The structure and activity of the glutathione reductase from Streptococcus pneumoniae

The three-dimensional crystal structure of the glutathione reductase from Streptococcus pneumoniae is reported at 2.56 Å resolution.

Abstract

The glutathione reductase (GR) from Streptococcus pneumoniae is a flavo­enzyme that catalyzes the reduction of oxidized glutathione (GSSG) to its reduced form (GSH) in the cytoplasm of this bacterium. The maintenance of an intracellular pool of GSH is critical for the detoxification of reactive oxygen and nitrogen species and for intracellular metal tolerance to ions such as zinc. Here, S. pneumoniae GR (SpGR) was overexpressed and purified and its crystal structure determined at 2.56 Å resolution. SpGR shows overall structural similarity to other characterized GRs, with a dimeric structure that includes an antiparallel β-sheet at the dimer interface. This observation, in conjunction with comparisons with the interface structures of other GR enzymes, allows the classification of these enzymes into three classes. Analyses of the kinetic properties of SpGR revealed a significantly higher value for K _m(GSSG) (231.2 ± 24.7 µM) in comparison to other characterized GR enzymes.

1. Introduction  

Glutathione (γ-glutamylcysteinyl-glycine) is a thiol-containing tripeptide that is present at high concentrations (1–10 mM) in eukaryotic and many prokaryotic species (Masip et al., 2006 ▸). Glutathione can exist in a thiol-reduced state (GSH) or in an oxidized state (GSSG), which consists of two GSH molecules linked together by a disulfide bond. GSH serves as a reductive cofactor for the antioxidant glutaredoxin enzymes (Ritz & Beckwith, 2001 ▸), with this activity converting glutathione from the reduced to the oxidized form. The ratio of GSH to GSSG in the bacterial cytoplasm is tightly regulated by the enzyme glutathione reductase (GR), which employs NADPH as the electron donor:

The Gram-positive bacterium Streptococcus pneumoniae (the pneumococcus) is the world’s foremost human bacterial pathogen. It is the leading cause of bacterial pneumonia, which accounts for 15% of all childhood disease mortality (Rudan et al., 2008 ▸, 2013 ▸). The acquisition of nutrients is essential to the ability of the pneumococcus to colonize and to mediate in vivo virulence. Analyses of genome sequences have revealed that S. pneumoniae lacks the genes required for de novo glutathione biosynthesis (Lanie et al., 2007 ▸). Instead, S. pneumoniae acquires GSH (∼10.9 ± 0.3 mM) from the extracellular environment via an ATP-binding cassette (ABC) transporter and the substrate-binding protein GshT (Potter et al., 2012 ▸; Begg et al., 2015 ▸). Isogenic deletion strains lacking either gshT or gor, which encodes GR, showed increased sensitivity to oxidative stress, albeit to differing extents (Potter et al., 2012 ▸). Further, the ΔgshT and Δgor strains were observed to be hypersensitive to intoxication by the divalent metal ions copper, zinc and cadmium (Potter et al., 2012 ▸; Begg et al., 2015 ▸).

Murine models of invasive pneumococcal infection have observed that infection triggers an innate immune response that includes increased levels of reactive oxygen and nitrogen species (ROS/RNS) and significant fluxes in the abundance of host metal ions such as zinc (Zn²⁺; McDevitt et al., 2011 ▸). This latter observation has been implicated in potentially exposing invading bacteria to zinc stress either directly or as a component of phagocytic cell killing (Botella et al., 2011 ▸; Martin et al., 2017 ▸; Eijkelkamp et al., 2014 ▸). Glutathione has been implicated in playing crucial roles in protection against these host-mediated stresses. ROS and RNS can be detoxified by GSH via direct interaction or by enzymatic processes, such as that catalyzed by glutathione peroxidase. Glutathione also forms metal complexes and thereby can contribute to the intracellular metal tolerance of ions such as zinc by forming a GSH–Zn²⁺ complex to moderate the abundance of the uncomplexed ion (Díaz-Cruz et al., 1998 ▸). These protective processes are dependent on cytoplasmic GSH, the concentration of which is dictated by S. pneumoniae GR.

