Purification and Nucleic Acid Binding Properties of a Fragment of Type C1/C2 Heterogenous Nuclear Ribonucleoprotein from Thymic Nuclear Extracts

A single-strand nucleic acid binding protein (C/F) that has an apparent molecular weight of 12 000 on SDS-polyacrylamide gel electrophoresis and that was originally thought to be the 12-kDa alpha -subunit of the AB form of terminal deoxynucleotidyl transferase (TdT) from calf thymus has been purified and identified as a fragment of the type C1/C2 hnRNP proteins. On the basis of NH sub(2)-terminal sequencing and mass spectrometric analysis, C/F contains similar to 94 residues and spans from residue 9 to approximately residue 102 in the type C1/C2 hnRNP proteins. C/F is presumably produced in vitro via limited proteolysis of the type C1/C2 hnRNP proteins following cell disruption. Since C/F corresponds almost exactly to the similar to 90-residue conserved ribonucleoprotein binding domain (RBD) that is shared by many eukaryotic RNA binding proteins, it provided an opportunity to better characterize the domain structure of the type C1/C2 hnRNP proteins and to compare the nucleic acid binding properties of the type C1/C2 and A1 RNA binding domains. Like the type A1 RBD, the type C1/C2 RBD has an apparent occluded site size of 6-7 nucleotides. The type C1/C2 RBD binds non-cooperatively to homopolynucleotides and has preferential affinity for RNA and for single as opposed to double-stranded nucleic acids. The type C1/C2 RBD has about a 100-fold higher affinity than the type A1 RBD does for RNA and some of this increased affinity results from additional ionic interactions. The latter account for similar to 50% of the free energy of binding of the type C1/C2 RBD. While the type C1/C2 hnRNP proteins exist in vivo as a very tight tetramer with the structure (C1) sub(3)C2, the isolated type C1/C2 RBD is a monomer. Hence, the determinants for tetramerization appear to lie outside the type C1/C2 RBD. Phenylalanine 19 was identified as the only point of photochemical cross-linking of the type C1/C2 RBD to [d(T)] sub(8). This residue corresponds to the major site of cross-linking of the A1 RBD to [d(T)] sub(8). Thus, even though the overall extent of direct sequence identity between the type A1 and C RNA binding domains is only 17%, both proteins appear to share a topologically similar RNP:oligonucleotide interface which does not appear to be substantially altered when it is placed within the context of a multidomain hnRNP protein.