drFrnE Represents a Hitherto Unknown Class of Eubacterial Cytoplasmic Disulfide Oxido-Reductases

Abstract

Aims:

Living cells employ thioredoxin and glutaredoxin disulfide oxido-reductases to protect thiol groups in intracellular proteins. FrnE protein of Deinococcus radiodurans (drFrnE) is a disulfide oxido-reductase that is induced in response to Cd²⁺ exposure and is involved in cadmium and radiation tolerance. The aim of this study is to probe structure, function, and cellular localization of FrnE class of proteins.

Results:

Here, we show drFrnE as a novel cytoplasmic oxido-reductase that could be functional in eubacteria under conditions where thioredoxin/glutaredoxin systems are inhibited or absent. Crystal structure analysis of drFrnE reveals thioredoxin fold with an alpha helical insertion domain and a unique, flexible, and functionally important C-terminal tail. The C-tail harbors a novel 239-CX₄C-244 motif that interacts with the active site 22-CXXC-25 motif. Crystal structures with different active site redox states, including mixed disulfide (Cys22-Cys244), are reported here. The biochemical data show that 239-CX₄C-244 motif channels electrons to the active site cysteines. drFrnE is more stable in the oxidized form, compared with the reduced form, supporting its role as a disulfide reductase. Using bioinformatics analysis and fluorescence microscopy, we show cytoplasmic localization of drFrnE. We have found “true” orthologs of drFrnE in several eubacterial phyla and, interestingly, all these groups apparently lack a functional glutaredoxin system.

Innovation and Conclusion:

We show that drFrnE represents a new class of hitherto unknown intracellular oxido-reductases that are abundantly present in eubacteria. Unlike other well-known oxido-reductases, FrnE harbors an additional dithiol motif that acts as a conduit to channel electrons to the active site during catalytic turnover. Antioxid. Redox Signal. 28, 296–310.

Innovation

drFrnE is a disulfide oxido-reductase that is over-expressed under Cd²⁺ stress, and its null mutant is sensitive to Cd²⁺ and γ-radiation. This study establishes FrnE as a new class of eubacterial intracellular oxido-reductases that could function under conditions when the other well-known intracellular disulfide oxido-reductase systems are absent or inhibited.

Introduction

Disulfide oxido-reductases of the thioredoxin superfamily have a central role in redox regulation of the cell. In the cytoplasm, thioredoxin and glutaredoxin systems, and low molecular weight (LMW) thiols (such as GSH, bacillithiols, mycothiols, coASH) play a critical role in maintaining free cysteine residues of intracellular proteins in the reduced state (11, 24, 47, 49). Recently, the thioredoxin system has been shown to be a target for the development of anti-microbial compounds, as it is essential for DNA synthesis (23, 26). The thioredoxin system consists of thioredoxin (Trx) and thioredoxin reductase (TrxR) proteins (25), whereas the glutaredoxin system consists of glutaredoxin (Grx) and glutathione reductase (GR) proteins (10). Thioredoxin and glutaredoxin were discovered in the years 1964 (22) and 1976 (16), respectively, and since then, no other intracellular oxido-reductase system has been identified in eubacteria. In comparison to Trx and Grx systems, Dsb proteins (DsbA–DsbG) are periplasmic disulfide oxido-reductases that help in the formation of native disulfide bonds in the secretory proteins (17).

The Trx superfamily is characterized by the presence of the Trx-fold and active site CxxC motif. The basic catalytic mechanism for Trx can be described as a bimolecular nucleophilic substitution reaction (5). The N-terminal nucleophilic cysteine of the CxxC motif forms a mixed disulfide bond between Trx and substrate protein. In the next step, C-terminal cysteine of the CxxC motif attacks the mixed disulfide, releasing a reduced substrate protein and oxidized Trx. The reduced state of the Trx active site is then regenerated by TrxR using electrons from NADPH. In case of Grx, GSH provides the necessary reducing equivalents.