The crystal structures of GR enzymes from Escherichia coli (Mittl & Schulz, 1994 ▸), Saccharomyces cerevisiae (Yu & Zhou, 2007 ▸), Plasmodium falciparum (Sarma et al., 2003 ▸; Böhme et al., 2000 ▸) and Homo sapiens (Schulz et al., 1978 ▸; Karplus & Schulz, 1989 ▸; Savvides & Karplus, 1996 ▸; Berkholz et al., 2008 ▸) have been described in conjunction with their kinetic properties (Mittl & Schulz, 1994 ▸; Savvides & Karplus, 1996 ▸; Yu & Zhou, 2007 ▸). All three enzymes are dimeric and bind one molecule of FAD per subunit. Here, we report the structural and kinetic characterization of the GR enzyme from S. pneumoniae (SpGR). Comparison of the SpGR structure with those from other sources reveals close similarities in the overall structures of these enzymes, with secondary-structural differences apparent at the dimer interface.

2. Materials and methods  

2.1. Macromolecule production  

Recombinant SpGR was generated by PCR-amplification of the SPD_0685 gene from S. pneumoniae serotype 2 strain D39 using ligation-independent cloning and the oligonucleotides GR_LIC1F (5′-TGGGTGGTGGATTTCCTAGAGAATATGATATCATTGCTATCGG) and GR_LIC1R (5′-TTGGAAGTATAAATTTCCACGCATGGTTACAAATTCTTC) to insert the gene into an N-terminal dodecahistidine tag-containing vector, pCAMnLIC01, to generate pCAMnLIC01-GR.

Heterologous expression of recombinant SpGR was performed in E. coli BL21(DE3) cells grown in lysogeny broth at 37°C. The expression of recombinant protein was induced by 0.5 mM IPTG at an optical density (OD) of 0.6. The cells were incubated at 25°C for 20 h and were then harvested by high-speed centrifugation. The cell pellet was suspended in buffer A (20 mM Tris pH 7.5, 150 mM NaCl) and disrupted using a TS series benchtop cell disruptor (Constant Systems). The His-tagged SpGR protein was purified using a two-step protocol consisting of immobilized metal-affinity chromatography (IMAC) followed by size-exclusion chromatography. The cell lysate was applied onto Ni-Sepharose 6 Fast Flow resin (GE Healthcare Life Sciences) equilibrated with buffer A and eluted with buffer A containing 300 mM imidazole. The eluted SpGR protein was incubated with 2 mM FAD for 2 h at 4°C before application onto a HiLoad 16/600 Superdex 200 size-exclusion column (GE Healthcare Life Sciences) pre-equilibrated with buffer A. The purified SpGR protein was concentrated to ∼20 mg ml⁻¹ as determined by a BCA assay (ThermoFisher Scientific) and was stored at −80°C until further use. Macromolecule-production information is summarized in Table 1 ▸.

Table 1. Macromolecule-production information.

Source organism	S. pneumoniae
DNA source	S. pneumoniae serotype 2 strain D39
Forward primer	TGGGTGGTGGATTTCCTAGAGAATATGATATCATTGCTATCGG
Reverse primer	TTGGAAGTATAAATTTCCACGCATGGTTACAAATTCTTC
Expression vector	pCAMnLIC01
Expression host	E. coli strain BL21(DE3)
Complete amino-acid sequence of the construct produced	MREYDIIAIGGGSGGIATMNRAGEHGAQAAVIEEKKLGGTCVNVGCVPKKIMWYGAQIAETFHQFGEDYGFKTTDLNFDFATLRRNRESYIDRARSSYDGSFKRNGVDLIEGHAEFVDSHTVSVNGELIRAKHIVIATGAHPSIPNIPGAELGGSSDDVFAWEELPESVAILGAGYIAVELAGVLHTFGVKTDLFVRRDRPLRGFDSYIVEGLVKEMERTNLPLHTHKVPVKLEKTTDGITIHFEDGTSHTASQVIWATGRRPNVKGLQLEKAGVTLNERGFIQVDEYQNTVVEGIYALGDVTGEKELTPVAIKAGRTLSERLFNGKTTAKMDYSTIPTVVFSHPAIGTVGLTEEQAIKEYGQDQIKVYKSSFASMYSACTRNRQESRFKLITAGSEEKVVGLHGIGYGVDEMIQGFAVAIKMGATKADFDATVAIHPTSSEEFVTMR

