Zum Inhalt springenZur Suche springen

An experimentally supported bioinformatics approach to analyze disease-associated 5 ' splice site mutations


Pathogenic splicing alterations are increasingly recognized as a widespread mechanism through which gene mutations cause disease. Thus, in assessing a splice site mutation's pathogenicity, reliable in silico prediction of its in vivo splicing outcome would increase the efficiency of genomic DNA based mutation detection assays and possibly resolve diagnostic dilemmas in patients. However, today's in silico implementation of the complex splicing machinery is still limited to a variety of independent algorithms scoring splice sites and/or cis-regulatory elements. In contrast, splicing regulatory elements are known to act in concert, and their interactions and dependencies play an important role in splice site functionality. Currently available in silico prediction tools have been applied to patients' splice site mutations and their predictive reliability has been assessed by endogenous RNA analyses and/or in splicing reporter assays. We exemplarily characterized the HIV-1 bidirectional splicing enhancer to get molecular insights into the splicing enhancer mechanisms to support the selective use of alternative splice sites.


Computational analysis of splice site mutations
Within a set of experimentally well documented splice sites from various human genes, web-based computer programs (S&S, MaxEnt, HBond and SD-score) consistently "predicted" aberrant splicing based upon the "weaker mutant than wild type" concept. This concept claims that, irrespective of the position within a splice site, any mutation that sufficiently reduces the intrinsic strength of the splice site leads to aberrant splicing. However, only HBond and SD scores provide an explicit binary prediction of aberrant splicing, while for the other scores - for the lack of a specific threshold value - any smaller score value for the mutant was considered indicative of aberrant splicing. In general, splice site strength assessments showed a high consistency both among the four different algorithms examined and compared to the RNA-based data, but their applicability is limited until specific thresholds will be defined.

Computational analysis of cis-active SRE mutations

Putative exonic splicing enhancer (ESE) and exonic splicing silencer (ESS) motifs were predicted for a variety of short exonic regions covering missense mutations in various human genes using the five most common web-based programs ESEfinder, RESCUE-ESE, FAS-ESS, PESX and ESRsearch with their default threshold values in order to quantify the mutation-induced change. Using these algorithms to identify splicing regulatory elements (SREs), we found a great variability between the individual predictions for the same nucleotide position both in wild type and mutant sequences. Even though we analyzed only a manageable number of missense mutations, in vivo and in silico concordance between all five programs for a given missense mutation was the exception rather the rule.

Characterization of the HIV-1 bidirectional splicing enhancer
We found that the guanosine-adenosine-rich exonic splicing enhancer (GAR ESE) identified in exon 5 of the human immunodeficiency virus type-1 (HIV-1) consisting of three individual binding sites for SF2/ASF and SRp40 fulfils a dual splicing regulatory function (i) by synergistically mediating exon recognition through its individual SR protein-binding sites and (ii) by effecting 3' ss selectivity within the 3' ss cluster preceding exon 5. Interestingly, every two of the three individual binding sites synergistically mediated the enhancer function, and mutation of any two of the three sites was detrimental for exon recognition, whereas any single mutation only slightly impaired splicing.


So far, a variety of computational algorithms have been developed to measure the strength ("score") of 5' or 3' splice sites, or SREs, thus addressing the splicing process, and indirectly allowing to assess mutation effects by the difference of wild type and mutant scores. Our study reveals current limits of the applied computational tools in the diagnostics of possibly pathogenic splicing mutations. In contrast, we found considerable divergence in the software-predicted SREs both in the reference sequences and in the mutated patient sequences. The apparent disagreement between computational SRE prediction and failure to affect splicing in vivo may reflect the view that SREs - as exemplified by the HIV-1 GAR enhancer - often act in concert, and that we do not know which of the motifs alone is essential for splicing, if any at all. Also, the type of aberrant splicing (exon skipping or activation of a cryptic splice site) can not yet be predicted by any of the examined algorithms. Therefore, any computational prediction of mutation effects still has to be complemented and verified by experimental RNA-based analyses as the current method of choice.

Aims for 2009/10
To characterize viral and human SREs involved in splice site and exon recognition, respectively, and to cure human pathogenic 5' splice site mutations which cause aberrant splicing by genetic therapy using U1 snRNAs with modified free 5'-ends.


2. Publikations belonging to the topics


Asang, C., Hauber, I., and H. Schaal. 2008. Insights into the selective activation of alternatively used splice acceptors by the Human Immunodeficiency Virus Type-1 bidirectional splicing enhancer. Nucl. Acids Res. 36:1450-1463


Hartmann, L., Theiss, S., Niederacher, D., and H. Schaal. 2008. Diagnostics of Pathogenic Splicing Mutations: Does bioinformatics cover all bases ? Front. Biosci. 13:3252-3272

3. Cooperations

  • Dept. of Obstetrics and Gynecology (Prof. Hans-Georg Bender / Dr. Dieter Niederacher): Analyses of disease-related sequence variations in BRCA1 and BRCA2 which might cause aberrant splicing

  • Institute of Human Genetics (Prof. Brigitte Royer-Pokora): Analyses of disease-related sequence variations in hMLH1 and hMSH2 which might cause aberrant splicing

  • BMFZ, Analytisches Zentrallabor (Dr. Sabine Metzger): Identification of proteins involved in splicing regulation.