Deinococcus radiodurans is a highly radiation-resistant bacterium that encodes both enzymatic and nonenzymatic components, contributing to its oxidative stress tolerance (28, 41, 42, 51). Recently, the role of the well-protected functional proteome in the recovery from radiation damage has been highlighted (20). The FrnE protein of D. radiodurans (drFrnE; NCBI Ref: WP_010887304) is induced on exposure to a sub-lethal concentration of Cd²⁺, and drfrnE::nptII null mutant showed reduced tolerance to cadmium and diamide (19, 32). These mutants were nearly six times less resistant to γ-radiation and displayed about twofold higher sensitivity to hydrogen peroxide (19). The molecular basis of drFrnE protecting D. radiodurans from Cd²⁺ metal toxicity is not yet clear. Generally, the key molecular events during oxidative stress induced by Cd²⁺ and Hg²⁺ exposure are expected to be protein S-glutathionylation, alteration in the ratio of reduced and oxidized LMW thiols, and reduced cellular accumulation of manganese and zinc ions (3, 14). For instance, bacillithiol (BSH) functions as a zinc buffer and BSH null mutants of Bacillus subtilis are found to be more sensitive to the toxic effects of Cu²⁺ and Cd²⁺ (8, 27, 40).

Here, we report the structural, biochemical, biophysical, and bioinformatics characterization of drFrnE. Structural and mutational analyses show a unique C-terminal tail harboring a novel 239-CX₄C-244 motif that can channel electrons to the active site 22-CXXC-25 motif. The structural studies also provide snapshots of the reaction pathway. drFrnE is more stable in oxidized form, is located in the cytoplasm, and is functional in the presence of Cd²⁺. drFrnE orthologs are abundant in several eubacterial phyla and interestingly, these FrnE-positive species lack a functional glutaredoxin system. Our results suggest that FrnE is a new class of hitherto unknown cytoplasmic oxido-reductase present in many eubacteria.

Results

Crystal structures suggest involvement of C-terminus in drFrnE function

The X-ray crystal structure of drFrnE shows that it belongs to the Trx superfamily, with the core domain (residues 12–52 and 179–222) consisting of a four-stranded β-sheet surrounded by three α-helices (Fig. 1a–d). drFrnE also harbors a conserved helical insertion domain (residues 53–178) and a unique C-terminal extension (223–252) with a second 239-CX₄C-244 motif. The characteristic active site 22-CPWC-25 motif in the canonical Trx fold is located on the loop connecting β-sheet 1 and α-helix 1, and it is shielded by the conserved cisPro loop (Val191-Pro192). The active site cysteines, in the crystal structures of drFrnE in the absence of reducing agent and in the presence of β-mercaptoethanol (βME), are in the oxidized form, with Cys22 forming a disulfide bond with Cys25 (Fig. 1e). In agreement with the crystal structure, drFrnE shows negligible absorbance at 412 nm in DTNB assay, suggesting that both the cysteine motifs are in oxidized form in the absence of reducing agent. The active site cysteines are in the reduced form in the crystal structure determined in the presence of nonthiol-reducing agent, tris(2-carboxyethyl)phosphine (TCEP) (Fig. 1f). Interestingly, in the crystal structure determined in the presence of 1,4-dithiothreitol (DTT), Cys22 forms a mixed intermolecular disulfide bond with Cys244 that is contributed by another drFrnE molecule (Fig. 1g).

FIG. 1.