2.2. Crystallization  

Initial crystallization trials were carried out with commercially available screens (Molecular Dimensions and Qiagen) by the sitting-drop vapour-diffusion method using equal volumes (0.2 µl) of protein sample (10–20 mg ml⁻¹ in 20 mM Tris, 0.15 M NaCl pH 7.5) and reservoir solution dispensed using a Crystal Gryphon robot (Art Robbins Instruments). Initial crystals grew after four weeks in 0.2 M potassium thiocyanate, 0.1 M bis-Tris propane pH 7.5, 20%(w/v) PEG 3350. Optimization of this condition yielded crystals of approximate dimensions 50 × 150 × 150 µm that were grown by hanging-drop vapour diffusion at 20°C by mixing equal volumes (1 µl) of protein solution (20 mg ml⁻¹ in 20 mM Tris, 0.15 M NaCl pH 7.5) and reservoir solution (0.1 M bis-Tris propane pH 7.0, 0.2 M potassium thiocyanate, 14% PEG 3350, 0.1 mM NADP⁺) and equilibrating the drops against 0.5 ml reservoir solution. A single SpGR crystal was transferred to cryoprotectant solution (0.1 M bis-Tris propane pH 7.0, 0.2 M potassium thiocyanate, 14% PEG 3350, 0.1 mM NADP⁺, 21% glucose) and flash-cooled in liquid nitrogen. Crystallization information is summarized in Table 2 ▸.

Table 2. Crystallization.

Method	Hanging drop
Plate type	VDXm plates
Temperature (K)	293
Protein concentration (mg ml−1)	16
Buffer composition of protein solution	20 mM Tris–HCl pH 7.5, 150 mM NaCl
Composition of reservoir solution	0.2 M potassium thiocyanate, 0.1 M bis-Tris propane pH 7.0, 14%(w/v) PEG 3350, 0.1 mM NADP+
Volume and ratio of drop	2 µl, 1:1
Volume of reservoir (µl)	500

2.3. Data collection and processing  

Diffraction data were recorded on an ADSC Quantum 315r detector at 100 K (in a nitrogen stream) on beamline MX2 at the Australian Synchrotron at a wavelength of 0.954 Å. Data were processed with XDS (Kabsch, 2010 ▸) and were merged and scaled with AIMLESS (Evans & Murshudov, 2013 ▸). Data-collection and processing statistics are summarized in Table 3 ▸

Table 3. Data collection and processing.

Values in parentheses are for the outer shell.

Diffraction source	MX2 beamline, Australian Synchrotron (Aragão et al., 2018 ▸)
Wavelength (Å)	0.954
Temperature (K)	100
Detector	ADSC Quantum 315r
Crystal-to-detector distance (mm)	400
Rotation range per image (°)	0.5
Total rotation range (°)	180
Exposure time per image (s)	0.5
Space group	P21
a, b, c (Å)	96.75, 172.67, 135.43
α, β, γ (°)	90, 91.05, 90
Mosaicity (°)	0.21
Resolution range (Å)	50–2.56 (2.60–2.56)
Total No. of reflections	531768
No. of unique reflections	141364
Completeness (%)	98.9 (84.1)
Multiplicity	3.8 (3.4)
〈I/σ(I)〉	8.7 (1.5)
R p.i.m. (%)	6.2 (39.0)
CC1/2	0.994 (0.641)
Overall B factor from Wilson plot (Å2)	39.3

2.4. Structure solution and refinement  

The structure of SpGR was solved by molecular replacement with Phaser (McCoy et al., 2007 ▸). A search model was generated with CHAINSAW (Stein, 2008 ▸) from the structure of E. coli GR (EcGR; 61% sequence identity; PDB entry 1ger; Mittl & Schulz, 1994 ▸). Refinement was carried out using REFMAC5 (Murshudov et al., 2011 ▸) with iterative cycles of model building in Coot (Emsley & Cowtan, 2004 ▸). Tight noncrystallographic symmetry (NCS) restraints were applied during the early stages of refinement. These were relaxed and removed entirely as refinement progressed. Model geometry was analysed by MolProbity (Chen et al., 2010 ▸). Refinement statistics are summarized in Table 4 ▸.

