Crystallization and preliminary X-ray diffraction analysis of the phosphate-binding protein PhoX from Xanthomonas citri

Here, the expression, purification and crystallization of the phosphate-binding protein PhoX from Xanthomonas axonopodis pv. citri and X-ray data collection at 3.0 Å resolution are described.

Abstract

Xanthomonas axonopodis pv. citri (X. citri) is an important bacterium that causes citrus canker disease in plants in Brazil and around the world, leading to significant economic losses. Determination of the physiology and mechanisms of pathogenesis of this bacterium is an important step in the development of strategies for its containment. Phosphate is an essential ion in all microrganisms owing its importance during the synthesis of macromolecules and in gene and protein regulation. Interestingly, X. citri has been identified to present two periplasmic binding proteins that have not been further characterized: PstS, from an ATP-binding cassette for high-affinity uptake and transport of phosphate, and PhoX, which is encoded by an operon that also contains a putative porin for the transport of phosphate. Here, the expression, purification and crystallization of the phosphate-binding protein PhoX and X-ray data collection at 3.0 Å resolution are described. Biochemical, biophysical and structural data for this protein will be helpful in the elucidation of its function in phosphate uptake and the physiology of the bacterium.

1. Introduction  

Periplasmic binding proteins from ATP-binding cassette transporters (ABC transporters) are essential components owing to their role in substrate uptake and delivery to the transporter. They consist of two domains connected by a hinge that form a cleft for substrate binding (Davidson et al., 2008 ▶). More than 100 structures of periplasmic binding proteins have been solved in different conformations, essentially demonstrating the changes that occur after ligand binding (Berntsson et al., 2010 ▶) and the essential residues for substrate binding and transport. In bacteria, the PstSABC system is the ABC transporter specific for inorganic phosphate (P_i), which contains the PstS protein as a model high-affinity phosphate-binding protein. At present, 18 structures of phosphate-binding proteins are available in the Protein Data Bank, including 12 structures of PstS from Escherichia coli, including the wild type and mutants (Wang et al., 1994 ▶, 1997 ▶; Yao et al., 1996 ▶; Hirshberg et al., 1998 ▶; Ledvina et al., 1998 ▶), structures of the orthologues from Yersinia pestis (Tanabe et al., 2007 ▶), Streptococcus pneumoniae (PDB entry 4exl; Center for Structural Genomics of Infectious Diseases, unpublished work), Borrelia burgdorferi (Brautigam et al., 2014 ▶) and Clostridium perfringens (PDB entry 4gd5; Center for Structural Genomics of Infectious Diseases, unpublished work), and the structures of PstS-1 and PstS-3 from Mycobacterium tuberculosis (Vyas et al., 2003 ▶; Ferraris et al., 2014 ▶). The structures showed similarities, including the two domains forming the cleft for the phosphate, conservation in the ligand-binding pocket and the importance of residues for high-affinity binding. The expression of pst genes is regulated by low phosphorus levels through the activation of the Pho regulon proteins (Van­Bogelen et al., 1996 ▶), which induces an adaptative response into the cell (Lamarche et al., 2008 ▶). Indirectly, the upregulation and downregulation of many genes mediated by phosphate levels has been correlated with bacterial virulence and pathogenicity, including the pstS genes (Crépin et al., 2011 ▶).

Curiously, in Xanthomonas axonopodis pv. citri (X. citri), the causative agent of citrus canker disease (Brunings & Gabriel, 2003 ▶), we identified the presence of two genes encoding phosphate-binding proteins, pstS (XAC1577) and phoX (XAC1578), which belong to the pstSCABU and oprOphoX operons, respectively (Pegos et al., 2014 ▶). Sequence analysis of the proteins revealed a sequence identity of 65%. A sequence identity of greater than 50% was obseved when the sequence of PhoX was compared with those of the PstS orthologues from E. coli, Y. pestis and S. pneumoniae (Table 1 ▶). PhoX also conserved the residues forming the putative active site found in these proteins. In previous studies, we have demonstrated that phoX is found in a large number of Xanthomonas strains and is identified in many species of proteobacteria. Recently, based on a proteomics analysis we have shown that the expression of PstS and PhoX is upregulated during depletion of phosphate, indicating that both proteins play an important function in bacterial growth (Pegos et al., 2014 ▶).

Table 1. Orthologues of X. citri PhoX with solved structures deposited in the Protein Data Bank and sequence comparison.

