Human genomic sequences that inhibit splicing

Mammalian genes are characterized by relatively small exons surrounded by a sea of introns. Sequences that resemble splice signals occur frequently within the introns of mammalian genes. How the cell distinguishes the real splice signals from the great multitude of false sites is unclear. Although much progress has been made in defining additional sequences that enhance the inclusion of an exon into the mRNA our knowledge is not such that we can accurately predict whether particular sites will be used. To increase our understanding of the role of exonic sequences in splice site selection we undertook a search for inhibitory sequences. To generate inhibitory sequences we developed a cell selection system that isolated sequences that would cause “exon skipping” when ligated into a test exon. Screening populations of human fragments with this criterion yielded the surprising discovery that these sequences are extremely common in human DNA. About one of three restriction fragments resulted in an aberrant splicing phenotype. This is in stark contrast to inserts from other sources like E. coli or random DNA where only one of the twenty-seven fragments tested resulted in inhibition. We went on to demonstrate that these inserts could function in a heterologous exon and used mutagenesis to identify the sequence elements within the inhibitors that were responsible for the inhibitory activity. We chose a polypyrimidine tract from one of our inhibitors for further study.; A mechanistic exploration of the insert-mediated inhibition indicated that U2AF65's binding the insert was important for its inhibitory function. Expanding the polypyrimidine tract by 11 more pyrimidines resulted in a total reorganization of the UV crosslinking profile of the probe. All interactions including that of U2AF65 were displaced in favor of a 35 Kd interaction. This 35 kD protein bound the polypyrimidine tract and inhibited splicing by blocking early complex formation at the upstream 3 ′ss. The U2AF65 mediated inhibition was also replicated in vitro though a particular site of action could not be identified unambiguously.