Structural and functional analysis of the RNA helicase Prp43 from the thermophilic eukaryote Chaetomium thermophilum

RNA helicases are essential key players in numerous cellular processes. Here, the crystal structure of the spliceosomal DEAH-box helicase Prp43 from C. thermophilum is reported and revealed to be capable of functionally replacing its yeast orthologue in spliceosomal disassembly assays.

Abstract

RNA helicases are indispensable for all organisms in each domain of life and have implications in numerous cellular processes. The DEAH-box RNA helicase Prp43 is involved in pre-mRNA splicing as well as rRNA maturation. Here, the crystal structure of Chaetomium thermophilum Prp43 at 2.9 Å resolution is revealed. Furthermore, it is demonstrated that Prp43 from C. thermophilum is capable of functionally replacing its orthologue from Saccharomyces cerevisiae in spliceosomal disassembly assays.

1. Introduction  

RNA helicases are ubiquitously distributed among all domains of life and are of crucial importance for numerous cellular processes such as pre-mRNA splicing, translation initiation, ribosome biogenesis and RNA transport (Cordin et al., 2006 ▸; Bleichert & Baserga, 2007 ▸; Ozgur et al., 2015 ▸). Helicases have been classified into six superfamilies (SFs) based on phylo­genetic sequence alignments (Fairman-Williams et al., 2010 ▸). The members of SF1 and SF2 share a central helicase core composed of two RecA-like domains. These adjacent domains provide the characteristic interface for DExD/H-box proteins and contain the eight conserved short sequence motifs I, Ia, Ib, II, III, IV, V and VI, in which motif II exhibits the eponymous amino-acid sequence. Mutagenesis and structural studies have unravelled the involvement of motifs I, II, V and VI in NTPase activity. Motifs Ia, Ib and IV are required for RNA binding and motif III couples NTP hydrolysis to RNA unwinding (Cordin et al., 2006 ▸; Hilbert et al., 2009 ▸). DEAH-box proteins belong to the SF2 helicases, which exhibit an additional N-terminal extension as well as three further C-terminal domains: a winged-helix (WH) domain, a ratchet and an oligosaccharide-binding fold (OB-fold) (He et al., 2010 ▸; Walbott et al., 2010 ▸). The DEAH-box family is capable of unwinding DNA as well as RNA substrates (Fairman-Williams et al., 2010 ▸). Contemporarily, it is assumed that DEAH-box helicases bind a single-stranded overhang of a nucleic acid substrate and unwind this substrate by a continuous movement mediated by the two RecA-like domains, i.e. they exhibit a certain level of processivity (Pyle, 2008 ▸). In contrast to the DEAD-box and Ski2-like helicases, DEAH-box proteins can utilize all nucleoside triphosphates, at least in vitro (Kim et al., 1992 ▸; Schwer & Guthrie, 1992 ▸; Tanaka & Schwer, 2005 ▸, 2006 ▸).

Eight conserved SF2 helicases belonging to the DExD/H-box and Ski2-like families are involved in pre-mRNA splicing. These eight helicases are key players in the accurate orchestration of the major compositional and conformational rearrangements which are undergone by the spliceosome during one cycle of intron removal. Prp43 (pre-mRNA processing factor 43) is required for proper disassembly of the yeast intron–lariat spliceosome (ILS) which is composed of the U2·U5·U6 snRNPs (Arenas & Abelson, 1997 ▸; Fourmann et al., 2013 ▸). To accomplish this function, Prp43 interacts with two cofactors: Ntr1 and Ntr2 (nineteen complex-related proteins 1 and 2) (Tanaka et al., 2007 ▸; Tsai et al., 2005 ▸, 2007 ▸; Boon et al., 2006 ▸). Ntr1 contains a G-patch motif (glycine-rich) that stimulates the ATPase and unwinding activities of Prp43 (Tanaka et al., 2007 ▸; Christian et al., 2014 ▸; Robert-Paganin et al., 2015 ▸).

