Abstract
The U1snRNP-A (U1-A) protein was used to select specific RNA sequences from a degenerate pool of transcripts using direct RNA binding and polymerase chain reaction amplification (PCR). Sequences were randomized in loops of 10 or 13 nucleotides or as a linear stretch of 25 nucleotides. From all three structural contexts, an unpaired ten nucleotide consensus sequence was obtained. A selected stem-loop structure that resembled the natural U1-A protein binding site on loop II of U1 RNA demonstrated the highest affinity of binding in comparison with the other structural contexts. A data profile of selected sequences identified U1 RNA upon searching the GenBank database. Thus, this method was useful in determining the sequence specificity of an RNA binding protein and may complement the use of phylogenetic comparisons to predict conserved recognition elements. These findings also suggest that the evolutionary conservation of loop II of U1 RNA results from constraints imposed by protein binding.