Expression, purification and preliminary X-ray crystallographic analysis of cyanobacterial biliverdin reductase

Biliverdin reductase (BVR) from Synechocystis sp. PCC6803 and its selenomethionine derivative were overexpressed and purified. X-ray diffraction data from an SeMet BVR microcrystal were collected to 3.0 Å resolution on microfocus beamline BL32XU at SPring-8.

Abstract

Biliverdin reductase (BVR) catalyzes the conversion of biliverdin IX α to bilirubin IX α with concomitant oxidation of an NADH or NADPH cofactor. This enzyme also binds DNA and enhances the transcription of specific genes. Recombinant cyanobacterial BVR was overexpressed in Escherichia coli, purified and crystallized. A native data set was collected to 2.34 Å resolution on beamline BL38B1 at SPring-8. An SeMet data set was collected from a microcrystal (300 × 10 × 10 µm) on the RIKEN targeted protein beamline BL32XU and diffraction spots were obtained to 3.0 Å resolution. The native BVR crystal belonged to space group P2₁2₁2₁, with unit-cell parameters a = 58.8, b = 88.4, c = 132.6 Å. Assuming that two molecules are present in the asymmetric unit, V _M (the Matthews coefficient) was calculated to be 2.37 ų Da⁻¹ and the solvent content was estimated to be 48.1%. The structure of cyanobacterial BVR may provide insights into the mechanisms of its enzymatic and physiological functions.

1. Introduction

Linear tetrapyrrole pigments are widely distributed throughout the kingdoms of life. In phototrophs, linear tetrapyrroles play a central role in both light harvesting and light sensing (Dammeyer & Frankenberg-Dinkel, 2008 ▶; Ikeuchi & Ishizuka, 2008 ▶). The biosynthesis of linear tetraphyrroles begins with haem degradation (Fig. 1 ▶); in phototrophs, the porphyrin ring of haem is cleaved at the α-meso carbon bridge by ferredoxin-dependent haem oxygenases to form biliverdin IX α (Frankenberg-Dinkel, 2004 ▶; Mantle, 2002 ▶), the common precursor of all phytobilins. In cyanobacteria, biliverdin IX α is further reduced to phytobilins by ferredoxin-dependent bilin reductases such as PcyA (Beale, 1993 ▶). The biosynthesis of phytobilin from haem has been extensively studied; we have been involved in studying the molecular mechanisms of cyanobacterial haem oxy­genase (Sugishima et al., 2004 ▶, 2005 ▶) and PcyA (Hagiwara et al., 2006 ▶, 2010 ▶) on the basis of their crystal structures.

Figure 1.

Biosynthesis of linear tetrapyrrole pigments in cyanobacteria (Synechocystis sp. PCC 6803).

In mammals, biliverdin reductase (BVR) catalyzes the reduction of biliverdin IX α using NADH or NADPH to produce the antioxidant bilirubin IX α. Intriguingly, whole-genome analysis revealed that many cyanobacteria possess the bvr gene, which encodes the mammalian BVR homologue. In vitro, recombinant cyanobacterial BVR converts biliverdin IX α into bilirubin IX α using NADH or NADPH as an electron donor (Schluchter & Glazer, 1997 ▶). Disruption of the bvr gene in Synechocytis sp. PCC6803 perturbs the biosynthesis of phycobiliproteins, which include light-harvesting pigments such as phycocyanobilin. Thus, it has been proposed that the reaction product of cyanobacterial BVR plays a regulatory role in phycobiliprotein biosynthesis (Schluchter & Glazer, 1997 ▶), even though bilirubin, which is the product of the in vitro BVR reaction, has never been detected in cyanobacterial cells (Ikeuchi & Ishizuka, 2008 ▶). Interestingly, mammalian BVR is a serine/threonine kinase that is activated by oxygen radicals and translocates into the nucleus in response to cGMP and oxidative stress (Maines et al., 2007 ▶; Salim et al., 2001 ▶). In addition, the dimeric form of human BVR binds to a DNA fragment corresponding to the haem oxygenase 1 promoter region and may play a role in transcriptional activation by oxidative stress (Ahmad et al., 2002 ▶). This finding raises the possibility that cyanobacterial BVR also regulates the transcription of the genes involved in phycobiliprotein biosynthesis in vivo.