Besides pre-mRNA splicing, Prp43 is additionally involved in ribosome biogenesis, in which it is required for the maturation of 18S and 25S pre-rRNAs (Lebaron et al., 2005 ▸; Bohnsack et al., 2009 ▸). Thereby, stimulation of Prp43 by the G-patch proteins Sqs1 (squelch of splicing suppression protein 1) and Gno1 (G-patch nucleolar protein 1) is essential. These two activator proteins exhibit no sequence identity to Prp43’s spliceosomal activator protein Ntr1 except for the G-patch motif (Aravind & Koonin, 1999 ▸; Pertschy et al., 2009 ▸).

The crystal structure of Prp43 from Saccharomyces cerevisiae (scPrp43) was solved in 2010 by two groups (He et al., 2010 ▸; Walbott et al., 2010 ▸). Here, we report the crystal structure of Prp43 from the thermophilic ascomycetal fungus Chaetomium thermophilum (ctPrp43) at 2.9 Å resolution. After publication of its genome (Amlacher et al., 2011 ▸), a continuously increasing number of protein structures (87 deposited in the PDB to date) from C. thermophilum have been released (Bock et al., 2014 ▸); however, it was not always demonstrated that a putative orthologue from C. thermophilum does actually functionally correspond to its mesophilic counterpart. To address this question for ctPrp43, we performed in vitro spliceosome disassembly assays which clearly demonstrate that ctPrp43 can fully replace scPrp43 in the yeast spliceosome.

2. Materials and methods  

2.1. Macromolecule production  

The identification of the potential homologue of scPrp43 in C. thermophilum was performed using BLAST (Altschul et al., 1990 ▸) against the complete C. thermophilum genome. The highest alignment score was achieved by a protein annotated as ‘hypothetical protein CTHT_0005780’ and referred to here as ctPrp43.

The gene encoding full-length ctPrp43 was amplified from genomic DNA of C. thermophilum var. thermophilum DSM 1495 and cloned into pGEX-6P-1 using the EcoRI and SalI restriction sites. All dispensable bases between the PreScission Protease cleavage site and the starting methionine of ctPrp43 were deleted via site-directed mutagenesis (QuikChange Site-Directed Mutagenesis Kit, Agilent Technologies). The gene of a truncated ctPrp43 construct comprising amino acids 61–764, generated for crystallization, was amplified from pGEX-6P-1-ctPrp43 and was subsequently cloned into pETM-13 with an additional C-terminal Strep-tag using NcoI and SalI restriction sites (see Table 1 ▸).

Table 1. Macromolecule production information for ctPrp43(61–764).

Source organism	C. thermophilum var. thermophilum DSM 1495
Expression vector	pETM-13
Expression host	E. coli Rosetta 2 (DE3)
Complete amino-acid sequence of the construct produced	MATTAKQAEAVEDSDINPWTGQRHSERYFKILKARRKLPVNKQRQEFLDLYHNNQILVFVGETGSGKTTQIPQYVLYDELPHQTGKLIACTQPRRVAAMSVAQRVADELDVKLGEEVGYSIRFENKTSSKTLLKYMTDGQLLREAMHDRDMSRYSCIILDEAHERTLATDILMALLKQLSERRKDLKIIVMSATLDAQKFQSYFFNAPLLAVPGRTHPVEIFYTPEAERDYVEAAIRTVLQIHACEPEGDILLFLTGEEEIEDACRRISLEVDEMIRESDAGPMSVYPLYGTLPPHQQQRIFEKAPQPFRPGGRPGRKCIVATNIAETSLTIDGIVYVVDPGFSKQKIYNPRTRVESLLVSPISKASAQQRAGRAGRTRPGKCFRLYTEEAFKKELIEQTYPEILRSNLSNTVLELKKLGVEDLVHFDLMDPPAPETMMRALEELNYLACLDDDGELTPLGNLASEFPLDPALAVMLISSPEFYCSNEILSITSLLSVPQIWVRPANARKRADEMKAQFAHPDGDHLTLLNAYHAYKGAEARGEDMKKWCHEHFLSYRHLSSADNVRAQLKKIMETHGIELVSTPFHDKNYYTNIRRALLAGFFMQVAMRESSNSKVYKTVKDEQLVLIHPSTTVTTPYEWVVYNEFVLTTKQYVRTVTNIRPEWLLEIAPVYYDLSTFQKGEIKNALTRVAEKIRRQQAMKASKAWSHPQFEK