Protein name	Organism	Sequence identity (%)	Query coverage (%)	E-value	PDB code and reference
Phosphate-binding protein PstS	Yersinia pestis	57	88	2.00 10122	2z22 (Tanabe et al., 2007 ▶)
Phosphate-binding protein	Escherichia coli	51	88	4.00 10121	2abh (Yao et al., 1996 ▶)
Phosphate-binding protein	Escherichia coli	56	88	4.00 10121	1ixh (Wang et al., 1997 ▶)
Periplasmic binding protein	Streptococcus pneumoniae Canada	55	25	2.00 107	4exl (Center for Structural Genomics of Infectious Diseases, unpublished work)
Phosphate receptor PstS-1	Mycobacterium tuberculosis	32	91	6.00 1038	1pc3 (Vyas et al., 2003 ▶)
Phosphate-binding protein PstS-3	Mycobacterium tuberculosis	31	91	7.00 1035	4lvq (Ferraris et al., 2014 ▶)
Phosphate-binding protein	Clostridium perfringens	27	59	4.00 107	4gd5 (Center for Structural Genomics of Infectious Diseases, unpublished work)
Phosphate-binding protein PstS	Borrelia burgdorferi	21	75	2.10 102	4n13 (Brautigam et al., 2014 ▶)

In order to obtain structural information for the PhoX protein from X. citri, the recombinant mature protein fused to an N-terminal His₆-tag sequence was expressed in E. coli Tuner (DE3) cells with an apparent molecular weight of 33 kDa and was purified using immobilized metal-affinity chromatography. The purified protein was crystallized in different conditions in the presence and absence of phosphate. The crystal structure of X. citri PhoX will be an important achievement for the study of phosphate uptake and regulation in the Xanthomonas genus, including the importance of the ABC transporter PstSCBA in the physiology and virulence of X. citri.

2. Materials and methods  

2.1. Cloning, expression and purification of PhoX protein  

The PhoX-encoding gene was amplified from X. axonopodis pv. citri genomic DNA (gi:21242328) using the following oligonucleotides: F_PhoX_NdeI, 5′-CATATGGCCGACGTCACCGGTG-3′, and R_PhoX_XhoI, 5′-CTCGAGTTAGTGCGGCAGATTCTTGGC-3′. The amplified fragment of 951 bp consisted of the sequence encoding the mature protein without 24 amino acids from the N-terminus flanked by the restriction enzymes NdeI and XhoI at the beginning and end of the gene, respectively, and was cloned into the pET-28a vector. Recombinant X. citri PhoX protein fused at the N-terminal end to a His₆-thrombin cleavage site tag was obtained after expression tests in E. coli Tuner (DE3) strain transformed with a pET-28 derivative carrying the phoX gene (pET_phoX). Induction of expression occurred when the cells reached an optical density at 600 nm of 0.5 after the addition of 0.1 mM IPTG for 4 h at 30°C. Cell extracts were obtained as previously reported for the alkanesulfonate-binding protein SsuA (Araújo et al., 2013 ▶) and the protein was purified by immobilized nickel-affinity chromatography with 150 mM imidazole in the purification buffer (10 mM Tris–NaCl pH 8.0, 500 mM NaCl). The purified fractions were concentrated and dialyzed against 10 mM Tris–HCl pH 8.0 and kept at −20°C at a final concentration of 10 mg ml⁻¹.

2.2. Protein crystallization  

Samples of recombinant X. citri PhoX were submitted to crystallization trials at the automatized crystallization facility at the Brazilian National Laboratory of Biosciences (LNBio; CNPEM, Campinas, SP, Brazil) using the hanging-drop vapour-diffusion method, mixing equal volumes (1:1 µl) of protein and reservoir solution and incubating at 290 K. Initial screening with protein at 10 mg ml⁻¹ was performed using commercial crystallization screening kits from Hampton Research and Jena Biosciences. Conditions that showed crystalline structures or needles were refined, varying the pH, precipitant concentration and protein concentration to yield suitable crystals. A list of all of the conditions that allowed crystal growth and refinement is shown in Table 2 ▶ and crystals are shown in Fig. 1 ▶.

Table 2. Crystallization screening and refinement conditions for X. citri PhoX in the absence and presence of phosphate.

The best crystals obtained after refinement are shown in Fig. 1 ▶.

Sample	Protein concentration (mgml1)	Screening	Refinement	Crystals	Resolution after refinement ()
PhoX + sodium phosphate	10	25% PEG 3350, 200mM sodium iodide, 100mM bis-tris pH 5.5	20% PEG 3350, 200mM sodium iodide	Fig. 1 ▶(a)	3.0
PhoX	8	20% PEG 3350, 200mM sodium iodide	17% PEG 3350, 200mM sodium iodide	Fig. 1 ▶ (b)	5.0
PhoX	10	20% PEG 3350, 200mM sodium iodide	17% PEG 3350, 250mM sodium iodide	Fig. 1 ▶(c)	6.0
PhoX + sodium phosphate	15	20% PEG 3350, 200mM sodium iodide	17% PEG 3350, 250mM sodium iodide	Fig. 1 ▶(d)	2.9
PhoX + sodium phosphate	15	17% PEG 3350, 250mM sodium iodide	15% PEG 3350, 250mM sodium iodide	Fig. 1 ▶(e)	3.0

Figure 1.

