Crystal Structure of Bacillus subtilis IolI Shows Endonuclase IV Fold With Altered Zn Binding

Introduction

Myo-Inositol is commonly found in nature, particularly in soil. Various microorganisms, including soil bacteria, are able to grow on inositol as the sole source of carbon.¹ These findings suggest that myo-inositol catabolism might be conserved among these different microorganisms, although the molecular genetics of the many genes involved have not yet been well characterized.¹ In B. subtilis genome, the inositol catabolism pathway consists of several genes including iolI. The iolI gene (NCBI accession P42419) codes for a conserved 278-residue protein of unknown function that shows no sequence homology with proteins of known structure [cluster 2500 in PROTOMAP, www.protomap.cs.huji.ac.il]. Therefore, it was suggested that this protein may display a previously unobserved structural fold. There are sequence homologies to other structurally undetermined proteins, predominately from the B. subtilis family, including one in the inositol catabolism pathway (iolH). PSI-BLAST² analysis revealed 335 sequence homologs of IolI in bacteria, archaea, and eukaryota (using inclusion threshold 0.005). As a part of the Midwest Center for Structural Genomics (MCSG) initiative (www.mcsg.anl.gov), we have determined the crystal structure of IolI protein at 1.6 Å resolution by using a semiautomated, high-throughput approach. After acquisition of the crystallographic data, structure solution and refinement were completed by using <6.5 h of CPU time. Rapid data acquisition and fast structure determination capabilities are essential components of the high-throughput structural genomics programs. Analysis of the IolI crystallographic structure, however, reveals that the protein adopts a beta-barrel (TIM) configuration³ [Fig.1(a)], suggesting structural homology to both endonuclease IV,⁴ a DNA repair enzyme, and xylose isomerase,⁵ a sugar-metabolizing enzyme (PDB accession 1QTW and 4XIS, respectively).

Fig. 1.

Structure of IolI protein and its metal-binding site. a: Ribbon diagram showing the spatial arrangement of α-helices and β-sheets comprising beta barrel conformation adopted by IolI protein. Side-chains of conserved residues are shown, including those coordinating Zn ion (green). View is approximately perpendicular to the axis of the barrel. b: 2Fo-Fc map of IolI protein in the presence of 2 mM ZnCl₂. c: 2Fo-Fc map of IolI protein in the presence of 2 mM MnCl₂. A water molecule (H₂O92) is residing in roughly the same location as the zinc atom of Figure 1(b).

Materials and Methods

The open reading frame of B. subtilis IolI protein was amplified from genomic DNA with Pfx DNA polymerase by using conditions and reagents provided by the vendor (Invitrogen, Carlsbad, CA). The gene was cloned into the pET30XaLIC (Novagen, Madison, WI) by using the Ligation Independent Cloning protocol.⁶ The resulting expression clone expressed a fusion protein with a thrombin-cleavable His-tag and a factor Xaclevable S-tag. The selenomethionine (SeMet) derivative protein was overexpressed in E. coli BL21-Gold(DE3) (Stratagene) as in Ref. 9. The cells were lysed in binding buffer (300 mM NaCl, 5% glycerol, 50 mM Na phosphate, pH 8.0, 10 mM imidazole, and 10 mM β-mercaptoethanol) by sonication after the addition of lysozyme to 1 mg/mL and 1 mM PMSF. The lysate was clarified by centrifugation, passed through a 0.2-µm filter and applied to Ni-NTA Superflow resin (Qiagen). The protein was eluted from the column with 250 mM imidazole and dialyzed into thrombin cleavage/capture buffer A (20 mM Tris-HCl, pH 8.4, 0.15 mM NaCl, 2.5 mM CaCl₂). The His-tag was cleaved from the protein by treatment with biotinylated thrombin by following the manufacturer’s protocol (Novagen) The cleaved protein was then purified from the cleaved His-tag and uncleaved protein by passing the mixture through a Ni-NTA column. SeMet-labeled protein was purified by using identical protocol.

Crystals of SeMet-derivitized IolI protein were grown in hanging drops containing 1.50–1.75 M ammonium sulphate, 0.050 M Tris buffer, pH 8.00, 0.10 M sodium chloride, 1.0 mM ethylenediaminetetraacetic acid (EDTA), and 3.0 mM dithiothreitol (DTT). A single crystal of approximately 0.3 × 0.3 × 0.2 mm was cryoprotected by using 25% w/v sucrose made up in reservoir solution, and flash-frozen in liquid nitrogen. The crystal belonged to orthorhombic space group P2₁2₁2, with cell dimensions a = 74.6 Å, b = 104.7 Å, c = 48.2 Å,α = β = γ = 90°. The absorption edge of Se was determined by using a fluorescence scan of the IolI crystal, followed by examination of the fluorescence data using CHOOCH.⁸ A three-wave-length MAD data set was collected in < 1 hour using the Structural Biology Center 19ID beamline of the Advanced Photon Source (APS), Argonne National Laboratory (ANL), and data were processed in near real time by using the HKL2000 suite.⁸ Diffraction data extended to 1.6 Å resolution. All six Se atoms in asymmetric unit were found by using SOLVE¹⁰ and CNS¹¹ programs. The 1.6 Å MAD map produced by CNS was of superior quality and allowed autotracing of 276 residues with wARP¹² and remaining two residues with QUANTA.¹³ Final refinement was completed by using CNS (annealing, water molecule identification, individual isotropic B refinement). The entire process of data processing, MAD phasing, autotracing, and refinement took a total of 6.5 h CPU time. Crystal characteristics, data collection, and structure solution results and structure refinement statistics are shown in Table I.