The fusion proteins GST-ctPrp43 and ctPrp43(61–764)-Strep were expressed from pGEX-6P-1 and pETM-13, respectively, in Escherichia coli Rosetta 2 (DE3) cells at 16°C for 18 h after induction with 0.5 mM IPTG at an optical density (OD₆₀₀) of 0.8. Subsequent to cell disruption via microfluidization (M-110S Microfluidizer) and the isolation of soluble proteins by ultracentrifugation at 35 000g for 30 min, GST-ctPrp43 was loaded onto Glutathione Sepharose 4B (GE Healthcare) and ctPrp43(61–764)-Strep was loaded onto StrepTactin HP Sepharose (GE Healthcare) in 400 mM NaCl, 50 mM Tris–HCl pH 7.5, 10 mM EDTA. After intensive washing with an additional 2 M LiCl, target protein elution was realised with 30 mM reduced glutathione (GST-ctPrp43) or 2.5 mM d-desthiobiotin [ctPrp43(61–764)-Strep]. GST-tag cleavage was realised by the addition of PreScission Protease [1:100(w/w), GE Healthcare]. Protein samples were purified to homogeneity by size-exclusion chromatography (Superdex 200, GE Healthcare) in 100 mM NaCl, 10 mM Tris–HCl pH 7.5, 2 mM MgCl₂ and concentrated to 40 mg ml⁻¹ (Amicon Ultra 50K, Millipore).

scPrp43 was expressed and purified as described previously by Christian et al. (2014 ▸).

2.2. Crystallization  

ctPrp43(61–764)-Strep was diluted to 4 mg ml⁻¹ with gel-filtration buffer and incubated with a tenfold molar excess of ADP. The protein was crystallized using the sitting-drop vapour-diffusion method at 293 K with droplets consisting of equal volumes of protein and reservoir [8%(w/v) Jeffamine M-2070, 0.17 M glycine, 16.7%(v/v) DMSO, 10 mM urea] solutions. Rod-shaped crystals with dimensions of up to 20 × 20 × 2000 µm were obtained after 3–4 d. The crystallization procedure is summarized in Table 2 ▸.

Table 2. Crystallization.

Method	Vapour diffusion
Plate type	Sitting drop
Temperature (K)	293
Protein concentration (mg ml−1)	4
Buffer composition of protein solution	10 mM Tris–HCl pH 7.5, 100 mM NaCl, 2 mM MgCl2
Composition of reservoir solution	8%(w/v) Jeffamine M-2070, 0.17 M glycine, 16.7%(v/v) DMSO, 10 mM urea
Volume and ratio of drop	2 µl; 1:1 ratio
Volume of reservoir (µl)	500

2.3. Data collection and processing  

The crystals obtained were cryoprotected in reservoir solution supplemented with 26%(v/v) glycerol and flash-cooled in liquid nitrogen prior to data collection. Diffraction data were collected on BL14.1 operated by the Helmholtz-Zentrum Berlin (HZB) at the BESSY II electron-storage ring (Berlin-Adlershof, Germany) (Mueller et al., 2012 ▸). Data were processed using the XDS package (Kabsch, 2010 ▸). Data collection and processing statistics are presented in Table 3 ▸.

Table 3. Data collection and processing.

Values in parentheses are for the outer shell.

Diffraction source	BL14.1, BESSY
Wavelength (Å)	0.918
Temperature (K)	100
Detector	Pilatus 6M
Crystal-to-detector distance (mm)	573.5
Rotation range per image (°)	0.2
Total rotation range (°)	60.6
Exposure time per image (s)	6.0
Space group	P65
a, c (Å)	152.2, 92.8
α, γ (°)	90.0, 120.0
Mosaicity (°)	0.361
Resolution range (Å)	50.00–2.92 (2.99–2.92)
Total No. of reflections	90272 (13924)
No. of unique reflections	26545 (4202)
Completeness (%)	99.3 (98.5)
Multiplicity	3.4 (3.3)
CC1/2 (%)	99.7 (67.8)
〈I/σ(I)〉	15.4 (2.0)
R meas	0.102 (0.770)
Overall B factor from Wilson plot (Å2)	63.6
Molecules per asymmetric unit	1

