Structure of the hDim1/U5-15kD Protein

by Kelly Hrywkiw

Introduction

The protein Dim1 goes by many names, hDim1/U5-15kD in humans, Dim1p in Schizosaccharomyces pombe (S. pombe), and Dib1p/Snu16p in Saccharomyces cerevisiae (S. cerevisiae)^[1][2][3]. Regardless of the name it is extraordinarily well conserved across the eukaryotic kingdom sharing 79% sequence identity between the human and S. pombe orthologs, and 66% sequence identity between the human and S. cerevisiae orthologs^[3][2]. Dim1 exhibits a thioredoxin like domain core, and is thought to play a multifunctional role in the spliceosome – the pre-mRNA splicing machine (Zhang 2003)^[3][2].

Pre-mRNA Splicing

Pre-mRNA contains exons (protein coding regions) separated by introns (non-coding regions)^[4]. The spliceosome catalyzes the removal of the introns and the ligation of the remaining exons to create mature mRNA. This reaction takes place through two transesterification reactions catalyzed by specific portions of the spliceosome^[5]. The overall structure of the spliceosome is highly dynamic and consist of five integral complexes known as small nuclear ribonucleoprotein particles (snRNPs) U1, U2, U4, U5 and U6 ^[4]. Each snRNP consist of a uridine rich small nuclear RNA (snRNA), seven sm or sm-like (Lsm) proteins, and several additional proteins^[6].

It has been proposed that the snRNPs associate in a stepwise manner ^[4]. The first step involves U1 snRNP recognizing the 5’ splice site, followed by an ATP dependent step in which U2 associates with the branch point ^[4][6]. The tri-snRNP comprised of U4/U6 U5 then binds, and through a series of rearrangements U6 replaces U1 at the 5’ splice site resulting in an activated complex that can perform the first of two transesterification reactions ^[4][6]. U1 and U4 then dissociate to yield the second active complex which completes the second reaction ^[4][6]. The remaining snRNPs then dissociate. It should be noted that these processive structural rearrangements would not be possible without the presence of ATP ^[6].

Role of Dim1 in Splicing

While the role of Dim1 in splicing is not well understood is has been shown, in multiple studies, to be a constituent of the tri-snRNP, and to strongly associate specifically with the U5 snRNP ^[3][2]. It is also required for the splicing of a non pre-mRNA, U3 RNA^[3]. Furthermore, hDim1 has been shown to interact with the proteins hnRNP F and hnRNP H’, which enhance tissue specific pre-mRNA splicing, as well as Npw38/PQBP-1 which can bind poly(rG) and co-localize with splicing co-activators when co-expressed with hDim1^[3]. hDim1 also contains RNA recognition motif-like sequences, which suggest that it could directly interact with poly(rG)^[3]. Dib1p strongly interacts with Prp6, a protein required for tri-snRNP accumulation. Overall, this suggests that Dim1 plays a multifaceted role in splicing biogenesis^[3][2].

Structure

Two structures of Dim1 have been solved, an oxidized full length form and a reduced dominant negative form. The oxidized full length form of Dim1 consists of 142 amino acids and its overall structure adopts a thioredoxin like core domain coupled with a C-terminal extension^[2]. The reduced dominant negative form contains 128 amino acids, where the C-terminal extension containing residues 129-142 have been removed^[3].

Oxidized Full Length Form (hDim1)

The thioredoxin-like fold of follows the arrangement of a five stranded β-sheet, consisting of parallel and antiparallel stands, surrounded by three α-helices, where the loop between β4 and α3 could not be resolved. When compared to human thioredoxin there are a total of 37 additional residues in hDim1, which result in several structural differences. For example, the N-terminus is extended by three residues, in the α2-β2 loop one residue is inserted, after β4 nine are inserted, before the α2 helix two are inserted, and . This results in an altered structure where β4 and β5 appear to be pulled away from the β-sheet leaving a between β3 and β4^[2].

In thioredoxin there is an N-terminal Cys-X-X-Cys motif which forms a functional disulfide bond, in hDim1 there is an . Interestingly, in hDim1 the Cys residue still participates in the formation of a disulfide bond with a residue located in β3 (Cys79) resulting in a located in a similar area as that in thioredoxin. Through a series of experiments it has been shown that hDim1 does not appear to participate in redox or protein disulfide isomerase activity as seen in thioredoxin. However, it still may be possible that hDim1 forms disulfide bonds with other spliceosomal proteins, or exhibits peroxiredoxin-like activity^[2].

On the surface of hDim1 there are , one located towards the N-terminal region of hDim1 containing residues: Trp12, Val14, Ile18, Lue19, Phe30, Phe69, and Phe84 (pink), and one located in the clef between β3 and β4 containing residues: Met72, Met82, Met91, Ile92, Lue94, Ile102, and Trp104 (green) separated by a region of the highly conserved residues Met72, His89, and Met91 (red). The exposed hydrophobic regions coupled with the presence of five highly conserved and exposed Met residues helps solidify the role of hDim1 in protein-protein interactions with other spliceosomal proteins. In addition to the hydrophobic regions, there is also a region of conserved (Arg86, Lys88, Arg121, Arg124, Lys125, and Arg127) which may be important to binding negatively charged RNA ^[2].

Reduced Dominant Negative Form (hDim1-128)

The removal of the C-terminal extension induces cell cycle arrest in G2, however does not affect localization, steady-state levels, or phosphorylation of the protein^[3]. This suggests that it may be the interactions to other proteins and substrates that are disrupted by the removal of the C-terminal extension ^[1]. It is therefore important to understand the functional changes that the presence of the C-terminal extension creates .

The structure of compared to hDim1 is remarkably similar, where the same mixed β-sheet flanked by three α-helices is seen. However a prominent difference includes the loss of the β-strand comprised of residues 129-131 and 91-93. By comparing the circular dichroism spectra of hDim1 to hDim1-128 there is a decrease in α helical structure upon truncation. This suggests that the C-terminal region consists of a partially α-helical region in solution, which contradicts the findings of the hDim1 crystal structure. However, it possible that this region is naturally flexible allowing it to adopt several conformations, thereby enabling it to interact with various regions of the spliceosome as it changes through a splicing event. Interestingly, when the basic residues , which were suggested to partake in the formation of a positively charged region in hDim1 (see above), were mutated, there was a major decrease in structural stability and cooperative folding^[3].