Overall structure and active site of drFrnE. (a) Cartoon diagram of the recombinant wild-type drFrnE. The Trx-like domain is shown in slate blue, the insertion helical subdomain is shown in brick-red, and the C-terminal extension is shown in wheatish color. Active site disulfide bond is also shown in stick. (b) The topology diagram. (c) A close-up view of an α-helix (residues 22–36) is shown to illustrate the quality of the electron density map. The electron density map computed with 2mfo-Dfc coefficients is contoured at the 1σ level. (d) Surface representation of drFrnE showing C-tail of the symmetry-related molecule forming a hook-like structure and binding to the active site. Color representation is same as (a). (e) The Trx-like active site of the wild-type drFrnE (and in the presence of βME, data not shown). The protein residues are shown as ball-and-sticks, in which cyan represents carbon, red represents oxygen, blue represents nitrogen, and yellow represents sulfur. The active site Cys22 and Cys25 are in the oxidized form in the native structure. (f) The fully reduced active site in the crystals grown in the presence of TCEP. (g) Mixed disulfide between Cys22 and Cys244 of neighboring molecule (shown in magenta) is observed in the crystals obtained in the presence of DTT. βME, β-mercaptoethanol; DTT, 1,4-dithiothreitol; TCEP, tris(2-carboxyethyl)phosphine. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Notably, in all the structures, the C-terminal tail forms a hook-like shape and anchors itself into the active site of another drFrnE molecule with the C-terminal Cys motif located close to the active site cysteines (Fig. 1d), thus forming an extended network in the crystal lattice. The C-terminal residues displayed high flexibility (RMSF ∼1 nm) in 2 ns molecular dynamics (MD) simulations. In the crystal structures, C-tail binds to the active site in two different modes. In one of the binding poses, Glu-237 is present in the active site and in the second pose, a mixed disulfide of Cys244 with Cys22 is formed. The corresponding C-tail peptides presented in the active site are 235-PAEGC-239 and 242-GQCAV-246 (these sites can be considered as P2′-P1′-P0-P1-P2 for further discussion). These peptides bind in the cleft created by residues Lys73 and Tyr74 (from insertion domain) on one wall and residues 189–191 (from Trx domain) as the other wall of the cleft (Fig. 2). It appears that binding specificity is provided by main-chain hydrogen bonds between residues 189–191 and β-strand conformation at P0, P1, and P2 sites (Fig. 2a, b). The P2′ binding site is also characterized by the presence of the indole ring of active site Trp24. The two peptides do not share sequence identity and C-tail displays highly variable sequences in the drFrnE orthologs, except for strictly conserved C-terminal cysteines (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/ars).

FIG. 2.

Similarity of C-tail/active site binding in drFrnE with DsbB/DsbA binding. (a) Cartoon representation of C-tail (wheatish) in the active site of drFrnE (Trx domain in slate blue and insertion domain in brick-red). Mixed disulfide bond between Cys22 and Cys244 is shown in sticks. Residues that are important in recognition are labeled. (b) Interactions of C-tail (241-DGQCAVP-247) in the active site of drFrnE-DTT structure. (c) Superposition of DsbA/DsbB (green) peptide structure (PDB id 4tky) onto drFrnE-DTT structure. For clarity, only DsbA residues are labeled. (d) Similarity of hydrogen bonds in DsbA-DsbB (green) peptide binding with drFrnE-C-tail binding. Residue labels are for DsbA residues. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

The structures also reveal interaction of a conserved Ser18 with the resolving Cys25 (Fig. 3a), where the Oγ atom of Ser18 is H-bond donor to Cys25. The Oγ atom of Ser18 also accepts H-bond from ɛ-NH2 of Lys129, which should constrain hydrogen of Ser18 toward Cys25 (angle NH2-Oγ-Sγ ∼134°; Fig. 3a). In the reduced state, as observed in drFrnE-DTT and drFrnE-TCEP structures, Cys25 shifts toward Ser18 to form a stronger hydrogen bond that could stabilize the thiolate form of Cys25 (Fig. 3b, c). Residue Asp19 is also strictly conserved in drFrnE orthologs. Asp19 belongs to the Trx domain and forms strong H-bonds with Leu54, His113, and Thr109 residues of the insertion domain (Fig. 3d). Interestingly, these residues are also conserved in drFrnE orthologs (Supplementary Fig. S1). These interactions might be important for the structural stability of drFrnE active site and relative arrangement of Trx and insertion domain.

FIG. 3.

Interactions of conserved residues in drFrnE active site. (a–c) Show interactions of the Oγ atom of Ser18 with Cys25 and Lys129 in different structures of drFrnE. The interactions are shown as dashed lines, with indicated distances (estimated coordinate error from phenix.refine is 0.16 Å). (a) Wild-type drFrnE. (b) drFrnE-TCEP. (c) drFrnE-DTT. (d) Strong hydrogen bonding interaction of conserved Asp19 from Trx domain with Leu54, Thr109, and His113 from insertion domain. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