2.4. Structure solution, refinement and structural analysis  

The structure of ctPrp43(61–764) was solved by molecular replacement using Phaser (McCoy et al., 2007 ▸) with chain A of scPrp43 (PDB entry 2xau; Walbott et al., 2010 ▸) as a search model. Manual model building was conducted with Coot (Emsley et al., 2010 ▸) and refinement was performed with PHENIX (Adams et al., 2010 ▸). During the refinement process, TLS refinement for chain A as a single group was performed and the coordination distances for the Mg²⁺ ion were restrained to 2.07 Å for water molecules and the Thr126 (OG1) side chain and to 2.09 Å for the β-phosphate (O2B), allowing a deviation of 0.2σ. After iterative cycles of model building and refinement, the final model (Table 4 ▸) was assessed for correctness using MolProbity (Chen et al., 2010 ▸). Figures were prepared with PyMOL (v.1.3; Schrödinger) and Chimera (Pettersen et al., 2004 ▸). Electrostatic surface potentials were calculated using PDB2PQR (Dolinsky et al., 2004 ▸) as well as APBS (Baker et al., 2001 ▸) and the surface conservation was visualized with AL2CO (Pei & Grishin, 2001 ▸).

Table 4. Structure solution and refinement.

Values in parentheses are for the outer shell.

Resolution range (Å)	43.94–2.92 (2.99–2.92)
Completeness (%)	99.3
σ Cutoff	F > 1.36σ(F)
No. of reflections, working set	26538 (1719)
No. of reflections, test set	2022 (143)
Final R cryst	0.190 (0.313)
Final R free	0.231 (0.384)
No. of non-H atoms
Total	5729
Protein	5661
Magnesium	1
ADP	27
Water	4
DMSO	36
R.m.s. deviations
Bonds (Å)	0.0043
Angles (°)	0.843
Average B factors (Å2)
Overall	76.5
Protein	76.3
Magnesium	66.8
ADP	79.5
Water	72.5
DMSO	103.8
Ramachandran plot
Most favoured (%)	96.44
Allowed (%)	3.56
Outliers (%)	0
Rotamer outliers (%)	0.32
MolProbity clashscore	7.50
MolProbity overall score	1.65
PDB code	5d0u

2.5. Functional yeast assays  

The in vitro reconstitution and disassembly assays from purified yeast spliceosomes were performed as described previously and were analysed on a linear 10–30%(v/v) glycerol gradient (Fourmann et al., 2013 ▸).

2.6. Protein melting-point determination  

The melting point was determined for ctPrp43 and scPrp43 via CD spectroscopy using a Chirascan CD spectrometer (Applied Photophysics) by monitoring the unfolding of α-helical protein regions. Far-UV spectra and melting curves were acquired in 10 mM Tris–HCl pH 7.5, 100 mM NaCl, 2 mM MgCl₂. Melting curves were determined at 222 nm for 12 s per °C between 20 and 84°C and far-UV spectra were recorded between 195 and 260 nm before (20°C) and after (84°C) denaturation of protein samples. The obtained melting curves were fitted according to Pace et al. (1998 ▸). Raw data were smoothed by a factor of two with SX-Pro-Data (Applied Photophysics).

3. Results and discussion  

3.1. Overall structure of Prp43 from C. thermophilum  

ctPrp43 crystallized (see Table 2 ▸) in the hexagonal space group P6₅ and its structure was refined at a resolution of 2.92 Å with R _work and R _free values of 0.190 and 0.231, respectively (Tables 3 ▸ and 4 ▸). Since the crystallization of a full-length ctPrp43 construct remained fruitless, a truncated variant was generated lacking the first 60 N-terminal residues, which are mainly present in loop regions as shown by Walbott et al. (2010 ▸). After molecular replacement using the yeast structure with PDB code 2xau (Walbott et al., 2010 ▸), the remaining 704 residues as well as the first amino acid from the attached C-terminal Strep-tag were traceable in the electron density. The Ramachandran plot reveals that 96.44% of all residues are in the most favoured regions and no residues are in disallowed regions. One monomer of ctPrp43(61–764) is present in the asymmetric unit, which also corresponds to its functional state in solution as determined by analytical gel filtration and multi-angle light scattering (data not shown). These findings are also true for the full-length protein.

