3). The G1976C mutation added a novel ESE site to the one already identified by ESEfinder as present in the wild type sequence, while no ESE motifs are predicted in either the normal or the mutated allele by both RescueESE and PESX. Surprisingly, only the G1976 mutation dramatically altered the rate of exon inclusion in the splicing assay (figures 2, 3 and table 1), while the A1963G mutation, which should have been responsible for the most severe effect on splicing, according to both ESEfinder and RescueESE, instead caused an increase in the exon inclusion rate.


In silico and in vivo splicing analysis of MLH1 and MSH2 missense mutations shows exon- and tissue-specific effects.

Lastella P, Surdo NC, Resta N, Guanti G, Stella A - BMC Genomics (2006)

Effects of the hMLH 1 exon 17 mutations on ESE sequences identified by the different algorithms. The complete sequence of exon 17 is shown (exonic sequence in capital bold). Numbering is relative to the nucleotide position in the ORF. The four exon 17 mutations are shown, wild type sequence underlined. The consequences of the four mutations on the predicted motif scores identified by ESEfinder, RescueESE and PESX are shown below.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC1590028&req=5

Figure 3: Effects of the hMLH 1 exon 17 mutations on ESE sequences identified by the different algorithms. The complete sequence of exon 17 is shown (exonic sequence in capital bold). Numbering is relative to the nucleotide position in the ORF. The four exon 17 mutations are shown, wild type sequence underlined. The consequences of the four mutations on the predicted motif scores identified by ESEfinder, RescueESE and PESX are shown below.
Mentions: Figure 2A,B shows the results of the RT-PCR experiments on Cos-7 cells transfected with the 20 different mutations analyzed and their respective normal controls consisting of the corresponding non mutated exon. As a positive control for the splicing assay we used the C6354T mutation in exon 51 of the FBN1 gene, which has already been reported to cause exon skipping both in vivo and in vitro [20]. Overall, the results of this analysis showed that even if all the mutations fall in ESE sites predicted from either ESEfinder, or RescueESE or PESX, less than half of them led to splicing alterations. In particular, aberrant splicing should be expected whenever a mutation abrogates one or more ESE sites without creating novel sequence motifs recognized by other SR proteins. However, even when the mutations were clustered in a small exonic region, they demonstrated a splicing proficiency not always corresponding to the one expected on the basis of the algorithms predictions. Paradigmatic results were obtained from the splicing assay of the four different hMLH1 mutations T1958G, C1961T, A1963G and G1976C, that all lie in exon 17. The first two lead to concurrent creation of novel ESE sites, in addition to those already present in the wild type sequence, and disruption of one ESE motif for the SRp55 and one for the SF2/ASF splicing factor, respectively (see figure 3 and table 1). In fact, these two mutations did cause only slight changes to the ratio of exon inclusion compared to the normal exon (fig. 2). The A1963G mutation abrogated the same two ESE motifs, as well as one of the two ESEs predicted by RescueESE, but A1963G while not creating any novel ESE sequence for ESEfinder and RescueESE, did generate an ESE sequence according to PESX (fig. 3). The G1976C mutation added a novel ESE site to the one already identified by ESEfinder as present in the wild type sequence, while no ESE motifs are predicted in either the normal or the mutated allele by both RescueESE and PESX. Surprisingly, only the G1976 mutation dramatically altered the rate of exon inclusion in the splicing assay (figures 2, 3 and table 1), while the A1963G mutation, which should have been responsible for the most severe effect on splicing, according to both ESEfinder and RescueESE, instead caused an increase in the exon inclusion rate.

Bottom Line: However, apart from the mutations in the donor and acceptor sites, the effects on splicing of other sequence variations are difficult to predict.However, computer predictions do not always correlate with in vivo splicing defects.Our results suggest that the available algorithms, while potentially helpful in identifying splicing modulators especially when they are located in weakly defined exons, do not always correspond to an obvious modification of the splicing pattern.

View Article: PubMed Central - HTML - PubMed

Affiliation: Section of Medical Genetics, Department of Biomedicine in Childhood, University of Bari, Italy. geneticamedica@medgene.uniba.it

ABSTRACT

Background: Abnormalities of pre-mRNA splicing are increasingly recognized as an important mechanism through which gene mutations cause disease. However, apart from the mutations in the donor and acceptor sites, the effects on splicing of other sequence variations are difficult to predict. Loosely defined exonic and intronic sequences have been shown to affect splicing efficiency by means of silencing and enhancement mechanisms. Thus, nucleotide substitutions in these sequences can induce aberrant splicing. Web-based resources have recently been developed to facilitate the identification of nucleotide changes that could alter splicing. However, computer predictions do not always correlate with in vivo splicing defects. The issue of unclassified variants in cancer predisposing genes is very important both for the correct ascertainment of cancer risk and for the understanding of the basic mechanisms of cancer gene function and regulation. Therefore we aimed to verify how predictions that can be drawn from in silico analysis correlate with results obtained in an in vivo splicing assay.

Results: We analysed 99 hMLH1 and hMSH2 missense mutations with six different algorithms. Transfection of three different cell lines with 20 missense mutations, showed that a minority of them lead to defective splicing. Moreover, we observed that some exons and some mutations show cell-specific differences in the frequency of exon inclusion.

Conclusion: Our results suggest that the available algorithms, while potentially helpful in identifying splicing modulators especially when they are located in weakly defined exons, do not always correspond to an obvious modification of the splicing pattern. Thus caution must be used in assessing the pathogenicity of a missense or silent mutation with prediction programs. The variations observed in the splicing proficiency in three different cell lines suggest that nucleotide changes may dictate alternative splice site selection in a tissue-specific manner contributing to the widely observed phenotypic variability in inherited cancers.

Show MeSH
Related in: MedlinePlus