TABLE I.

Summary of IolI Crystal Data, MAD Data Collection, and Refinement

Unit cell parameters (angstroms, degrees)		a = 74.29 Å, b = 104.77 Å, c = 48.42 Å,   α = β = γ = 90
Space group		P 21212 (#18)
Molecular weight [278 residues (SeMet)]		31,937 Da
Molecules per asymmetric unit (a.u.)		1
Selenomethionine residues per a.u.		6
MAD data		Edge energy	Peak energy		High energy
Wavelength (Å)		0.9791	0.9793		0.9466
Resolution limit (Å)		1.60	1.60		1.60
No. of unique reflections:		49,179	49,182		49,100
Overall data completeness (%)		96.9	96.9		96.7
Overall data redundancy		6.35	7.22		7.34
Overall Rmerge (%)		9.0	8.4		6.7
Figure of merit (FOM)			0.709
Phasing power		2.43	2.85		1.60
	Resolution range (Å)		50.0–1.6
	No. of reflections (all)		96,476
	No. of reflections (observed)		90,306
	Percent reflections observed		93.6
	σ Cutoff		0.0
	Overall R-value (%)		0.213
	Free R-value (%)		0.217
	RMSD from ideal geometry
	Bond length		0.006
	Angle		1.5
	Dihedral		21.8
	Improper		0.75
	No. of protein non-hydrogen atoms		2,229
	No. of water molecules		182
	Mean B-factor (Å 2)		12.5
	Ramachandran plot statistics (%)
	Residues in most favored regions		92.5
	Residues in additional allowed regions		7.1
	Residues in generously allowed regions		0.4
	Residues in disallowed region		0.0

IolI protein shows structural homology with E. coli endonuclease IV and xylose isomerase. All three proteins show a TIM barrel fold and bind metal ions. However, there are striking differences in the metal-binding pocket. Endonuclease IV binds three Zn⁺² ions,⁴ whereas xylose isomerase binds two Mn⁺² ions.⁵ To test the metal-binding properties of the IolI protein, crystals were soaked for 2 h in 2 mM ZnCl₂ or 2 mM MnCl₂. Single wavelength data sets were collected to 1.8 Å resolution. The resulting structures revealed a single Zn ion bound via residues Asp174, His177, His200, and Glu246 (Fig. 1, Table II). The apparent inability to bind Mn⁺² ions suggests the metal-binding region of the IolI protein better resembles endonuclase IV catalytic site. However, IolI binds only one Zn ion (site 3) and apparently is unable to bind Zn at sites 1 and 2. The superposition of the two proteins shows that endonuclease IV’s His69 residue coordinating Zn in site 1 has been replaced with Asn66 in IolI, whereas the Zn2 site has His231 replaced with Arg217. Analysis of the Zn3-binding site, however, shows that all four zinc-binding residues are preserved in both endonuclease IV and IolI; it is this site that actively binds a single zinc ion in IolI protein (Fig. 1). Therefore, introduction of just two mutations, His66→Asn and His217→Arg, can explain the lost of the Zn1 and Zn2 binding. The substitution of critical binding-pocket residues at zinc sites 1 and 2 suggest IolI may have a different biochemical function than endonuclase IV. In fact, among IolI sequence homologs, several sugar/alcohol-metabolizing enzymes are found such as hexulose-6-phosphate isomerase, sugar-phosphate isomerase, D-tagatose 3-epimerase, 4-hydroxyphenylpyruvate dioxygenase, hydroxy-pyruvate isomerase. Our data show that catalytic sites of proteins can undergo significant protein sequence modifications that alter the metal ion-binding properties. This may have significant role in evolution of protein function.

TABLE II.

Comparison of Zinc-Binding Site Residues in E. coli Endonuclease IV and B. subtilis IolI

Endonuclease IV (E. coli)	IOLI (B. subtilis)
His69 Zn1	Asn66
Glu145 Zn1	Glu142
Asp229 Zn2	Asp216
His231 Zn2	Arg217
Asp179 Zn3	Asp174Zn1
His182 Zn3	His177 Zn1
His216 Zn3	His200 Zn1
Glu261 Zn3	Glu246Zn1