Recombinant production, purification and crystallization of the Toxoplasma gondii coronin WD40 domain

The T. gondii coronin WD40 domain with the conserved region was recombinantly produced, purified and crystallized. A diffraction data set was collected to 1.65 Å resolution.

Abstract

Toxoplasma gondii is one of the most widely spread parasitic organisms in the world. Together with other apicomplexan parasites, it utilizes a special actin–myosin motor for its cellular movement, called gliding motility. This actin-based process is regulated by a small set of actin-binding proteins, which in Apicomplexa comprises only 10–15 proteins, compared with >150 in higher eukaryotes. Coronin is a highly conserved regulator of the actin cytoskeleton, but its functions, especially in parasites, have remained enigmatic. Coronins consist of an N-terminal actin-binding β-propeller WD40 domain, followed by a conserved region, and a C-terminal coiled-coil domain implicated in oligomerization. Here, the WD40 domain and the conserved region of coronin from T. gondii were produced recombinantly and crystallized. A single-wavelength diffraction data set was collected to a resolution of 1.65 Å. The crystal belonged to the orthorhombic space group C222₁, with unit-cell parameters a = 55.13, b = 82.51, c = 156.98 Å.

1. Introduction  

Toxoplasma gondii is an apicomplexan parasite that is one of the most successful parasitic organisms in the world, able to infect and multiply in a wide variety of warm-blooded animals. Approximately one-third of the human population is infected with T. gondii, but it only causes a disease, toxoplasmosis, in the developing foetus and in individuals suffering from a weakened immune defence. T. gondii can sexually reproduce only in cats, which are the primary host of the parasite (Innes, 2010 ▶). All members of the Apicomplexa phylum are obligate parasites, many of which are of medical or economic importance because they cause disease in either humans or cattle. The most notorious members of the phylum are the Plasmodium species, which cause malaria, one of the deadliest diseases affecting mankind.

Apicomplexan parasites share a common method of actin-dependent cellular movement, called gliding motility, which is also thought to be closely linked to host-cell entry (Dobrowolski & Sibley, 1996 ▶). Gliding requires extraordinarily short actin filaments, located beneath the parasite plasma membrane, that are linked to transmembrane receptors that recognize surface molecules on the host cell (reviewed in Sattler et al., 2011 ▶). No specialized motor organelles or visible changes in the cell shape are involved (Håkansson et al., 1999 ▶). Despite the importance of actin in gliding, the parasite cytoplasm is practically free of microfilaments (Schüler & Matuschewski, 2006 ▶). Recently, it has been suggested that, as opposed to gliding, host-cell entry may not be directly dependent on actin or other components of the gliding machinery (Andenmatten et al., 2013 ▶; Meissner et al., 2013 ▶), which is a partly controversial result and calls for a re-evaluation of many of the current mechanistic hypotheses concerning host-cell invasion by apicomplexan parasites.

Actin dynamics and regulation in apicomplexan parasites differ significantly from higher eukaryotes. The set of actin-interacting regulatory proteins is limited to ∼10–15 in Apicomplexa compared with at least 150 in higher eukaryotes. Furthermore, some classical regulators, mainly nucleation and branching factors, such as Arp2/3, Wiskott–Aldrich syndrome protein (WASp) and WASp-homology 2 (WH2) domains, are missing in apicomplexan parasites (Sattler et al., 2011 ▶). The apicomplexan actins are among the least conserved of all actins and have peculiar polymerization properties. These actins form only very short filaments, and T. gondii actin has been proposed to polymerize in an isodesmic manner, which is a fundamental difference from all other actins (Skillman et al., 2013 ▶). The set of regulatory proteins must have evolved hand in hand with actin, and the apicomplexan actin-binding proteins have divergent sequences, structures and functions compared with canonical actin regulators (Sattler et al., 2011 ▶).