Table 4. Structure solution and refinement.

Values in parentheses are for the outer shell.

Resolution range (Å)	50–2.56 (2.62–2.56)
Completeness (%)	98.7
σ Cutoff	None
No. of reflections, working set	134126 (8684)
No. of reflections, test set	7170 (434)
Final R cryst	0.180 (0.326)
Final R free	0.233 (0.338)
Cruickshank DPI	0.229
No. of non-H atoms
Protein	27516
Ligand	430
Solvent	1766
Total	29712
R.m.s. deviations
Bond lengths (Å)	0.007
Angles (°)	1.079
Average B factor (Å2)	43.34
Ramachandran plot
Most favoured (%)	96.8
Allowed (%)	99.7
PDB code	6du7

3. Results and discussion  

3.1. Overall structure of SpGR  

SpGR crystallized in space group P2₁ with eight molecules in the asymmetric unit (arranged as four homodimers; Fig. 2b). All residues (1–448) were modelled in chains A, B, C, D, F and G, while chains E and H included 447 and 446 residues, respectively. Refinement of the model to 2.56 Å resolution converged with residuals of R = 18.0% and R _free = 23.3% with excellent statistics and model geometry. The atomic coordinates and structure factors have been deposited in the Protein Data Bank (PDB) with accession code 6du7 (Table 4 ▸).

The observation of dimeric SpGR assemblies in the crystal is consistent with the analysis of the purified protein by size-exclusion chromatography (Fig. 1 ▸), which showed that SpGR eluted at an elution volume corresponding to a molecular weight of approximately 100 kDa. These data confirm that SpGR was also present as a dimer in solution (Fig. 1 ▸) and are consistent with the structural characterization of GR enzymes from E. coli (EcGR; Mittl & Schulz, 1994 ▸), H. sapiens (hGR; Schulz et al., 1978 ▸; Karplus & Schulz, 1989 ▸; Savvides & Karplus, 1996 ▸; Berkholz et al., 2008 ▸), P. falciparum (PfGR; Sarma et al., 2003 ▸; Böhme et al., 2000 ▸) and S. cerevisiae (yGR; Yu & Zhou, 2007 ▸). There are no significant differences between the backbone structures of the eight molecules in the asymmetric unit, as shown by r.m.s.d. values ranging from 0.2 to 0.4 Å for the superposition of 446 common C^α atoms between each combination of monomer pairs.

Figure 1.

Purification of SpGR. The elution profile of purified SpGR relative to the indicated protein strandards analysed by size-exclusion chromatography. The molecular weight of the SpGR protein is estimated to be 100 kDa, which corresponds to a homodimer.

The SpGR monomer consists of three distinct domains: an NADPH-binding domain (residues 139–263), an FAD-binding domain (residues 1–138 and 264–334) and a dimerization domain (residues 335–448) (Fig. 2 ▸ a). The NADPH-binding and FAD-binding domains both show Rossmann folds, and the dimerization domain comprises five α-helices and five β-strands. GR enzymes contain a redox-active disulfide bond that is reduced by NADPH via FAD (Pai & Schulz, 1983 ▸), which is observed between residues Cys41 and Cys46 in the SpGR structure, and is positioned adjacent to the isoalloxazine ring of the FAD cofactor. The FAD cofactor is bound in an extended conformation within the FAD-binding domain.

Figure 2.

Structure of the glutathione reductase from S. pneumoniae. (a) The structure of the SpGR monomer is shown in cartoon representation with the FAD-binding domain in grey, the NADPH-binding domain in red and the interface domain in gold. The FAD cofactor is represented as yellow spheres. (b) Structure of the SpGR dimer: both monomers are coloured according to their domain structures as in (a).

The SpGR coordinates were used to perform a secondary-structure search of the PDB with PDBeFold (http://www.ebi.ac.uk/msd-srv/ssm; Krissinel & Henrick, 2004 ▸) and revealed close structural similarities between the SpGR structure and those of EcGR, hGR, yGR and PfGR, in addition to those of the GR enzymes from Vibrio parahaemolyticus (VpGR; PDB entry 5u1o), Yersinia pestis (YpGR; PDB entry 5vdn) and S. mutans (SmGR; PDB entry 5v36), all of which are unpublished (Table 5 ▸). In addition, the SpGR structure is closely similar to that of the glutathione amide reductase from Chromatium gracile (Table 5 ▸; Van Petegem et al., 2007 ▸).

