Silence Is Not Always Golden

Current strategies for clinical genetic testing for heritable disorders may fail to detect an important set of pathogenic gene mutations. This deficiency reflects the facts that limited sequencing strategies, such as exon-directed (exome) analysis, may exclude transcriptional regulatory, as well as intronic, regions and more extensive sequencing and high-resolution single nucleotide polymorphism microarray approaches are still relatively expensive (Maxmen, 2011). Additionally, silent mutations in exons encoding the open-reading frame may be dismissed as nonpathogenic polymorphisms because they are not predicted to alter the protein product. In this regard, mutations that affect the RNA splicing patterns of a gene transcript may be particularly difficult to uncover because splicing is regulated by a complex array of pre-mRNA sequence elements at the 5′ (donor) and 3′ (acceptor) splice sites, as well as other degenerate sequence motifs located in both exons and introns (exonic and intronic enhancer and silencers).

In this issue, Chung and colleagues (Brenner et al., 2011) highlight this genetic testing concern by describing a novel RNA splicing mutation in the MITF gene from a large family with autosomal dominant Waardenburg Syndrome (WS). Mutations in six genes (MITF, PAX3, SNAI2, SOX10, EDN3, EDNRB) are associated with the four types of WS (WS1-4), which show phenotypic variability in characteristic WS disease features, including sensorineural hearing loss, white forelock, dystopia canthorum, and abnormal pigmentation of skin, hair, and iris. Although genetic testing had failed to reveal the causative mutation, linkage analysis led these authors to the MITF (microphthalmia-associated transcription factor) gene, and subsequent sequencing combined with expression analysis uncovered a c.1212 G>A (Thr404Thr) synonymous mutation in all affected, but not unaffected, family members. Using several web splice site prediction resources (ESEfinder, Berkeley Drosophila Genome Project Splice Site Prediction by Neural Network), the authors discovered that this mutation generates an alternative 3′ (or acceptor) splice site (ss) in MITF exon 9. The introduction of this 3′ ss results in a 52-base pair deletion, a frameshift, premature translational termination, and a truncated MITF C-terminus. Based on this study, Chung and colleagues argue for routine inclusion of all sequence variants in clinical reports and for the use of computational ss prediction tools to evaluate potential effects of synonymous variants on splicing. Given the escalating number of disease-associated mutations that alter RNA splicing patterns (Cooper et al., 2009), this is an excellent suggestion.