drFrnE exists as a dimer

The size exclusion chromatography (SEC) analysis suggests that drFrnE exists as an oligomer in solution with an estimated Mr of ∼100 kDa (Fig. 4a). Glutaraldehyde cross-linking experiments show an additional band of ∼60 kDa corresponding to the dimer of drFrnE (Supplementary Fig. S2a). Molecular weight of drFrnE in solution estimated using average static light scattering data is ∼73 kDa (Supplementary Fig. S2b). These observations cumulatively suggest that drFrnE has a dimeric quaternary structure in solution. The Native-PAGE shows drFrnE band between bovine serum albumin (66 kDa) and chicken egg albumin (45 kDa) (Supplementary Fig. S2c). For C239S/C244S and ΔC222 truncation (residues 223–252 deleted) mutants, SEC analyses indicate similar quaternary structures (Fig. 4a). The size of drFrnE particle (7.8 nm; Fig. 4b) was not affected by the presence of 10 mM DTT, suggesting that the oligomeric status is not due to intermolecular disulfide bond, as also observed from SDS-PAGE of reduced and nonreduced drFrnE (Fig. 4c). The dimer is stable in the presence of 500 mM salt or 0.2% triton, as evident from DLS profiles (data not shown). However, the presence of 0.6% CHAPS (zwitter-ionic detergent) or 0.6% sarcosyl (anionic detergent) breaks this dimeric assembly, suggesting that both hydrophobic and ionic interactions may be involved in the dimer formation (Supplementary Fig. S3).

FIG. 4.

drFrnE forms oligomers in solution, and oxidized drFrnE displays higher thermal and chemical stability. (a) An overlay of size exclusion chromatography profiles of recombinant wild-type drFrnE (filled squares) and C239S/C244S (filled circles) and ΔC222 truncation mutants (filled triangles). The retention volumes of the molecular weight markers (Ferritin, 440 kDa; Conalbumin, 75 kDa; Carbonic anhydrase, 29 kDa) are shown at the top of the figure. (b) A profile overlay of DLS of oxidized drFrnE at pH 7 (filled squares) and of reduced drFrnE in the presence of 10 mM DTT (open triangles). (c) Twelve percent SDS-PAGE analysis of drFrnE in the presence or absence of reducing agent (DTT). Lane 1, markers; lane 3, drFrnE+DTT+heat; lane 4, drFrnE+DTT+No heat; lane 6, drFrnE+No DTT+heat; lane 7, drFrnE+No DTT+No heat. (d) The first derivatives of the thermal denaturation curves of oxidized (filled circles) and reduced (open triangles, in the presence of 2 mM DTT) drFrnE monitored by ellipticity at 220 nm. (e) A profile overlay of DSC heating scans of oxidized (solid line) and reduced (dash line) drFrnE in the presence of 2 mM DTT. (f) Equilibrium GdmHCl unfolding profile of oxidized drFrnE (filled squares) and reduced drFrnE (open triangles, in the presence of 5 mM DTT) monitored by ellipticity at 220 nm. DSC, differential scanning calorimetry.

drFrnE is stable in oxidized form

The drFrnE protein does not show significant changes in the secondary structure on exposure to reducing agent (Supplementary Fig. S4a). However, the oxidized drFrnE (in the absence of DTT) shows a typical transition profile from native to unfolded state with T _M of ∼59°C, whereas it shifts to a lower value of 52°C in the presence of DTT (Fig. 4d and Supplementary Fig. S4b). Differential scanning calorimetry (DSC) analyses shows T _M of ∼57°C for drFrnE, which shifts to 52°C in the presence of DTT (Fig. 4e). Taken together, both the studies clearly suggest higher thermal stability of drFrnE in oxidized form, compared with the reduced form.

The thermodynamics of unfolding of drFrnE driven by guanidine hydrochloride (GdmHCl) reveals characteristic of the two-state model (Fig. 4f). GdmHCl unfolding of drFrnE commences at about 0.5 M, followed by a sharp transition to the unfolded end of the profile at about 2 M of GdmCl. The unfolding profiles of oxidized and reduced drFrnE were found to be similar, although with a small change in free energy values; an average of 8.6 ± 0.38 kcal/mol for oxidized drFrnE, as compared with 7.9 ± 0.23 kcal/mol for reduced drFrnE. It further supports higher stability of oxidized drFrnE.

drFrnE displays tolerance toward Cadmium (II) ions

The insulin reductase activity was estimated from the accumulated aggregates of reduced insulin, which, in turn, also depends on the nucleation-dependent process contributing toward initial lag time. Reducing agent DTT (330 μM) was used as electron donor in insulin reduction assays. drFrnE shows insulin reductase activity with a lag time of ∼25 min, which is essentially independent of enzyme concentration (Supplementary Table S1). The C-terminal Cys mutant (C239S/C244S double mutant) shows a similar lag time but a slightly higher rate of insulin precipitation (increased slope), compared with that of wild-type drFrnE. The ΔC222 truncation mutant, on the other hand, shows an increased lag phase and a significantly higher rate as compared with wild-type drFrnE (Fig. 5a). Similar activities of C239S/C244S double mutant and wild-type drFrnE suggest that C-terminal cysteines are not involved directly in catalysis. The presence of DTT in the reaction mixture, which can directly reduce active site cysteines, obviates the need of C-terminal cysteines.

