Crystal structure of the spliceosomal DEAH-box ATPase Prp2

Abstract

The DEAH-box ATPase Prp2 plays a key role in the activation of the spliceosome as it promotes the transition from the B^act to the catalytically active B* spliceosome. Here, four crystal structures of Prp2 are reported: one of the nucleotide-free state and three different structures of the ADP-bound state. The overall conformation of the helicase core, formed by two RecA-like domains, does not differ significantly between the ADP-bound and the nucleotide-free states. However, intrinsic flexibility of Prp2 is observed, varying the position of the C-terminal domains with respect to the RecA domains. Additionally, in one of the structures a unique ADP conformation is found which has not been observed in any other DEAH-box, DEAD-box or NS3/NPH-II helicase.

Keywords: DEAH-box; Prp2; RNA helicase; spliceosome.

Figure 1

Overall structure of Prp2 from C. thermophilum. The model of ctPrp2 is depicted as a cartoon. The remaining amino acids from the truncated N-terminal extension (270–296) are shown in black, the RecA1 domain (297–475) is in orange, the RecA2 domain (476–652) is in blue, the winged-helix domain (WH; 653–720) is in grey, the helix-bundle domain (HB; 721–852) is in wheat and the oligonucleotide-binding fold (OB; 853–920) is in green.

Figure 2

Structural comparison of ctPrp2 and scPrp43 (PDB entry 3kx2; He et al., 2010 ▸). (a) Superposition of ctPrp2 (blue) and scPrp43 (yellow), depicted here as ribbon models, reveals a high structural similarity, resulting in an r.m.s.d. of 1.3 Å for 573 common C^α atoms. (b) The most prominent difference is found at the N-termini of the proteins. The N-termini of ctPrp2 and scPrp43 are shown as cartoons, whereas the protein core of scPrp43 is shown as a surface (grey). The N-terminal extension of ctPrp2 (blue) protrudes from the protein core, whereas the N-terminus of scPrp43 (red) wraps around the RecA1 domain and parts of the C-terminal domains.

Figure 3

Flexibility of the C-terminal entity. (a) A superposition of all of the ctPrp2 structures via the RecA2 domain highlights the flexibility of the C-terminal domains with respect to the RecA2 domain. While the helicase cores superpose well, the C-terminal domains reveal various positions. (b) Despite the observed flexibility, ten common interdomain interactions between the C-terminal domains and the helicase core can be found among the four structures. The residues involved in these conserved interactions are displayed as spheres and labelled accordingly. ctPrp2 is shown as a ribbon model and the domains are coloured according to Fig. 1 ▸.

Figure 4

Comparison of the different β-hairpin conformations. (a) All ctPrp2 structures are superposed via their RecA2 domains. The different β-hairpins are presented as cartoon models using the colouring of Fig. 3 ▸(a), while the remaining part of the protein is shown as a surface (nucleotide-free structure only). The OB-fold domain was omitted for clarity. The superposition reveals that the β-hairpin of ctPrp2 can adopt different conformations. (b) Superposition of ctPrp43 structures in different functional states via their RecA2 domains (PDB entries 5d0u, 5lta, 5ltj and 5ltk; Tauchert et al., 2016, 2017 ▸). The structures are presented according to (a) using different colours.

Figure 5

Comparison of differently bound ADP molecules. ADP molecules are presented as ball-and-stick models and the magnesium ion and water molecules as spheres. ctPrp2 is depicted as a cartoon representation with highlighted residues as ball-and-stick models. (a) depicts the mF _o − DF _c electron-density OMIT map for the magnesium ion with the four coordinated waters of CF3 (grey) contoured at 3σ. This OMIT map is representative of all ADP-bound structures. The OMIT maps at the same contour level for the ADP molecules of CF2 (orange) and CF3 are displayed in (b) and (c), respectively. (d) On superposing the RecA1 domains of CF2 and CF3 the phosphate moieties align almost identically, but the adenosine is present in a previously undescribed conformation, leading to different adenosine interactions in CF2. (e) A superposition of the different ADP molecules via the ribose reveals that the adenine is present in a syn (CF2) and an anti (CF3) conformation and that the torsions along the C4′—C5′ bond and the C5′—O5′ bond are changed by 167 and 62°, respectively. (f) Depending on the ADP conformations described in (d) and (e), a loop in the nearby motif VI adopts distinct conformations. A superposition of the RecA2 domains reveals a flipped conformation of this loop in the NT-free and CF2 structures compared with CF1 and CF3, which exhibit an ADP-binding pattern comparable to other DExD/H-box helicases.

Figure 6

The nucleotide-free Prp2 structure with a bound sulfate molecule. A superposition of the RecA1 domains of the NT-free and CF1 structures shows that a sulfate that is present at the active site of the nucleotide-free structure adopts the same position as the β-phosphate of the bound ADP.