Primary structure and binding activity of the hnRNP U protein: binding RNA through RGG box.

Abstract

Heterogeneous nuclear ribonucleoproteins (hnRNPs) are thought to influence the structure of hnRNA and participate in the processing of hnRNA to mRNA. The hnRNP U protein is an abundant nucleoplasmic phosphoprotein that is the largest of the major hnRNP proteins (120 kDa by SDS‐PAGE). HnRNP U binds pre‐mRNA in vivo and binds both RNA and ssDNA in vitro. Here we describe the cloning and sequencing of a cDNA encoding the hnRNP U protein, the determination of its amino acid sequence and the delineation of a region in this protein that confers RNA binding. The predicted amino acid sequence of hnRNP U contains 806 amino acids (88,939 Daltons), and shows no extensive homology to any known proteins. The N‐terminus is rich in acidic residues and the C‐terminus is glycine‐rich. In addition, a glutamine‐rich stretch, a putative NTP binding site and a putative nuclear localization signal are present. It could not be defined from the sequence what segment of the protein confers its RNA binding activity. We identified an RNA binding activity within the C‐terminal glycine‐rich 112 amino acids. This region, designated U protein glycine‐rich RNA binding region (U‐gly), can by itself bind RNA. Furthermore, fusion of U‐gly to a heterologous bacterial protein (maltose binding protein) converts this fusion protein into an RNA binding protein. A 26 amino acid peptide within U‐gly is necessary for the RNA binding activity of the U protein. Interestingly, this peptide contains a cluster of RGG repeats with characteristic spacing and this motif is found also in several other RNA binding proteins. We have termed this region the RGG box and propose that it is an RNA binding motif and a predictor of RNA binding activity.

Rrp8p is a yeast nucleolar protein functionally linked to Gar1p and involved in pre-rRNA cleavage at site A2

“…The GAR domain is found in the nucleolar proteins Nop1/fibrillarin (Schimmang et al+, 1989;Lapeyre et al+, 1990), Gar1p (Girard et al+, 1992), Nsr1p (Lee et al+, 1991), Ssb1p (Clark et al+, 1990), nucleolin (Lapeyre et al+, 1987), gar2 (Gulli et al+, 1995), and in hnRNP proteins such as hnRNP A1, Nop3p/Npl3, and Hrp1p (Beyer et al+, 1977;Kiledjian & Dreyfuss, 1992;Russell & Tollervey, 1992, Birney et al+, 1993Henry et al+, 1996)+ The GAR domain (also called RGG domain) has RNA-helix-destabilizing properties in vitro (Ghisolfi et al+, 1992), is implicated in nonspecific protein-RNA interactions (Burd & Dreyfuss, 1994a, 1994b and in proteinprotein interactions (Cartegni et al+, 1996, Bouvet et al+, 1998)+ So far, however, no phenotype has ever been associated with the removal of the GAR domain(s) of a given protein (Girard et al+, 1994;Sicard et al+, 1998)+ The two GAR domains of the essential Gar1p protein are not required for viability and, consistent with this, Gar1p lacking both GAR domains (Gar1p⌬GAR) accumulates in the nucleolus (Girard et al+, 1994)+ To assess whether removal of the GAR domains of Gar1p results in subtle growth defects, we compared the growth at various temperatures of a gar1-null strain complemented by the gar1⌬GAR allele (JG535-2D-d; see Table 1) with that of the same gar1-null strain complemented by the GAR1 wild-type allele (JG535-2D-c)+ A growth defect is clearly observed for the strain complemented by the mutant allele (Fig+ 2A)+ To determine whether this growth delay is correlated with a pre-rRNA processing defect, we performed pulse-chase labeling experiments with (methyl-3 H)methionine to label methylated RNAs (Fig+ 2B)+ We can detect a slight accumulation of the 32S pre-rRNA and a delay in mature 18S and 25S rRNA production (see Fig+ 1 for a cartoon of the pre-rRNA processing pathway in yeast)+ These data demonstrate that Gar1p lacking its two GAR domains is not fully functional+ Identification of Rrp8p, a novel functional partner of Gar1p, by a synthetic lethal screen, and cloning of RRP8 Deletion of the GAR domains, which partially impairs Gar1p function but still allows cells to grow, could cause synthetic lethality if another component that physically interacts or functionally overlaps with Gar1p becomes mutated+ Thus such components could be identified by searching for the genes that, when mutated in a gar1⌬GAR background, lead to a lethal phenotype+ To screen for such mutations, we used an ade2-ade3-based colony sectoring assay …”