Besides the protein ctPrp43(61–764), one ADP molecule, one magnesium ion, four water molecules and nine DMSO molecules are present in the final structure (Fig. 1 ▸). ctPrp43 can be divided into six domains, namely the N-terminal extension (residues 1–96), the RecA1 (97–273) and RecA2 (274–459) domains, the WH domain (459–526), the ratchet (527–640) and the OB-fold (641–764). A characteristic feature of all DEAH-box helicases is the presence of a β-hairpin, which is located in the RecA2 domain and comprises residues 401–420 in ctPrp43. While the two RecA-like domains are a typical feature of all SF1 and SF2 helicase members, the three C-terminal domains can only be found in the DEAH helicase family.

Figure 1.

Crystal structure of ctPrp43(61–764) at 2.92 Å resolution. The overall structure of the N-terminally truncated ctPrp43(61–764) is depicted as a cartoon model. The remaining part of the N-terminal extension (amino acids 61–96) is shown in brown, the RecA1 domain (97–273) in light green, the RecA2 domain (274–458) in violet, the WH domain (459–526) in red, the ratchet (527–640) in blue and the OB-fold (641–764) in orange. The bound ADP molecule is shown in ball-and-stick mode and C atoms are coloured yellow, N atoms blue, O atoms red, P atoms orange and the Mg²⁺ ion light green.

The surface representation of ctPrp43 reveals a tunnel inside the molecule which was proposed to be an RNA-binding site (Walbott et al., 2010 ▸) by deduction from a superposition with the Ski2-like helicase Hel308 bound to DNA (PDB entry 2p6r; Büttner et al., 2007 ▸). From the electrostatic surface potential (Fig. 2 ▸), the orientation of an RNA molecule in this tunnel can be predicted. Since the surface of the ratchet domain in this binding tunnel exhibits a highly negative charge, one can assume that the bases of the RNA are oriented in this direction. The RecA1 and RecA2 domains provide basic patches which are compatible with the binding of phosphate groups from the RNA backbone. This binding mode would also explain why Prp43 can bind to RNA in a sequence-independent manner (Tanaka & Schwer, 2006 ▸). The surface potential of ctPrp43 also exhibits an additional RNA-binding site which was previously identified biochemically for scPrp43 (Walbott et al., 2010 ▸). Mutagenesis or deletion of this region, located in the OB-fold, was shown to drastically decrease the RNA-stimulated ATPase activity as well as the affinity towards RNA.

Figure 2.

Electrostatic surface potential of ctPrp43(61–764). The electrostatic surface potential was calculated using APBS (Baker et al., 2001 ▸) and is depicted at a contour level of ±6k _B T. Blue indicates positive charge and red negative charge. The side (left) and back (right) view of ctPrp43 are shown and (proposed) RNA-binding sites as well as the nucleotide-binding site are indicated by dashed circles. For further information, refer to §3.1.