Crystals of X. citri PhoX obtained in the absence and presence of phosphate (see Table 2 ▶ for details).

2.3. Data collection and processing  

PhoX crystals were mounted in nylon loops and flash-cooled in mother liquor using liquid nitrogen; the high PEG concentration was sufficient to provide cryoprotection. The PhoX data set was collected on the W01B-MX2 protein crystallography beamline at Laboratório Nacional de Luz Síncrotron (LNLS), Brazil using a wavelength of 1.459 Å and a MAR Mosaic 225 CCD detector. Diffraction data were indexed and integrated using XDS (Kabsch, 2010 ▶) and merged and scaled using AIMLESS (Evans, 2011 ▶) from the CCP4 package (Winn et al., 2011 ▶). Molecular replacement was performed with PHENIX (Adams et al., 2010 ▶) using PstS from E. coli K12 (PDB entry 2abh; Yao et al., 1996 ▶), which shares 51% amino-acid identity with PhoX, as a model. The phosphate, water molecules, H atoms and atoms with zero occupancy were removed from the 2abh structure to provide a better search model. Data-collection statistics are shown in Table 3 ▶.

Table 3. Data-collection and processing statistics.

Values in parentheses are for the highest resolution shell.

Diffraction source	W01B-MX2, LNLS
Wavelength ()	1.459
Temperature (K)	110
Detector	MAR Mosaic 225 CCD
Crystal-to-detector distance (mm)	120
Rotation range per image ()	0.5
Total rotation range ()	164.5
Exposure time per image (s)	30
Space group	P21
a, b, c ()	81.56, 115.60, 133.30
, , ()	90.00, 90.33, 90.00
Mosaicity ()	0.45
Resolution range ()	87.363.00 (3.113.00)
Total No. of reflections	131283 (12689)
No. of unique reflections	46810 (4633)
Completeness (%)	94.3 (95.5)
Multiplicity	2.8 (2.7)
I/(I)	8.4 (2.5)
R meas	0.145 (0.554)
Overall B factor from Wilson plot (2)	39.95

3. Results and discussion  

PhoX was produced from E. coli cells in a soluble form and purified in a single step of purification using immobilized metal-affinity chromatography (Fig. 2 ▶). The purified protein was submitted to crystallization trials in the absence and presence of phosphate, resulting in needles and very small crystals that were less than 15 µm in the longest dimension. Most of the crystals were grown in PEG 3350 and sodium iodide. Refinement of these conditions improved the size and shape of the crystals, especially reducing the percentage of PEG 3350 and increasing the concentration of sodium iodide (Table 2 ▶). The presence of phosphate also increased the hardness of the crystals, suggesting a possible rearrangement of the structure upon ligand binding. The best diffraction pattern was obtained from a crystal grown in 200 mM sodium iodide, 20% PEG 3350 (Fig. 3 ▶). This crystal diffracted to a maximum resolution of 3.0 Å and the data were processed using XDS (Kabsch, 2010 ▶). The crystals showed the symmetry and systematic absences of the primitive monoclinic space group P2₁. All data-collection statistics are shown in Table 3 ▶.

Figure 2.

Expression and purification of X. citri PhoX. PhoX was expressed from E. coli Tuner (DE3) cells after IPTG induction and was purified by immobilized metal-affinity chromatography using nickel columns. Lane MW, molecular-weight markers; lane E, cell extract; lane P, cell pellet; lane FT, flowtrough of the immobilized nickel-affinity chromatography; lane 1, eluted samples from washing with 5 mM imidazole; lanes 2 and 3, eluted samples using 50 and 150 mM imidazole, respectively.

Figure 3.

Crystal and diffraction pattern of X. citri PhoX. (a) Crystal of PhoX obtained in 20% PEG 3350, 200 mM sodium iodide and sodium phosphate with protein at 10 mg ml⁻¹. (b) Diffraction pattern of the crystal to 3.0 Å resolution (circle).

The range of possibilities for the Matthews coefficient (Matthews, 1968 ▶) of between 1.83 and 3.30 ų Da⁻¹ indicated the presence of between nine and five protein molecules in the asymmetric unit, respectively. After molecular replacement and analysis of the data, the best number of molecules in the asymmetric unit with no packing clashes was eight, for which the Matthews coefficient was calculated to be 2.01 ų Da⁻¹ and the solvent content to be 40.3%. The molecular-replacement procedure was performed using the E. coli phosphate-binding protein structure modified by CHAINSAW (Stein, 2008 ▶) as a search model (Yao et al., 1996 ▶; PDB entry 2abh) in Phaser (McCoy et al., 2007 ▶). The initial maps produced are encouragingly good, clearly showing electron density for the phosphate. The model is being rebuilt and refined.