Prp43p contains a processive helicase structural architecture with a specific regulatory domain

Abstract

The DEAH/RNA helicase A (RHA) helicase family comprises proteins involved in splicing, ribosome biogenesis and transcription regulation. We report the structure of yeast Prp43p, a DEAH/RHA helicase remarkable in that it functions in both splicing and ribosome biogenesis. Prp43p displays a novel structural architecture with an unforeseen homology with the Ski2-like Hel308 DNA helicase. Together with the presence of a beta-hairpin in the second RecA-like domain, Prp43p contains all the structural elements of a processive helicase. Moreover, our structure reveals that the C-terminal domain contains an oligonucleotide/oligosaccharide-binding (OB)-fold placed at the entrance of the putative nucleic acid cavity. Deletion or mutations of this domain decrease the affinity of Prp43p for RNA and severely reduce Prp43p ATPase activity in the presence of RNA. We also show that this domain constitutes the binding site for the G-patch-containing domain of Pfa1p. We propose that the C-terminal domain, specific to DEAH/RHA helicases, is a central player in the regulation of helicase activity by binding both RNA and G-patch domain proteins.

Figure 1

Structure of yeast Prp43p. (A) Schematic representation of the domain organization of Prp43p. Structural domain names and boundaries are indicated on top, conserved sequence motifs (roman numerals) and domains (hatched areas) beneath. (B) Stereo view of the overall structure of Prp43p in complex with ADP-Mg²⁺. The six structural domains of Prp43p are coloured as in (A). The β-hairpin element (light blue) protrudes from the RecA2 domain (dark blue) between motifs V and VI. The ADP molecule is shown as magenta sticks, and the magnesium ion is represented as a magenta sphere.

Figure 2

Detailed comparison of the ATPase sites of Prp43p and the Hel308 homologue Hjm. The RecA1 and RecA2 domains are coloured as in Figure 1. Residues that directly interact with the ADP molecule (magenta sticks) are shown as sticks and labelled. The conserved sequence motifs are also indicated. (A) Close-up stereo view of the Prp43p ADP-binding site. The magnesium ion is pictured as a magenta sphere. (B) Close-up stereo view of the Hel308 homologue Hjm ADP-binding site (PDB 2ZJ5; Oyama et al, 2009) in the same orientation as in (A).

Figure 3

RNA unwinding by the Prp43p helicase. (A) Overview of the Hel308/DNA complex (PDB 2P6R; Buttner et al, 2007) in the same orientation and colour code as in Figure 1B. The DNA strands are shown in magenta (threaded strand) and pink cartoons. (B) Model of the Prp43p/DNA complex modelled against the Hel308/DNA complex by simple superposition of the proteins (same orientation and colour code as in (A)). (C) Surface representation of Prp43p in complex with modelled DNA (magenta and pink), coloured according to its electrostatic potential calculated with APBS (Baker et al, 2001) and presented in the same orientation as in (B). (D) Close-up view of the nucleic acid-binding cavity entrance of the Prp43p/DNA complex. The β-hairpin element and the OB-fold domain are coloured light blue and red, respectively. The loops L12 (dark green) and L45 (black) and specific positively charged residues (pictured as sticks) of the OB-fold, which are mutated in the probed Prp43p variants, are also indicated. (E, F) Close-up views of the conformational rearrangements of each of the Prp43p domains (E, initial conformation) to adopt the Hel308 conformation (F, final conformation). The Prp43p domains and modelled DNA are pictured as in (B). The two black arrows in (E) show the concerted movement of the RecA-like and ratchet domains, which provides a possible model for RNA translocation and unwinding mechanism of Prp43p helicase: the ratchet domain seems to pull the threaded single strand inside the cavity (left arrow), whereas the β-hairpin slices through the nucleic acid duplex (right arrow).

Figure 4

Analysis of the production and RNA binding of Prp43p variants. (A) Coomassie blue staining (left) and western blot analysis (right) using anti-histidine antibodies of purified histidine-tagged wild-type Prp43p and the indicated histidine-tagged Prp43p variants. (B) Band retardation assays performed with the 5′ half of snR5 snoRNA at 1 nM hybridized to a 21 nucleotide-long radiolabelled ssDNA (Lebaron et al, 2009) and wild-type Prp43p (left), Prp43pΔL12 (middle) or Prp43p(1–657) (right) at the indicated concentrations.

Figure 5

Analysis of the ATPase activity of Prp43p variants. ATP hydrolysis time courses were performed with the indicated proteins either in the absence of RNA (A), in the presence of RNA (B) or in the absence of RNA and presence of Pfa1p (C). Wild-type or Prp43p variants at 100 nM were incubated for the indicated times with cold ATP at 100 μM, mixed with α-³²P-ATP tracer. In addition, total yeast RNA was added at 150 μM in (B) and Pfa1p at 500 nM in (C). Aliquots of reaction mixes were then subjected to thin layer chromatography. ATP and ADP were quantified by phosphorimager to derive ADP concentrations. The abbreviations 664, 704, 708, ΔL12, ΔL45 and 1–657 refer to Prp43pR664A-His, Prp43pK704A-His, Prp43pR708A-His, Prp43pΔL12-His, Prp43pΔL45-His and Prp43p(1–657)-His, respectively.

Figure 6

Prp43p(1–657) fails to interact with the G-patch-containing domain of Pfa1p. The indicated combinations of His-tagged wild-type or truncated Prp43p and Pfa1p proteins were incubated at 33 nM for the wild type and 130 nM for the truncated proteins and precipitated with anti-Prp43p antibodies. Proteins in the pellet (IP) and supernatant (Sup) fractions were detected by western blot analysis using anti-histidine antibodies.