FIG. 5.

Reduction activity, phylogenetic tree, and cytoplasmic localization of drFrnE. (a) drFrnE (10 μM) can reduce insulin disulfides in the presence of 50 μM (red) and 100 μM (green) Cd²⁺. The C-tail truncation mutant (black) displays higher activity than the wild-type recombinant protein (blue). The rate of insulin precipitation was monitored as the increase in turbidity at 650 nm for 60 min. Activity with DTT alone (control) is shown in gray color. Data presented are average of three replicates. (b) Wild-type drFrnE (blue) reduces DTNB more efficiently than C239S/C244S (red) and ΔC222 truncation mutants (green) in the presence of ecTrxR and NADPH, suggesting that C-tail residues are involved in electron channeling. drFrnE alone shows negligible activity in the absence of electron donor (violet). (c) Maximum-likelihood phylogenetic unrooted tree computed by using Mega7 program. The tree with the highest log likelihood (−3811.72) is shown and is drawn to scale, with branch lengths measured in the number of substitutions per site. The “true” FrnE sequences were identified from the NCBI database by using BLAST search against each of the 27 eubacterial phyla and drFrnE as query sequence. Only one FrnE sequence from each phylum is used in the phylogenetic analysis. Sequences are identified based on a three-letter code for species and phylum. (d) Fluorescence microscopy of drFrnE-GFP. The left panel shows Deinococcal cell harboring pVHS559 plasmid (control), whereas the right panel shows drFrnE-GFP expression in Deinococcus radiodurans R1 harboring pFrnEGFP plasmid. The boxed areas are magnified in the lower panels. Bar 50 pixel. DTNB, 5,5-dithio-bis-(2-nitrobenzoic acid). To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

The effect of Cd²⁺ was probed by using 50 or 100 μM of CdCl₂. The drFrnE displays maximal activity in the presence of Cd²⁺ in both the assays (Fig. 5a), though slope decreases and lag phase increases at a higher Cd²⁺ concentration (Supplementary Table S1). In comparison, recombinant Escherichia coli thioredoxin (Trx-1) was irreversibly inhibited in the presence of Cd²⁺ and accumulated activity did not reach the maximum (Supplementary Fig. S5).

drFrnE accepts electrons through its C-tail

The reductase activity of drFrnE and its mutants was also measured by DTNB [5,5-dithio-bis-(2-nitrobenzoic acid)] assay using E. coli thioredoxin reductase (ecTrxR) and NADPH as a coupled electron source. Significant reduction of DTNB by drFrnE suggests that drFrnE accepts electrons from NADPH/ecTrxR couple (Fig. 5b). The C239S/C244S and ΔC222 truncation mutants, however, display a marked decrease in this reduction capability by using NADPH/ecTrxR couple as electron donor (Fig. 5b).

“True” orthologs of drFrnE are abundant in many eubacterial phyla

Both the crystal structures and activity analyses suggested the functional role of C-terminal cysteine motif. Therefore, we used conserved C-terminal Cys motif as characteristic for drFrnE type oxido-reductases and used that as a probe to identify “true” orthologs of drFrnE. We find that drFrnE orthologs are abundantly present among seven eubacterial phyla. To our surprise, we did not detect the Grx system and the proteins involved in the glutathione biosynthesis (GshA, GshB, and GshF) in majority of the species carrying frnE gene (Supplementary Table S2).