Coronin is one of the few actin regulatory proteins present in apicomplexan parasites. Coronin was originally isolated from Dictyostelium discoideum (de Hostos et al., 1991 ▶) and is a WD-repeat protein that typically contains a seven-bladed β-propeller domain in its N terminus (Appleton et al., 2006 ▶), followed by a conserved region and a unique region, including a coiled-coil domain, in the C-terminal end (Uetrecht & Bear, 2006 ▶). The β-propeller domain contains the primary binding site for F-actin, and the C-terminal domains mediate putative interactions with other actin-regulatory proteins and actin as well as self-assembly. The coiled-coil domain has also been predicted to have a secondary, lower-affinity binding site for actin (Gandhi et al., 2009 ▶). To date, there are only two crystal structures from murine coronin 1 available in the Protein Data Bank (PDB), and both of them lack major parts of the unique region and the coiled-coil domain (Appleton et al., 2006 ▶). T. gondii coronin shares less than 30% sequence identity with mouse coronin 1A (Fig. 1 ▶).

Figure 1.

Sequence alignment of T. gondii coronin and mouse coronin 1A. The WD40 domain is shown in red, the conserved region in green and the coiled-coil domain in orange. Identical residues are displayed as white letters surrounded by blue boxes.

We have expressed and purified a truncated version of T. gondii coronin (hereafter TgCorΔcc), which contains the WD40 and conserved domains, using baculovirus-infected Sf21 insect cells and purification based on affinity and size-exclusion chromatography. The recombinant coronin was folded, as analyzed by circular-dichroism (CD) spectroscopy, was successfully crystallized, and a high-resolution diffraction data set was collected.

2. Materials and methods  

2.1. Recombinant protein production and characterization  

A synthetic gene (UniProt ID Q5Y1E7) coding for full-length T. gondii coronin (Fig. 1 ▶), codon-optimized for insect-cell expression, was ordered from Eurofins MWG Operon, Germany. Several constructs, including that encoding amino acids 2–392, were cloned into an insect-cell expression vector using a sequence- and ligation-independent method (SLIC; Li & Elledge, 2012 ▶). The pFastBacNKI-his-3c-LIC vector (NKI Protein Facility, the Netherlands) was amplified in DH5-α Escherichia coli cells and purified using a QIAprep Spin Miniprep Kit (Qiagen). The vector was then linearized using the KpnI enzyme. The coding region of each construct was amplified by PCR using the primers described in Table 1 ▶ for TgCorΔcc. DpnI was added to the PCR products to digest the template DNA. The PCR-amplified insert and the linearized vector were purified from an agarose gel using the QIAquick gel extraction kit (Qiagen). To create single-stranded overhangs, the digested vector and inserts were treated separately with T4 polymerase (New England Biolabs) at 296 K for 20 min. The reaction was terminated by adding 25 mM ethylenediaminetetraacetic acid (EDTA), followed by inactivation of the T4 polymerase at 348 K for 20 min. The T4-treated vector was mixed with the insert, and the annealing reaction was performed at 338 K for 10 min, after which the reaction mixture was slowly cooled to 295 K in order to improve the annealing efficiency. The reaction mixture was then used to transform NEB 5-alpha heat-competent cells (New England Biolabs) in order to amplify the plasmid. DNA sequencing at Eurofins MWG Operon was used to check the correctness of the clones after amplification.

Table 1. Recombinant protein production information.

In the primers, the sequence given in upper-case letters denotes the insert-specific sequence, and the lower-case letters denote the overhang required for the SLIC vector. In the protein sequence, the coronin sequence is indicated in bold, and the 6×His tag and the 3C cleavage site are in italics. Thus, the final protein contains an additional GPG in the N terminus after proteolytic cleavage of the 6×His tag.

