In a widely accepted model， the steroid receptor RNA activator protein (SRA protein; SRAP) modulates the transcriptional regulatory activity of SRA RNA by binding a specific stem-loop of SRA. We first confirmed that SRAP is present in the nucleus as well as the cytoplasm of MCF-7 breast cancer cells， where it is expressed at the level of about 10(5) molecules per cell. However， our SRAP-RNA binding experiments， both in vitro with recombinant protein and in cultured cells with plasmid-expressed protein and RNA， did not reveal a specific interaction between SRAP and SRA. We determined the crystal structure of the carboxy-terminal domain of human SRAP and found that it does not have the postulated RRM (RNA recognition motif). The structure is a five-helix bundle that is distinct from known RNA-binding motifs and instead is similar to the carboxy-terminal domain of the yeast spliceosome protein PRP18， which stabilizes specific protein-protein interactions within a multisubunit mRNA splicing complex. SRA binding experiments with this domain gave negative results. Transcriptional regulation by SRA/SRAP was examined with siRNA knockdown. Effects on both specific estrogen-responsive genes and genes identified by RNA-seq as candidates for regulation were examined in MCF-7 cells. Only a small effect (~20% change) on one gene resulting from depletion of SRA/SRAP could be confirmed. We conclude that the current model for SRAP function must be reevaluated; we suggest that SRAP may function in a different context to stabilize specific intermolecular interactions in the nucleus.