In the evolutionary tree, all the “representative” FrnE sequences cluster together (Fig. 5c). The structurally related prokaryotic HCCA isomerase, mitochondrial GST class-kappa (GSTk), and bacterial Dsb proteins form independent clades. Structural similarity (based on DALI Z-score) could not discriminate the evolutionary relationship between drFrnE, GSTk, and Dsb proteins. The phylogenetic analysis, however, suggests that FrnE proteins are more closely related to mitochondrial GSTk (or bacterial HCCA isomerase) proteins than the periplasmic DsbA proteins, despite the absence of active site Cys residues in GSTk proteins. The presence of C-tail with a second Cys motif is unique to drFrnE; however, an extra helical domain insertion in the trx-fold is present in all the three proteins at the same site. Importantly, length and topology of the insertion domains in FrnE and GST class-Kappa match with each other and are different from the insertion domain of DsbA (Supplementary Fig. S6). GSTk proteins are considered to be evolved independently by convergent evolution and not to be part of the GST gene family (33). Based on their structural similarity, we hypothesize that drFrnE and GSTk proteins may share a common ancestor.

drFrnE localizes in cytoplasm

Glutathione redox couple was used to determine drFrnE-glutathione redox equilibrium by monitoring intrinsic fluorescence of Trp residues. The equilibrium constant for the drFrnE/glutathione system was found to be 2.78 × 10⁻⁴ M on nonlinear regression (Supplementary Fig. S7). The apparent redox potential value for drFrnE at 25°C and pH 7 (E _o ^FrnE) was estimated by using the Nernst Equation to be −187 mV, which is close to cytoplasmic thiol oxidoreductase proteins.

SignalP4.1, which uses the neural network-based approach to provide reliable identification of signal peptides, did not detect signal sequence in the representative “true” FrnE sequences, except for FrnE of Deinococcus proteolyticus (Supplementary Table S3). The Cello2go server predicted with high confidence cytoplasmic localization for all the FrnE proteins and poor scores for extra-cellular and periplasmic localization. For instance, percentage probability of cytoplasmic localization for E. coli FrnE was 82.2%, as compared with extra-cellular and periplasmic probabilities of 1.8% and 5.1%, respectively (Supplementary Table S3). Thus, both the servers predicted cytoplasmic localization for FrnE orthologs with high confidence.

D. radiodurans cells harboring pFrnEGFP were induced with IPTG, and the synthesis of the drFrnE-GFP fusion protein was confirmed by immunoblotting using GFP-specific antibodies (data not shown). A microscopic examination of D. radiodurans cells expressing drFrnE-GFP fusion shows GFP fluorescence distributed throughout the cell, akin to fluorescence by GFP alone, suggesting localization of drFrnE protein in the cytoplasm of D. radiodurans (Fig. 5c and Supplementary Fig. S8).

Discussion

Thioredoxin and glutaredoxin are two well-known cytoplasmic thiol oxido-reductase systems that help to maintain intracellular redox status in eubacteria. The drFrnE belongs to the thioredoxin superfamily and displays both disulfide isomerase and insulin reductase activities (19). In Deinococcus, frnE gene is over-expressed by sevenfold on exposure to Cd²⁺ (19). Expectedly, we have observed purified drFrnE protein to be tolerant toward a considerably high concentration of Cd²⁺. Native-like insulin reductase activity of drFrnE approaching total accumulation, although with the increase in lag time, was observed even in the presence of 100 μM Cd²⁺ (Fig. 5a).

X-ray crystal structure of drFrnE shows a canonical 22-CPWC-25 active site motif and an additional 239-CxxxxC-244 Cys motif in a unique extended C-terminal tail (Fig. 1a, b). Crystal structure of drFrnE and DTNB assay show that both the cysteine motifs are in oxidized form in the absence of reducing agent.

The stability of oxidized vs reduced states of thiol oxido-reductases provides driving force and determines the thermodynamically favorable direction of the chemical reaction (5). Trx is stable in oxidized state and acts as a reducing agent in the bacterial cytoplasm. In comparison, DsbA is more stable in the reduced form and functions as an oxidant in the bacterial periplasm. The oxidized state of drFrnE (in the absence of reducing agents) displays higher thermal and chemical stability as observed by circular dichroism (CD) and DSC analyses (Fig. 4d–f). The apparent redox potential value for drFrnE was estimated by using the glutathione couple to be −187 mV (Supplementary Fig. S7), which is similar to that of cytoplasmic Grx 3 (−198 mV) and is significantly different from periplasmic Dsb proteins (DsbA, −122 mV; DsbC, −130 mV). These data pointed out that drFrnE is likely to be a cytoplasmic disulfide reductase. Bioinformatics analysis also suggested with high confidence that “true” FrnE orthologs may be localized in the cytoplasm (Supplementary Table S3). The localization of drFrnE in the cytoplasm was confirmed by fluorescence microscopy studies using expression of drFrnE-GFP fusion protein in Deinococcus (Fig. 5d).