Source organism	T. gondii
DNA source	Synthetic (codon-optimized for insect cells)
Forward primer	5′-cag gga ccc ggt GCC GAC GCT GTA GAC G
Reverse primer	5′-cga gga gaa gcc cgg tta AGG TTT GAC ACT ACG TCT AC
Cloning vector	pBS II SK(+)
Expression vector	pFastBacNKI-his-3c-LIC
Expression host	Spodoptera frugiperda (Sf21)
Complete amino-acid sequence of the construct produced	MAHHHHHHSAALEVLFQGPG ADAVDVPLIKNLYAEAWKQQYSDLRLSTKQTESCGLAANTEYIAAPWDVGGGGVLGILRLADIGRNPAVAKIKGHTASIQDTNFSPFYRDILATACEDTIVRIWQLPEEVTGTTELKEPIATLTGALKKVLSAEWNPAVSGILASGCFDGTVAFWNVEKNENFASVKFQESLLSAKWSWKGDLLACTTKDKALNIVDPRAAQVVGSVACHDGSKACKCTWIDGLAGRDGHVFTTGFGKMQEREMAIWDTRKFDKPVYHAEIDRGSSPLYPIFDETTGMLYVCGKGDSSCRYYQYHGGTLRSVDAYRSSVPIKNFCFIPKLAVDQMRAEIGRMLKQENGNVLQPISFIVPRKNQDVFQADLYPPAPDVEPSMTAEEWFKGENKAIRRRSVKP

Recombinant TgCorΔcc protein with a hexahistidine tag was expressed in baculovirus-infected insect cells using the protocol described by Bieniossek et al. (2008 ▶). DH10EMBacY (EMBL, Grenoble) cells were used for bacmid generation. Sf21 (Invitrogen) cells were transfected with the TgCorΔcc bacmid, using the FuGene 6 transfection reagent (Promega), and virus particles were harvested after 3 d of infection. The recombinant primary virus was amplified to a high-titre viral stock. Approximately 9 µl (1 × 10⁶ cells) of the high-titre virus were used to infect the insect-cell culture at a cell density of 0.6–0.8 × 10⁶ cells per millilitre. The cells were harvested after 72 h and the cell pellets were stored at 253 K. For purification, the cells were resuspended in ice-cold lysis buffer consisting of 100 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 5 mM β-mercaptoethanol (BME), complemented with 10 mM imidazole, 8 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and the cOmplete Mini protease-inhibitor cocktail (Roche). The cells were disrupted by sonication, after which the cell debris was removed by centrifugation at 277 K.

The His-tagged protein in the clarified lysate was bound to Co²⁺-charged iminodiacetic acid (Co–IDA) agarose (Jena Bioscience) in a gravity-flow column, washed with 20 mM imidazole, 500 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 5 mM BME and eluted with 500 mM imidazole, 100 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 5 mM BME. The pooled fractions containing the fusion protein were concentrated to 2.5 ml, applied onto a PD-10 desalting column (GE Healthcare) to remove imidazole, and eluted with 100 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 5 mM BME. Up to this point, the purification steps were performed at room temperature using cold buffers. The N-terminal 6×His tag was cleaved using HRV 3C protease (Novagen) for 16 h at 277 K. 5–10 µl of the protease in 45–65% glycerol was added to 3.5 ml of the uncleaved 6×His-TgCorΔcc solution from the desalting step. 500 µl of 50% Co–IDA slurry was added to the cleaved protein solution in order to bind the cleaved tag, the protease used, and other remaining contaminants that bound to the Co²⁺ matrix, after which the solution was filtered. Final purification was performed at 277 K by size-exclusion chromatography using a Superdex 200 10/300 GL column (GE Healthcare), equilibrated with 100 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 1 mM tris(2-carboxyethyl)phosphine (TCEP). Pure TgCorΔcc was used freshly for following experiments or stored on ice for a few days.

A CD spectrum of the purified and cleaved TgCorΔcc from 260 to 190 nm was measured at 293 K using an Applied Photophysics Chirascan Plus spectropolarimeter equipped with a thermal control unit (Quantum Northwest, TC125), a direct temperature probe and a 0.5 mm path-length quartz cuvette (Hellma). The spectrum was recorded at a concentration of 0.25 mg ml⁻¹ in 25 mM NaCl, 25 mM ammonium sulfate, 12.5 mM sodium phosphate buffer pH 6.5, 0.25 mM TCEP. The DichroWeb server (Lobley et al., 2002 ▶) was used for secondary-structure determination using the CDSSTR algorithm and set 4 optimized for 190–240 nm as a reference data set (Compton & Johnson, 1986 ▶).