Table 5. Structural comparisons between SpGR and homologues.

Organism	PDB code	R.m.s.d. (Å)	N align	Sequence identity (%)
Vibrio parahaemolyticus	5u1o	0.74	444	59
Streptococcus mutans	5v36	0.95	441	60
Escherichia coli	1ger	0.94	445	61
Homo sapiens	1gra	1.23	444	51
Yersinia pestis	5vdn	0.95	444	59
Plasmodium falciparum	1onf	1.11	425	44
Saccharomyces cerevisiae	2hqm	1.06	436	47
Chromatium gracile †	2r9z	1.21	437	49

^†Glutathione amide reductase (Van Petegem et al., 2007 ▸).

3.2. Structure of the subunit interface  

The subunit interface of SpGR shows a total buried surface area of 3502 Ų per monomer, comprising ∼16% of the surface area of each protomer. Two main structural regions contribute to dimer formation: residues 46–93 from the FAD-binding domain (buried surface area of 1276 Ų) and residues 408–448 at the C-terminus of the dimerization domain (buried surface area of 1481 Ų).

Residues 70–79 from each monomer align at the dimer interface in an antiparallel β-sheet structure (Fig. 3 ▸). This closely mirrors the structure of the dimerization interface of EcGR, but is distinct from those observed for yGR and hGR. In hGR an intersubunit disulfide bridge between residues Cys90 and Cys90′ is present in this position, while the yGR structure features a short α-helix connected by two loose loops owing to an insertion of six residues in this region of the structure (Fig. 3 ▸). The determination of the SpGR structure described here divides GR enzymes into three classes based on the structures of the dimer interfaces (Fig. 3 ▸). These are disulfide-linked (hGR), looped (yGR) and β-sheet (EcGR, SpGR, PfGR and other bacterial homologues). The bacterial enzymes therefore represent the major class (as defined by the structures of their dimer interfaces), all of which share the same structural features in this region (Fig. 3 ▸). Despite these structural differences, the composition of the GR dimer interface does not appear to determine the enzyme activity. Mutagenesis of the EcGR protein to introduce an intersubunit disulfide bond, similar to that observed in the hGR enzyme, did not have an impact on its catalytic activity or thermal stability (Scrutton et al., 1988 ▸).

Figure 3.

The structures of the dimer interfaces for GR enzymes (highlighted in red and blue). (a) SpGR (this work), (b) hGR (PDB entry 3dk4), (c) yGR (PDB entry 2hqm).

The catalytic mechanism and the location of the substrate-binding sites in hGR have been extensively investigated through X-ray crystallographic analyses of crystals soaked with NADPH and glutathione (Karplus & Schulz, 1989 ▸). A comparison of the SpGR structure with that of hGR shows that the dimer interface contributes to the structure of the GSSG binding pocket, which is composed of residues from both monomers. The binding pocket for NADPH is separated from the GSSG binding site by the Cys41–Cys46 disulfide linkage. Despite the fact that the SpGR crystallization conditions contained NADP⁺, we did not observe evidence for bound NADP⁺ in the electron density. The presence of the disulfide linkage in the SpGR structure described here confirms that this structure was solved in the oxidized (resting) state. Extensive analyses of the activities of GR enzymes from other sources have demonstrated the importance of this linkage in enzyme catalysis, where it mediates hydride transfer from NADPH to GSSG (Karplus & Schulz, 1989 ▸).

3.3. Kinetic analyses of the SpGR enzyme  

Michaelis–Menten analyses of the glutathione reductase activity of SpGR for the substrates NADPH and GSSG revealed the kinetic parameters K _m(NADPH) = 23.2 ± 3.3 µM, K _m(GSSG) = 231.2 ± 24.7 µM and V _max = 319.3 ± 15.9 µmol min⁻¹ mg⁻¹ (Table 6 ▸). The values of V _max and K _m(NADPH) for SpGR are consistent with those for the other characterized enzymes in this family; however, the determined value of K _m(GSSG) for SpGR is significantly higher than those reported for the majority of kinetically characterized GRs (Table 6 ▸). The only exception is the GR enzyme from spinach, which has a comparable K _m(GSSG) to that of SpGR.

