 |
SINGLE-STRANDED NUCLEIC ACID BINDING PROTEINS
Gorgio Manzini, Antonella Bandiera and Gianluca Tell, Department of Biochemistry, Biophysics and Macromolecular Chemistry, University of Trieste.
A relevant part of single-stranded nucleic acid binding proteins in eukaryots is given by the hnRNP proteins in a broad sense. On the basis of their electrophoretic mobility and/or their sequence homologies they have been assigned to several families. Notwithstanding, their fairly high number (likely in the hundreds according to evidences from human genome sequencing) not many of them have been deeply characterised from both structural and functional points of view. Although most of them are, or appear, involved in RNA processing, evidences that some of them are acting at the genome level (as transcription factors or in telomere functioning) are accumulating (Krecic and Swanson, 1999, Curr. Opin. Cell Biol., 11, 363-71, and references therein).
Since several years a research line active in this laboratory aims to isolating and characterising nuclear proteins able to interact specifically with telomeric type repeated sequences, which could be involved in telomere functioning in vivo. In particular attention has been focused on proteins able to specifically recognize these sequences in the single-stranded form. Initially, through electrophoretic mobility shift assays and UV-crosslinking experiments, a quite specific binding activity towards the d(CCCTAA)n repeated motif has been found in nuclear extracts from HeLa cells (Marsich et al., 1996, Nucleic Acids Res., 24, 4029-33). Subsequently this type of activity has been shown to be present also in extracts from various mammalian sources, both tissues and cell lines, and its degree of specificity assessed with a series of competitive EMSAs (Marsich et al., 1998, Eur. J. Biochem., 258, 93-9).
In order to identifying the protein(s) responsible for this activity, chicken erythrocytes have been chosen as a convenient source to isolate and characterize such molecules. Affinity chromatography to telomeric-type ssDNA has allowed the recovery of three main protein components, that, after purification, have been identified by IS and MALDI-TOF mass spectrometric analyses (Marsich et al., 2001, Eur. J. Biochem., 268, 139-48). The higher molecular weight component displayed the less sequence-specific activity and resulted to be very likely the yet undescribed avian homolog of mammalian hnRNP K, whereas the lower MW components coincided with two isoforms from alternative splicing of another hnRNP protein, already described in the literature as a likely liver specific transcription factor, termed ssDBF (Smidt et al., 1995, Nucleic Acids Res., 23, 2389-95).
A plasmid clone of the cDNA of the smaller isoform has been obtained and the recombinant protein expressed. Notwithstanding a minor alteration at the N-terminus and the absence of post-translational modifications, certainly present in the natural form, the recombinant has shown to maintain its DNA binding activity. Moreover, the protein has been shown to be expressed also in an avian cell line, BML2, and the larger isoform to appear only after DMSO treatment. This protein shows sequence homology with a number of hnRNPs, namely those of the A/B and D subfamilies, which are characterized by the presence of two RBDs (RNA Binding Domain) in tandem, plus a C-terminal Glycine-rich domain and, in some cases an N-terminal poorly characterized domain. In this context it is worth noting that the tertiary structure of the two RBDs of hnRNP A1 complexed with the telomeric DNA sequence d(TTAGGGTTAGGG) has been resolved by X-ray crystallography (Ding et al., 1999, Genes Dev., 13, 1102-15).
Fig. 1, left shows the folding of the b-a-b-b-a-b motif in the tertiary structure of a RBD.
Figure1. Left: tertiary structure of the RBD domain; right: tertiary structure of the KH domain.
After setting up the isolation and identification procedures, the most recent activity has been devoted to isolating and identifying analogous proteins from human sources. In particular nuclear protein extracts from K562 cells have been chosen as starting material. Eight protein components have been recovered by affinity chromatography and isolated by poly-acrylamide gel electrophoresis. Seven of them could be identified by mass spectrometry. Six resulted to be hnRNP proteins, as expected. One was beta-actin, captured indirectly, almost certainly through its interaction with one of the others. Unexpectedly none of them was the human homolog of avian ssDBF, which has already been described, although no function has assigned to it up to now. Apart from hnRNP K, two isoforms of hnRNP I, and a protein termed JKTBP, which appeared to bind telomeric-type DNA with relatively low specificity, two components appeared to maintain after purification a clear preference for this DNA.
They resulted to be hnRNP E1, alias PCBP1, a protein from an intronless gene, already proposed to be involved in several apparently unrelated activities, and one of the four known splicing isoforms of hnRNP D0. This evidence appears interesting since very recently it has been shown that different isoforms of this protein exhibit binding properties towards either telomeric-type repeat single-strands (Eversole and Maizels, 2000, Mol. Cell. Biol., 20, 5425-32). Recombinant vectors for both PCBP1 and two of the isoforms of hnRNP D0 have been obtained and their expression is on the way. As to sequence homology, PCBP1 is a member of the hnRNP E and K subfamilies, which are characterized by the presence of three KH (K Homology) domains. The folding of the b-a-a-b-b-a motif within the tertiary structure of a KH domain is shown in Fig. 1, right. As to hnRNP D0, it shares sequence homologies and tertiary structure features, like the presence of two RBDs, with the hnRNP A/B and hnRNP D members.
|