2.2. Crystallization  

Crystallization conditions for TgCorΔcc were screened using the sitting-drop vapour-diffusion method in Swissci MRC 2 96-well plates. Crystallization experiments were set up by manually mixing 0.3 µl of both reservoir solution and protein solution at a concentration of 4–9 mg ml⁻¹, as determined by UV absorption at 280 nm and the extinction coefficient calculated from the TgCorΔcc sequence, and the plates were incubated at 293 K. The following screens were used: PEG/Ion, Crystal Screen Lite (Hampton Research, Aliso Viejo, USA) and ProPlex (Molecular Dimensions, Altamonte Springs, USA). Optimization of promising conditions was performed by screening the effect of various salts with polyethylene glycol (PEG) 3350 or polyethylene glycol monomethyl ether (PMME) 2000 at two different pH values (6.5 and 7.5). Data-collection-quality crystals grew in 200 mM magnesium acetate, 20–25% PMME 2000, 100 mM Tris–HCl pH 7.5 at 293 K. For identifying protein crystals, an X-taLight 100 UV source (Molecular Dimensions, Altamonte Springs, USA) was used to detect the tryptophan fluorescence of protein crystals (Dierks et al., 2010 ▶). Crystallization details are given in Table 2 ▶.

Table 2. Crystallization details of TgCorΔcc.

Method	Sitting-drop vapour diffusion
Plate type	Swissci MRC 2 96-well plates
Temperature (K)	295
Protein concentration (mg ml−1)	8
Buffer composition of protein solution	100 mM NaCl, 100 mM ammonium sulfate, 50 mM sodium phosphate buffer pH 6.5, 1 mM TCEP
Composition of reservoir solution	200 mM magnesium acetate, 25% PMME 2000, 100 mM Tris–HCl pH 7.5
Volume and ratio of drop	0.3 µl protein + 0.3 µl reservoir
Volume of reservoir (µl)	40

2.3. Data collection and processing  

Crystals of TgCorΔcc were flash-cooled in liquid nitrogen after soaking in a cryoprotectant solution with 25%(v/v) glycerol in the reservoir solution. Preliminary X-ray diffraction tests and native data collection to 1.65 Å resolution were performed on the EMBL beamline P14 at PETRA III, Hamburg, Germany. Diffraction images were recorded on a PILATUS 6M detector (DECTRIS Ltd, Switzerland), using a wavelength of 0.976 Å, an oscillation angle of 0.2° per frame and a crystal-to-detector distance of 266.6 mm. The diffraction images were indexed, integrated and scaled using the XDS package (Kabsch, 2010 ▶) and XDSi (Kursula, 2004 ▶).

3. Results and discussion  

TgCorΔcc (394 residues; molecular weight 43.3 kDa) was successfully overexpressed in Sf21 insect cells and purified to homogeneity using Co–IDA agarose in gravity-flow columns followed by size-exclusion chromatography (Fig. 2 ▶ a). TgCorΔcc eluted from the Superdex 200 10/300 GL column at an elution volume corresponding to the molecular weight of a monomeric WD40 domain, which was expected in the absence of the predicted coiled-coil domain presumed to be responsible for coronin self-assembly into dimers or trimers. The identity of the purified protein was verified by peptide-mass fingerprinting using mass spectrometry at the Biocenter Oulu Proteomics Core Facility. The typical yield was approximately 3.5 mg of pure TgCorΔcc from 1 l of Sf21 cells at a density of 0.6–0.8 × 10⁶ cells per millilitre at the point of infection. The cell count normally doubled from that before growth arrest.

Figure 2.

Purification and CD analysis of TgCorΔcc. (a) Size-exclusion chromatogram and Coomassie-stained denaturing gel of the peak fractions eluting at ∼17 ml from a Superdex 200 10/300 GL column. The positions of molecular-weight markers (in kDa) are indicated at the side of the gel. (b) The CD spectrum of TgCorΔcc from 260 to 190 nm measured at 293 K.