Table 6. Kinetic parameters for GR enzymes from various organisms.

Organism	V max (µmol min−1 mg−1)	K m(NADPH) (µM)	K m(GSSG) (µM)
Streptococcus pneumoniae †	319 ± 16	23.2 ± 3.3	231 ± 25
Homo sapiens ‡	147§	8.5	65
Escherichia coli ¶	360.6§	16	66
Saccharomyces cerevisiae ††	166.7§	15	74.6
Plasmodium falciparum ‡‡	204§	4.8 ± 0.4	69 ± 2
Calf liver§§	190§	21.1 ± 1	101 ± 7
Rat liver¶¶	207§	8.2 ± 0.8	26.3 ± 5.7
Phaeodactylum tricornutum §	190§	14	60
Pea	26§	3	62
Spinach†	246§	2.78 ± 0.34	196 ± 40

^†This work.

^‡Savvides & Karplus (1996 ▸), Storey et al. (1998 ▸).

^§Errors were not reported.

^¶Mittl & Schulz (1994 ▸).

^††Yu & Zhou (2007 ▸).

^‡‡Böhme et al. (2000 ▸).

^§§Carlberg & Mannervik (1981 ▸).

^¶¶Carlberg et al. (1981 ▸).

^†††Arias et al. (2010 ▸).

^‡‡‡Connell & Mullet (1986 ▸).

^§§§Halliwell & Foyer (1978 ▸).

Although we were unable to determine the structure of the SpGR enzyme in the presence of the substrates NADPH and GSSG, we predicted the binding locations of these molecules by superposition with the structure of hGR, which has been solved in complex with NADP⁺ and GSSG (PDB entry 3dk4; Berkholz et al., 2008 ▸). We observed that the structures of the respective NADPH and GSSG binding sites are almost identical in the EcGR, hGR, yGR and SpGR enzymes. The hGR complex structure defined nine residues that directly interact with the substrate NADPH (Ala195, Tyr197, Ile198, Glu201, Arg218, Arg224, Gly290, Leu337 and Val370; Karplus & Schulz, 1989 ▸), eight of which are conserved in the sequences of SpGR, EcGR and yGR (Fig. 4 ▸). The single substitution is that of Leu337 in hGR for Glu in SpGR, EcGR and yGR. Based on the kinetic data [K _m(NADPH) = 23.2 ± 3.3 µM], this substitution does not appear to affect the affinity of binding of NADPH to SpGR. In the structure of hGR in complex with NADP⁺ and GSSG, the interaction of this residue with the substrate NADPH is side-chain-independent, being mediated by a hydrogen bond between its amide nitrogen and NADPH. The close structural similarities between the GSSG binding sites in the SpGR, hGR, EcGR and yGR enzymes indicate no obvious features that would explain the differences in the kinetic parameters [specifically K _m(GSSG)] for these enzymes.

Figure 4.

Sequence alignment of SpGR against the EcGR, hGR and yGR homologues. The alignment was performed using T-Coffee and ESPript (http://tcoffee.crg.cat/apps/tcoffee/do:expresso; Robert & Gouet, 2014 ▸). The secondary-structural elements were identified from PDB entries 1ger, 2hqm and 3dk4 using ESPript and are displayed at the top of the alignment. The α-helices and β-sheets are denoted α and β, respectively. Conserved residues are indicated by white lettering on a red background. Resides that form the NADPH-binding site are indicated with black stars.

4. Conclusion  

The crystal structure of SpGR determined at 2.56 Å resolution shows high structural conservation with GR enzymes from human, E. coli (and other bacterial organisms) and S. cerevisiae. Despite the significant structural conservation, kinetic studies reveal a weaker affinity of the enzyme for its substrate, GSSG. Further investigation of the links between the structural and kinetic properties of SpGR must await its structure solution in the presence of bound substrates.

Supplementary Material

PDB reference: glutathione reductase from Streptococcus pneumoniae, 6du7

Funding Statement

This work was funded by National Health and Medical Research Council grant GNT1080784 to Megan Maher, , and David Aragão.