The amino-acid sequences of human (accession No. AAH05902), rat (NP_446302) and cyanobacterial (NP_440117) BVRs share ∼60% similarity over the N-terminal 150 residues. In mammals, BVR contains a leucine zipper-like motif (residues 100–160; Ahmad et al., 2002 ▶), a DNA-binding motif that consists of repeats of five residues separated by six amino acids. In contrast, cyanobacterial BVRs do not contain a canonical leucine-zipper motif, although the corresponding region contains periodic leucine residues as well as many basic and hydrophobic residues. Recently, it has been reported that cyanobacterial BVR forms a homodimer (Hayes & Mantle, 2009 ▶), which is consistent with the behaviour of a typical DNA-binding protein.

The crystal structures of the apo forms of rat and human BVR and the NAD⁺- and NADP⁺-bound forms of rat BVR have been determined (Kikuchi et al., 2001 ▶; Whitby et al., 2002 ▶); the configurations of the bound NAD⁺/NADP⁺ and the surrounding residues were also elucidated. The substrate-binding model based on the crystal structure predicted that a conserved Tyr97 in rat BVR might play a central role in the catalytic reaction; however, when this residue was mutated (Y97F) the activity was only reduced by 50% compared with the wild type (Whitby et al., 2002 ▶). The identities of both the catalytic residues and the substrate-binding site remain controversial because the binding configuration of the substrate, biliverdin, remains unknown. The DNA-binding site on BVR also remains unknown.

In order to understand the structural basis for DNA binding, as well as biliverdin binding, of cyanobacterial BVR, we initiated the determination of its structure by X-ray crystallographic analysis. In this paper, we describe the overexpression, purification and crystallization of cyanobacterial BVR, as well as our preliminary X-ray diffraction studies.

2. Materials and methods

2.1. Production of native and selenomethionine-substituted BVR

The slr1784 gene encoding BVR was amplified by PCR from the genomic DNA of Synechocystis sp. PCC6803. NdeI and BamHI restriction sites (shown in bold) were introduced into the oligo­nucleotide primers 5′-CATATGTCTGAAAATTTTGCAGTTG-3′ and 5′-GGATCCGAGCTCAATTTTCAACCTTATATCCACA-3′, respectively. The amplified gene, including the original stop codon, was cloned into pCR2.1-TOPO (Invitrogen) to yield pCR2.1-TOPO/bvr. After confirmation of its DNA sequence, pCR2.1-TOPO/bvr was digested with NdeI and BamHI and cloned into pET21a for nontagged BVR and pET15b (Novagen) for overexpression of the protein with a His₆ tag at its N-terminus.

To overexpress native BVR, Escherichia coli strain C41 (DE3) cells (Miroux & Walker, 1996 ▶) transformed by heat shock with pET21a/bvr were grown at 310 K until the A ₆₀₀ reached 0.6 in Terrific Broth containing ampicillin (50 µg ml⁻¹). Overexpression was con­ducted overnight under shaking at the same temperature after induction by the addition of 1 mM IPTG. Cells were harvested by centrifugation at 3000g for 30 min. E. coli strain B834 (DE3) cells transformed with pET15b/bvr were grown at 310 K to overexpress selenomethionine-substituted (SeMet) His₆-BVR. A 20 ml aliquot of a culture grown overnight at 310 K in LB medium was centrifuged at 5000g for 1 min and the pellet was then resuspended in 1 ml LeMaster medium and added to 5 l of the same medium containing ampicillin (50 µg ml⁻¹) and selenomethionine (100 µg ml⁻¹). Overexpression was conducted for 48 h at the same temperature after induction by the addition of 1 mM IPTG. Cells were harvested by centrifugation at 3000g for 30 min.