The folding and secondary-structure content of the purified TgCorΔcc were analyzed using CD spectroscopy. The CD spectrum indicates the protein to be folded (Fig. 2 ▶ b). TgCorΔcc was estimated to contain 41% β-strands and 22% turns, which is similar to the corresponding region in the homologous murine coronin 1 crystal structure, which consists of 42% β-strands and 15% turns (PDB entries 2aq5 and 2b4e; Appleton et al., 2006 ▶).

Screening for crystallization conditions was performed using commercially available screens. No crystals were obtained directly from the screens, but promising-looking drops with microcrystalline precipitate were found, especially from the PEG/Ion screen. Screening was continued by optimizing these conditions using various salts with PEG 3350 or PMME 2000 at two different pH values, 6.5 and 7.5. Crystals with the longest dimension varying from 50 to 200 µm were obtained from the optimized condition within 7 d (Fig. 3 ▶). A peculiar feature of the crystals was that they all grew using salt crystals as a nucleation point. In the drops, plate-shaped salt crystals appeared immediately after setting up the drops, and approximately after 7 d, additional crystals were detected (Fig. 3 ▶ a). The crystals were observed under UV light, where a clear tryptophan fluorescence signal was visible (Fig. 3 ▶ b). The protein crystals could be separated from the salt crystals for X-ray diffraction measurements. A native data set to 1.65 Å resolution was recorded and used for molecular-replacement trials.

Figure 3.

TgCorΔcc protein crystals nucleate from salt crystals. (a) Visible light image of a TgCorΔcc crystal attached to the surface of a salt crystal. The protein crystal size is approximately 100 × 100 × 50 µm. (b) Corresponding image showing the detection of the protein tryptophan fluorescence signal while illuminating the crystal with UV light.

The TgCorΔcc crystals belong to space group C222₁, with unit-cell parameters a = 55.1, b = 82.5, c = 156.98 Å, α = β = γ = 90°. The Matthews coefficient of 2.06 ų Da⁻¹ with a solvent content of 40.3% (Matthews, 1968 ▶) suggests that the asymmetric unit contains one TgCorΔcc molecule. Data-collection and processing statistics are given in Table 3 ▶. The data quality was evaluated using phenix.xtriage (Adams et al., 2010 ▶) and no pathological behaviour was detected. Structure determination will be started using molecular replacement using the homologous murine coronin 1 structure (PDB entry 2aq5) as a model.

Table 3. Data-collection and processing statistics.

Values in parentheses are for the outer shell.

Diffraction source	P14 at PETRA III/DESY
Wavelength (Å)	0.976
Temperature (K)	100
Detector	PILATUS 6M
Crystal-to-detector distance (mm)	266.6
Rotation range per image (°)	0.2
Total rotation range (°)	200
Space group	C2221
Unit-cell parameters (Å, °)	a = 55.13, b = 82.51, c = 156.98, α = β = γ = 90
Resolution range (Å)	50–1.65 (1.7–1.65)
Total No. of reflections	168873
No. of unique reflections	43134
Completeness (%)	99.2 (99.1)
Multiplicity	3.9
〈I/σ(I)〉†	12 (1.2)
R meas ‡ (%)	7.7 (134)
CC1/2 § (%)	99.9 (35.3)
Overall B factor from Wilson plot (Å2)	29.7

^†The mean I/σ(I) falls below 2.0 at 1.75 Å resolution.

^‡ R _meas is the redundancy-independent R factor as defined by Diederichs & Karplus (1997 ▶) and Weiss & Hilgenfeld (1997 ▶). .

^§CC_1/2 is defined as the correlation coefficient between two random half data sets, as described by Karplus & Diederichs (2012 ▶).

The upcoming structure of TgCorΔcc will be the first crystal structure of a parasite coronin to be determined and only the second for a WD40 domain from the whole coronin protein family. The crystallized domain from T. gondii coronin shares only 30% sequence identity with the corresponding murine coronin 1. Thus, it is expected to reveal valuable information on the diversity of coronin WD40 domains and how, for example, the suggested actin-binding patches located within the WD40 domain of higher eukaryotic coronins differ in apicomplexan coronins.