3.2. ADP binding of ctPrp43  

The prototypical domain motifs (Fairman-Williams et al., 2010 ▸; Walbott et al., 2010 ▸) of all SF2 helicase members were identified for ctPrp43 and are highlighted in Fig. 3 ▸. Motifs Ia (¹⁴⁹TQPRRVAA¹⁵⁶), Ib (¹⁹⁵TDGQLLR²⁰¹) and IV (³¹⁰LLFLTG³¹⁵) have been reported to interact with substrate nucleic acid and motif III (²⁵⁰SAT²⁵²) couples nucleoside triphosphate hydrolysis to substrate unwinding. Motifs I (¹²²GSGKT¹²⁶) (also denoted as the P-loop) (Rudolph et al., 2006 ▸), II (²¹⁸DEAH²²¹), V (³⁸¹TNIAETSLT³⁸⁹) and VI (⁴²⁸QRAG­RAGR⁴³⁵) are involved in nucleotide binding. The ADP molecule is sandwiched between the RecA1 and RecA2 domains. Here, the adenine is bound via π-electron stacking and cation–π interaction between the side chains of Arg162 and Phe360 (Fig. 4 ▸). These two residues are not part of the conserved motifs but also appear to be relevant for nucleotide binding. In Prp43, as well as all other spliceosomal DEAH-box proteins, the base of the nucleoside triphosphate is not specifically recognized, in contrast to the DEAD-box and Ski2-like helicases, which contain an additional binding motif, the Q-motif, that elicits adenine specificity. In ctPrp43, the ribose forms hydrogen bonds (for bonding distances, see Supplementary Table S1) from O2′ and O3′ to Asp391 (motif V) and Arg435 (motif VI), respectively. The involvement of Arg435 in hydrogen bonding is plausible owing to Gly122, the main-chain carboxyl group of which attracts the proton of the O3′ into its direction, thereby allowing O3′ to form a hydrogen bond to a hydrogen of Arg435. Furthermore, the α-phosphate interacts via hydrogen bonds with the main chain and side chain of Thr127, which is adjacent to motif I, and water 1. The β-phosphate exhibits intensive hydrogen bonding to the Lys125 and Thr126 side chains, as well as to Gly122, Gly124 and the Lys125 main-chain amides, all belonging to motif I, and to water molecules 3 and 4. In addition to this, the β-phosphate is involved in the coordination of the Mg²⁺ ion, the hexavalent coordination sphere of which comprises four additional water molecules and the Thr126 side chain (motif I). The residues of eponymic motif II, the DEAH motif, do not exhibit any direct interaction with the ADP molecule. Instead of interacting with the ADP molecule, Asp218 coordinates water molecule 4, whereas Glu219 participates in the coordination of water molecules 2, 3 and 4. At least in the ADP-bound state, Ala220 and His221 do not form any contacts with the bound nucleotide or a water molecule. Moreover, only one residue (Arg435) of motif VI, which comprises eight amino acids, interacts with the bound nucleotide. Here, the lack of interactions might suggest pronounced conformational changes of the RecA-like domains of Prp43 between the ADP-bound and ATP-bound states, which can be also caused by RNA substrate or G-patch protein binding. In contrast to DEAD-box helicases, the conformational changes of which have been extensively studied (for reviews, see Hilbert et al., 2009 ▸; Ozgur et al., 2015 ▸), the conformational flexibility regarding the relative position and orientation of the two RecA-like domains appears to be restricted by the additional C-terminal domains of the DEAH-box proteins.

Figure 3.

Conserved motifs of ctPrp43. The conserved binding motifs of ctPrp43 are presented in blue (nucleotide binding), red (nucleic acid substrate binding) and green (coupling of NTP hydrolysis to substrate unwinding). The N-terminal extension and the RecA1 domain are visualized in light grey, the RecA2 domain in dark grey and the three C-terminal domains in pale blue. The ADP molecule is coloured according to Fig. 1 ▸.

Figure 4.

ADP-binding site of ctPrp43. The ADP molecule is sandwiched between the RecA1 (green) and RecA2 (violet) domains. C atoms are shown in yellow/green/violet, N atoms in blue, O atoms in red, P atoms in orange, the Mg²⁺ ion in light green and water molecules in cyan. Residues which are involved in ADP, water or Mg²⁺ binding are presented in ball-and-stick mode and are labelled according to the ctPrp43 sequence. Polar interactions are visualized as dashed black lines. (a) The adenine moiety is bound via π-electron stacking and the ribose by hydrogen bonding. (b) The α- and the β-phosphates participate intensively in hydrogen bonding. The central Mg²⁺ ion is coordinated by four water molecules, the Thr126 side chain and an O atom of the β-­phosphate. The residues Asp218 and Glu219 of motif II are involved in the coordination of water molecules and do not interact directly with the bound nucleotide. For more detailed information, see §3.2 and Supplementary Table S1 for bonding distances.