2.2. Purification

2.2.1. Purification of native BVR

The bacterial pellet (typically 15 g wet weight) was suspended in 40 ml 50 mM Tris–HCl pH 7.8 (buffer A). Cells were disrupted by sonication and centrifugation (8000g for 30 min at 277 K). The supernatant containing the soluble native BVR was subjected to ammonium sulfate fractionation at 40% saturation and centrifuged (27 000g for 30 min at 277 K). The supernatant was loaded onto a HiPrep 16/10 Butyl FF column (GE Healthcare) equilibrated with 40% ammonium sulfate in buffer A. After the column had been washed with buffer A, the protein was eluted with a decreasing linear gradient of ammonium sulfate (40–0%). Fractions containing BVR were collected and dialyzed against buffer A. The sample was loaded onto a HiLoad 26/10 Q Sepharose HP column (GE Healthcare) equilibrated with buffer A. The column was washed with buffer A and the protein was eluted with an increasing linear gradient of NaCl (0–500 mM). Fractions containing BVR were concentrated using a Vivaspin filter (GE Healthcare) and loaded onto a HiPrep 16/60 Sephacryl S-200 HR gel-filtration column (GE Healthcare) equilibrated with 150 mM NaCl in buffer A. Fractions containing BVR were monitored by SDS–PAGE analysis. The protein concentration was determined by the Bradford assay using BSA as a reference.

2.2.2. Purification of SeMet BVR

The bacterial pellet (typically 20 g wet weight) was suspended in 40 ml 50 mM Tris–HCl pH 7.8, 300 mM NaCl and 20 mM imidazole (buffer B). Cells were disrupted by sonication and centrifugation (8000g for 30 min at 277 K). The supernatant containing the soluble His-tagged protein was mixed with His-Accept resin (Nacalai) and incubated for 30 min at 277 K under rotation. The lysate/resin mixture was washed five times with buffer B. Protein elution was carried out five times with 50 mM Tris–HCl pH 7.8, 300 mM NaCl and 300 mM imidazole. The eluted fractions were collected and analyzed by SDS–PAGE.

Nontagged SeMet BVR was prepared by digestion of His₆-tagged SeMet BVR with 10 units of thrombin (GE Healthcare) per milligram of protein in PBS buffer (10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 2.7 mM KCl and 140 mM NaCl) pH 7.4 for 48 h at 293 K. The nontagged SeMet BVR was separated from the remaining His-tagged protein by mixing it again with His-Accept resin and incubating for 30 min. The flowthrough fraction containing nontagged SeMet BVR and thrombin was concentrated using a Vivaspin filter and loaded onto a HiPrep 16/60 Sephacryl S-200 HR gel-filtration column (GE Healthcare) equilibrated with PBS buffer.

2.3. Crystallization

The purified native and SeMet BVR were concentrated using Vivaspin filters. All crystallization trials were carried out using the hanging-drop vapour-diffusion method at 277 K. Crystallization drops consisting of 1 µl protein solution and 1 µl precipitant solution were equilibrated against 200 µl precipitant solution. The initial trials were performed using commercially available sparse-matrix screening kits: Crystal Screen, Crystal Screen 2, Crystal Screen Lite, PEG/Ion and PEG/Ion 2 (Hampton Research), Wizard I–III (Emerald BioStructures) and JB Screen 1–6 (Jena Bioscience GmbH). The initial conditions that produced crystals were optimized by varying the protein concentration, precipitant, buffer system and pH.

