ABSTRACT
The function(s) and RNA binding properties of vigilin, a ubiquitous protein with 14 KH domains, remain largely obscure. We recently showed that vigilin is the estrogen-inducible protein in polysome extracts which binds specifically to a segment of the 3′ untranslated region (UTR) of estrogen-stabilized vitellogenin mRNA. In order to identify consensus mRNA sequences and structures important in binding of vigilin to RNA, before vigilin was purified, we developed a modified in vitro genetic selection protocol. We subsequently validated our selection procedure, which employed crude polysome extracts, by testing natural and in vitro-selected RNAs with purified recombinant vigilin. Most of the selected up-binding mutants exhibited hypermutation of G residues leading to a largely unstructured, single-stranded region containing multiple conserved (A) n CU and UC(A) n motifs. All eight of the selected down-binding mutants contained a mutation in the sequence (A) n CU. Deletion analysis indicated that approximately 75 nucleotides are required for maximal binding. Using this information, we predicted and subsequently identified a strong vigilin binding site near the 3′ end of human dystrophin mRNA. RNA sequences from the 3′ UTRs of transferrin receptor and estrogen receptor, which lack strong homology to the selected sequences, did not bind vigilin. These studies describe an aproach to identifying long RNA binding sites and describe sequence and structural requirements for interaction of vigilin with RNAs.
ACKNOWLEDGMENTS
We are grateful to L. Kunkel for the dystrophin cDNA clone, to A. Martı́nez del Pozo for the gift of α-sarcin, and to R. Gumport for helpful comments on the manuscript.
This research was supported by NIH grants DK-50080 and HD-16720.