3.3. ctPrp43 can functionally replace scPrp43  

Prp43 from C. thermophilum exhibits high sequence similarity and identity to the homologues from Homo sapiens (66.2 and 56.5%, respectively) and S. cerevisiae (77.6 and 68.1%, respectively). The explicit degree of conservation of surface-exposed residues between ctPrp43 and scPrp43 is shown in Fig. 5 ▸, which illustrates that the RecA1, RecA2 and WH domains are especially highly conserved (each domain exhibits about 90% sequence similarity and 80% identity; for the exact values, see Supplementary Table S2), while the level of conservation is lower in the ratchet and the OB-fold (approximately 70 and 55%, respectively). Moreover, the crystal structure of ctPrp43 superposes very well on that of full-length scPrp43 (PDB entry 2xau) after the removal of its first 57 N-terminal residues, the corresponding residues to which are missing in our crystal structure (see Fig. 6 ▸). R.m.s.d. values of 0.90 Å (566 C^α) for chain A and 0.85 Å (574 C^α) for chain B were calculated after aligning both structures.

Figure 5.

Surface conservation between ctPrp43 and scPrp43. The surface conservation was mapped onto ctPrp43(61–764) using AL2CO (Pei & Grishin, 2001 ▸) after structure alignment of the ctPrp43 and scPrp43 sequences in T-Coffee (Notredame et al., 2000 ▸). ctPrp43 is presented as front (left) and back (right) views. Highly conserved regions are coloured magenta, poorly conserved regions light blue and unaligned residues pale wheat.

Figure 6.

Superposition of C. thermophilum and S. cerevisiae Prp43. ctPrp43 is coloured as in Fig. 1 ▸. scPrp43 (PDB entry 2xau) is shown in pale wheat. The superposition was calculated for chain B of scPrp43 with 574 common C^α atoms and an r.m.s.d. of 0.85 Å. The front view (left) and side view (right) are presented. The region of scPrp43 which does not superpose on our structure is the N-terminal extension, which is missing in our truncated construct.

Owing to the high sequence identity and the conservation of a large number of residues located on the surface as well as the virtually identical structure, we wanted to analyse whether ctPrp43 is capable of functionally replacing its yeast homologue in spliceosomal disassembly assays. Proving that ctPrp43 is the authentic orthologue of scPrp43 would increase the importance of spliceosomal RNA helicases from C. thermophilum for further crystallographic studies and for the determination of their exact modus operandi. The disassembly assays were carried out with purified yeast intron–lariat spliceo­somes (ILSs) and recombinant ctPrp43 according to Fourmann et al. (2013 ▸). When ILSs were incubated solely with ATP and subsequently fractionated on a glycerol gradient, only 5% of intron–lariat RNA was released (Fig. 7 ▸ a). This indicates that ILSs are stable complexes and only a minor amount of RNA is released upon gradient centrifugation. The addition of ctPrp43 to ILSs lead to a slight increase in disassembly (22%; Fig. 7 ▸ b), but in the presence of its spliceosomal activator protein Ntr1 ctPrp43 is now able to dissociate larger amounts of ILSs (45%; Fig. 7 ▸ c). For these experiments, the heterodimer of Ntr1·Ntr2 from yeast was utilized, which indicates that ctPrp43 is also able to interact with Ntr1 from S. cerevisiae. Ntr2 is required to recruit Prp43 to its target substrate by binding to Brr2, which is part of the U5 snRNP. Finally, to demonstrate that our employed construct of ctPrp43, which lacks 60 N-terminal amino acids, is still fully functional, we also performed a spliceosome disassembly assay for this variant. Interestingly, the truncated construct leads to the highest rates of ILS disassembly in the presence of the scNtr1·scNtr2 dimer (70%, Fig. 7 ▸ d) compared with the wild-type protein, which suggests involvement of the N-terminal extension in Prp43 regulation.

Figure 7.