2.4. Data collection

The crystals were transferred into a cryoprotectant solution con­sisting of reservoir solution containing 15%(v/v) glycerol for a few seconds and then flash-cooled in a cryostream at 100 K. A complete data set for native BVR was collected with an ADSC Quantum 4R detector using synchrotron radiation (λ = 1.000 Å) on SPring-8 beamline BL38B1 (Hyogo, Japan). The crystal-to-detector distance was 200 mm and 180 images were recorded at 1° intervals with an exposure time of 15 s per image. The intensity data were processed and scaled using HKL-2000 (Otwinowski & Minor, 1997 ▶).

Owing to the small size of the SeMet BVR crystal, data collection was carried out on the RIKEN targeted protein beamline BL32XU at SPring-8, which is dedicated to data collection from protein microcrystals (Hirata et al., 2010 ▶). The intensity data set for SeMet BVR was collected with a Rayonix MX225HE CCD detector using a beam size of 1 × 10 µm (horizontal × vertical), a wavelength of 0.9740 Å and a crystal-to-detector distance of 220 mm. After the collection of each two sequential images, the beam-irradiation position was shifted by 1.0 µm with inverse-beam geometry using the SPring-8 data-collection software BSS (Ueno et al., 2005 ▶). A total of 508 images were collected with an oscillation angle of 1.0°. The intensity data were processed and scaled using HKL-2000 (Otwinowski & Minor, 1997 ▶).

3. Results

3.1. Purification

When the native form of BVR was overexpressed in E. coli strain C41 (DE3), SDS–PAGE of whole cell extracts showed a major Coomassie Blue-stained band at ∼36 kDa, which is consistent with the mass of BVR estimated from its DNA sequence. Purification of the native BVR yielded ∼15 mg protein, the purity of which was evaluated to be greater than 95% using a Coomassie Blue-stained gel (Fig. 2 ▶).

Figure 2.

SDS–PAGE analysis of the purified cyanobacterial BVR. Lane M, molecular-weight markers (kDa); lane 1, purified native BVR; lane 2, SeMet His₆-BVR; lane 3, purified nontagged SeMet BVR.

We initially attempted to overexpress SeMet BVR in E. coli strain B834 (DE3) using pET21a/bvr, which encodes nontagged BVR. Because the resulting protein level was very low, we could not obtain purified SeMet BVR in a sufficient quantity for crystallization trials. Hence, we overexpressed SeMet BVR with a histidine tag in B834 (DE3) cells in order to achieve efficient purification. Nickel-affinity purification successfully produced nearly homogeneous SeMet His₆-BVR. The first 17 amino acids from the N-terminus of SeMet His₆-BVR were cleaved with thrombin and subsequent purification steps com­pletely removed the N-terminal His-tag region (Fig. 2 ▶). The resulting SeMet BVR contained three additional amino acids (i.e. Gly-Ser-His) at the N-terminus compared with native BVR and three selenomethione residues in each subunit.

3.2. Crystallization

Initial crystallization trials using native BVR produced small crystals in several drops containing polyethylene glycol (PEG) as a precipitant. The crystallization conditions ultimately established consisted of protein at 12 mg ml⁻¹ in 50 mM Tris–HCl pH 7.8 con­taining 150 mM NaCl and a reservoir solution consisting of 15%(w/w) PEG 4000, 50 mM Tris–HCl pH 7.25, 0.2 M sodium acetate and 0.2 mM Cymal-2. Crystals grew to maximum dimensions of 0.1 × 0.1 × 0.5 mm (Fig. 3 ▶ a).

Figure 3.

Typical crystals of cyanobacterial BVR. (a) Crystals of native BVR. Approximate crystal dimensions are 0.1 × 0.1 × 0.5 mm. (b) Microcrystals of nontagged SeMet BVR. Approximate crystal dimensions are 300 × 10 × 10 µm.

Fortunately, microcrystals of nontagged SeMet BVR could be obtained under conditions similar to those that produced the crystals of native BVR. In order to improve the size and shape of the crystals, it was necessary to optimize the crystallization conditions. The best crystals of nontagged SeMet BVR were grown in a drop containing a 1:1 mixture of protein solution (12.7 mg ml⁻¹ in PBS buffer pH 7.4) and reservoir solution [15%(w/v) PEG 4000, 0.1 M Tris–HCl pH 8.5 and 0.2 M sodium acetate]. The crystals grew to maximum dimensions of 300 × 10 × 10 µm in a week (Fig. 3 ▶ b).

