SR proteins are a family of proteins that have a common epitope recognized by a monoclonal antibody (MAb104) that binds active sites of polymerase II transcription. Four of the SR family members have been shown to restore activity to an otherwise splicing-deficient extract (S100 extract). Here we show that two untested SR proteins, SRp20 and SRp75, can also complement the splicing-deficient extract. We isolated a cDNA encoding SRp75 and found that this protein, like other SR proteins, contains an N-terminal RNA recognition motif (RRM), a glycine-rich region, an internal region homologous to the RRM, and a long (315-amino-acid) C-terminal domain composed predominantly of alternating serine and arginine residues. The apparent molecular mass of dephosphorylated SRp75 is 57 kDa, the size predicted from the cDNA clone. We also detected mobility shifts after dephosphorylating SRp55, SRp40, SRp30a, and SRp30b; the sizes of the shifts are proportional to the length of the SR domain, suggesting that serines in this domain are phosphorylated.