Distinct functions of the closely related tandem RNA-recognition motifs of hnRNP A1

“…A: Schematic domain structure of human hnRNP A1+ hnRNP A1 consists of two domains; the N-terminal UP1, which is a proteolytic product of hnRNP A1, and the C-terminal Gly-rich domain (G domain)+ UP1 consists of two copies of canonical RRMs (RRM1 and RRM2), each of which possesses the conserved RNP-2 and RNP-1 submotifs+ The G domain comprises an RGG box (Kiledjian & Dreyfuss, 1992) and the M9 nuclear localization motif (Siomi & Dreyfuss, 1995;Weighardt et al+, 1995) in 12 imperfect Gly-rich repeats )+ Amino acid residues are numbered from the initiation codon Met 1+ B: Purified recombinant hnRNP A1 wild-type and variant proteins+ All the proteins are untagged; their structures are shown schematically at left+ Proteins (60 pmol each) were analyzed by 12% SDS-PAGE and Coomassie Blue staining+ General or nonspecific RNA binding by hnRNP A1 and the six variant proteins was also assayed using a 501-nt b-globin pre-mRNA splicing substrate (data not shown)+ Although the binding affinity of wild-type protein was about 10-fold lower than for binding to the high-affinity Rd6-1 probe, the relative affinities of wildtype and variant proteins were nearly identical with the two kinds of RNA probes+ However, binding by A1-D(RRM2) was more similar to that of wild type when the b-globin probe was used+…”

Separable roles in vivo for the two RNA binding domains of a Drosophila A1-hnRNP homolog

“…The large GRD at the C terminus contributes many of the unusual properties of A/B hnRNP binding+ It is ;120-200 residues long in proteins of this class and has a low sequence conservation, though a high conservation of glycine richness (40-50% gly), and a semiregular spacing of flexible, aromatic, and positively charged or polar amino acids (Cobianchi et al+, 1986) resulting in binding determinants along its entire length (Pontius, 1993)+ The GRD of HRB87F/hrp36 contains two RGG boxes (Kiledjian & Dreyfuss, 1992) near its N terminus and an M9-like nuclear shuttling signal near its C terminus+ In addition to binding RNA directly (Kumar et al+, 1990, Casas-Finet et al+, 1993, the GRD of A1 hnRNP is responsible for protein-protein interactions (independent of RNA) between proteins with a GRD (Casas-Finet et al+, 1993;Cartegni et al+, 1996) and for the cooperative binding of A1 to RNA Nadler et al+, 1991)+ We find that the GRD plays an important role in vivo, not only in nuclear localization of the protein, but also in specifying the quantitative deposition properties of the protein at sites of transcription and in protein sequestration at the 93D locus during heat shock+ Figure 1A shows the structure of the wild-type and mutant HRB87F/hrp36 proteins used in this study+ Transformed homozygous fly stocks were established in a wild-type background, each of which expressed a different mutant Hrb87F transgene under the control of the hsp70 promoter+ All of the proteins were epitope tagged at their N terminus with the Flag epitope so that they could be distinguished immunologically from the endogenous HRB87F protein after transformation into flies+ To abrogate the RNA binding properties of one or the other RBD, two highly conserved phenylalanines that occur in the RNP-1 box of the RBDs and participate directly in RNA binding (Merrill et al+, 1988;Oubridge et al+, 1994) were changed to aspartic acid residues, as was done previously for similar studies with human A1 hnRNP (Mayeda et al+, 1994)+ Although we have not done in vitro RNA-binding assays on these mutated proteins, there is ample evidence from previous studies that these changes to RNP-1 will severely compromise RNA-binding ability of the RBD as mediated via its RNA-binding b-sheet platform (e+g+, Mayeda et al+, 1994;Bouvet et al+, 1997;Deardorff & Sachs, 1997)+ This disruption of RNA-binding ability by mutation of conserved aromatics is also predicted by biophysical (Merrill et al+, 1988) and X-ray crystallographic analyses (Oubridge et al+, 1994)+ The phenylalanines were substituted by aspartic-acid residues, which are expected to remain solvent exposed without disrupting the structure of the domain, but unable to participate in the base-stacking interactions important to interaction of RNA with the RNA-binding platform of the RBD+ The wil...…”