3.3. Preliminary X-ray analaysis

An X-ray diffraction experiment using a native BVR crystal was performed on the BL38B1 beamline. The crystal belonged to the orthorhombic space group P2₁2₁2₁, with unit-cell parameters a = 58.8, b = 88.4, c = 132.6 Å. A complete data set was collected to a resolution of 2.34 Å (Table 1 ▶). The unit-cell parameters are compatible with the presence of two BVR molecules per asymmetric unit and a solvent content of 48.1% (V _M = 2.37 ų Da⁻¹; Matthews, 1968 ▶). We attempted to solve the structure of BVR by molecular-replacement methods using apo rat BVR (PDB code 1gcu; 21% sequence identity; Kikuchi et al., 2001 ▶), NADH-bound rat BVR (PDB entry 1lc0; 21% sequence identity; Whitby et al., 2002 ▶) and human BVR (PDB code 2h63; 21% sequence identity; K. Kavanagh, J. Elkins, E. Ugochukwu, K. Guo, E. Pilka, P. Lukacik, C. Smee, E. Papagrigoriou & U. Oppermann, unpublished work) as search models using the programs MOLREP (Vagin & Teplyakov, 1997 ▶), EPMR (Kissinger et al., 1999 ▶) and MrBUMP (Keegan & Winn, 2007 ▶). Unfortunately, all of the molecular-replacement trials were unsuccessful.

Table 1. Summary of crystallographic statistics.

Values in parentheses are for the outermost shell.

	Native BVR	SeMet BVR
X-ray source	BL38B1	BL32XU
Wavelength (Å)	1.0000	0.9740
Space group	P212121	P212121
Unit-cell parameters (Å)	a = 58.8, b = 88.4, c = 132.6	a = 58.4, b = 87.8, c = 133.7
Resolution range (Å)	50–2.34 (2.42–2.34)	25–3.00 (3.05–3.00)
No. of observations	204988	286322
No. of unique reflections	29740	14380
Rmerge† (%)	5.6 (31.8)	9.8 (31.9)
Multiplicity	6.9 (6.7)	19.9 (20.7)
Completeness (%)	99.9 (99.9)	100.0 (100.0)
Mean I/σ(I)	11.9 (7.1)	13.6 (3.5)

^† R _merge = , where 〈I(hkl)〉 is the mean intensity for multiply recorded reflections.

To detect the anomalous signals in the diffraction data, a microcrystal of nontagged SeMet BVR was mounted on beamline BL32XU at SPring-8, where we used a 1 × 10 µm (horizontal × vertical) beam with a photon flux density of 2 × 10¹⁰ photons s⁻¹ µm⁻². Irradiation of the single needle-shaped crystal on the loop using a microbeam with exceptionally high brightness permitted diffraction to 3.0 Å resolution (Fig. 4 ▶). However, this microbeam caused serious radiation damage to the crystal; the intensities of the reflections in the higher resolution shells decreased rapidly after a couple of frames. We overcame this problem by shifting the beam-irradiation position along the needle axis of the crystal by 1.0 µm after the collection of every two images. This change in irradiation position effectively led to migration of the radiation damage and an intensity data set could be collected from one microcrystal (Table 1 ▶). Structure determination is now under way using the SAD and SIRAS methods; we also plan to collect MAD data from the microcrystals.

Figure 4.

X-ray data collection from SeMet BVR microcrystals. (a) A needle-shaped crystal mounted in a loop in a cryogenic stream on SPring-8 beamline BL32XU. Approximate crystal dimensions are 300 × 10 × 10 µm. (b) A typical diffraction frame from an SeMet BVR microcrystal.