The researchers focused on a brain protein that binds to RNA called Nova. Darnell and his colleagues first identified Nova in 1993 as the target protein in a neurodegenerative disease termed POMA (paraneoplastic opsoclonus-myoclonus ataxia) that is also associated with several types of cancers. Since then the laboratory has focused on identifying RNA sequences -- and in particular, identifying alternatively spliced pre-mRNAs -- that Nova binds to. In the last three years, in work published in Science and Nature Genetics, the Darnell lab identified over 50 Nova-regulated alternatively spliced exons, using new techniques developed at Rockefeller specifically to find Nova RNA targets, and validating their results in "knockout" mice that were missing Nova.
In the new study, Darnell, with co-first authors Jernej Ule and Giovanni Stefani, took these 50 RNA transcripts and searched them for clusters of sequences they had previously identified as Nova binding sites through biochemical and, in collaboration with former Rockefeller University structural biologist Stephen Burley, X-ray crystallographic studies. Unexpectedly, this search revealed four discrete peaks where the binding clusters locate. Furthermore, the location of the peaks correlated with Nova's action on regulating whether the alternative exon is spliced in or out.
The researchers tested whether this RNA map was valid by asking whether it could predict how Nova would act on RNA transcripts that had yet to be discovered. They took a bioinformatics approach, using a database of all alternatively spliced RNAs compiled by co-authors Terry Gaasterland and Bahar Taneri, to search for new genes that had clusters of Nova binding sites. Of the 50 or so transcripts with such clusters, 30 turned