In vitro yeast spliceosomal disassembly assay. The disassembly of the intron–lariat spliceosome was analyzed as described previously (Fourmann et al., 2013 ▸). (a) The negative control without ctPrp43 revealed 5% dissociated spliceosomes. (b) In the presence of ctPrp43 and ATP 22% were disassembled. (c) The addition of the activator heterodimer scNtr1–scNtr2 increased the amount of disassembly events to 45%. (d) The N-terminally truncated construct ctPrp43(61–764), which was used for crystallization, exhibited an even higher number of dissociated spliceosomes (70%). Quantification was performed with ImageQuant (Molecular Dynamics). Numbers represent the percentage of intron–lariat RNA released in the top fractions (sum of fractions 1–11) or remaining associated with the ILS (unreleased; sum of fractions 12–22) relative to the intron–lariat RNA distributed in all 22 fractions, the sum of which was set to 100%.

3.4. Thermophilic adaptation of ctPrp43  

C. thermophilum was introduced to the structural biology community in 2011 by Amlacher and coworkers as a thermophilic eukaryote which grows at temperatures of up to 60°C (Amlacher et al., 2011 ▸). We wanted to analyse the thermophilic adaption of ctPrp43 in comparison to its mesophilic counterpart from yeast. For this purpose, we performed CD measurements at 222 nm between 20 and 84°C to monitor the unfolding of α-helices (see Fig. 8 ▸). Owing to the highly similar overall structure, direct comparison of the obtained melting curves is possible. The melting points of ctPrp43 (56°C) and scPrp43 (40°C) differ significantly by 16°C. The molecular basis for the thermophilic adaption of ctPrp43 cannot be easily determined since neither the number of salt bridges nor of hydrogen bonds in ctPrp43 is significantly increased compared with scPrp43 nor does the secondary-structure content or the number of solvent-exposed residues change, which would facilitate entropic stabilization of the protein.

Figure 8.

CD spectroscopic melting-point determination of ctPrp43 and scPrp43. Unfolding of α-helices was assayed at 222 nm between 20 and 84°C. Curves were fitted according to Pace et al. (1998 ▸) and exact melting points were calculated. Far-UV spectra of ctPrp43 and scPrp43 were measured to ensure complete protein denaturation at 84°C (see Supplementary Fig. S3). Residuals indicate smoothing by a factor of two in the corresponding colour to the melting curves.

4. Conclusions  

The presented 2.9 Å resolution crystal structure of Prp43 from C. thermophilum exhibits high structural similarity to the homologue from S. cerevisiae (for r.m.s.d. values for the superposition, see §3.3). Despite an almost identical overall structure, ctPrp43 shows a thermophilic adaption and exhibits a melting temperature which is elevated by 16°C compared with that of its orthologue from yeast. Nevertheless, ctPrp43 is capable of functionally replacing the orthologue from S. cerevisiae in yeast-based spliceosome disassembly assays. Interestingly, the construct we used for crystallization, which lacks the first 60 N-terminal residues, shows an increased capability to dissociate spliceosomes compared with the wild-type ctPrp43, which raises the question of the role of the N-terminal extension in regulation of Prp43. All previously described structural elements, namely the N-terminal extension, the RecA1 and RecA2 domains, the WH domain, the ratchet and the OB-fold (see Fig. 1 ▸), are also identifiable in the orthologue from C. thermophilum as well as all conserved DEAH-box protein motifs (see Fig. 3 ▸; Tanaka & Schwer, 2006 ▸; Walbott et al., 2010 ▸; He et al., 2010 ▸; Cordin et al., 2006 ▸). The structure of ctPrp43 bound to ADP is very likely to be in a post-catalytic state (see §3.2), as suggested, among other reasons, by the low number of contacts formed by several motifs to the bound nucleotide, which have been reported to be crucial for the ATPase activity of Prp43. However, the structure of Prp43 in its activated state with bound ATP, RNA and G-patch proteins is as yet unknown. Owing to this, the mechanism of activation and RNA translocation still remains elusive. The DEAH-box proteins from C. thermophilum might provide an alternative source to tackle the structure determination of their functional complexes.

Supplementary Material

PDB reference: RNA helicase Prp43 from Chaetomium thermophilum bound to ADP, 5d0u