We have determined drFrnE crystal structures in three different redox states. In the structures of native protein and in the presence of βME, the active site is oxidized with an intra-molecular disulfide bond between Cys22 and Cys25 residues (Fig. 1e). The active site is in reduced state in the presence of TCEP (Fig. 1f), whereas an intermolecular disulfide bond between Cys22 and Cys244 residues is observed in the presence of DTT (Fig. 1g). These structures are likely to represent snapshots of reaction pathway.

By comparison with Trx family members, Cys22 can be expected to be activity linked with lower pKa and Cys25 can be expected to be resolving residue. We have observed that mutations of these Cys residues abrogate insulin reductase activity of drFrnE. The first step of the reaction mechanism in disulfide reductases, such as Trx, is to attack the disulfide of the incoming substrate by the thiolate form of catalytic cysteine. Therefore, to function as a disulfide reductase in vivo, the active site of drFrnE needs to be in the reduced state and this pose is observed in the structure in the presence of TCEP (PDB id 5E59, Fig. 1f and Fig. 6: stage 1). In this state, reduced Cys22 is in an optimal position to react with the incoming substrate disulfide bond. Resolving Cys25 in this state forms a hydrogen bond with Ser18, and this interaction can modulate pK_a of Cys25 required for its function as a resolving partner (Fig. 3b). A similar interaction of resolving cysteine with a conserved threonine in thioredoxin from Bacillus cereus was found to increase the enzymatic activity by stabilizing the resolving thiolate anion (43). Release of the substrate protein in reduced form in the next stage will leave drFrnE in oxidized form, which is represented by the structure of drFrnE in the absence of reducing agent (PDB id 5CNW) and in the presence of βME (PDB id 5COH) (Fig. 1e and 6: stage 2). The presence of C22–C25 and C239–C244 disulfides in drFrnE was confirmed by DNTB assay. The Ser18-Cys25 hydrogen bond distance increases from 3.0 to 3.5 Å in this structure. These stages are similar to the basic catalytic mechanism of the Trx superfamily.

FIG. 6.

Schematic representation of different stages of drFrnE catalytic cycle. Stage 1 shows reduction of disulfide in substrate protein by drFrnE active site cysteines. Stage 2 comprises reduction of C-terminal cysteines of drFrnE by cytosolic electron donor. Stage 3 includes regeneration of reduced 22-CxxC-25 motif by C-terminal cysteine motif through intermolecular disulfide bond between Cys22 and Cys244. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

In Trx, regeneration of the reduced form of active site cysteines takes place by direct interaction with thioredoxin reductase (TrxR) through the formation of the inter-molecular Trx-TrxR disulfide bond. Analogous intermolecular disulfide bond is observed in the crystal structure of DsbA-DsbB complex. The intermolecular disulfide bond between Cys22 and Cys244 residues seen in drFrnE structure in the presence of DTT (PDB id 5CO3, Fig. 1g and 6: stage 3) suggests that the regeneration of drFrnE, however, most likely is facilitated by its C-tail, through the conserved 239-Cx4C-244 motif. The remarkable similarity in the binding modes of drFrnE-C-tail binding with DsbA-DsbB peptide complex (Fig. 2c, d) also supports the role of C-terminal cysteines in the regeneration of the active state (reduced form) of drFrnE. Similar redox behavior of CxxxxC motif and its interaction with CxxC motif have been described in several other enzymes such as eukaryotic TrxR, GR, and SBPase (25).

The role of C-terminal cysteines as a conduit to transfer electrons to the active site cysteines is evident from the loss of DTNB reduction activity in drFrnE C239S/C244S and ΔC222 truncation mutants when ecTrxR/NADPH were used as electron donor. The activity decreases to <15% in ΔC222 truncation mutant, compared with wild-type protein. The critical role of C-tail residue in electron channeling is also consistent with the observation that drFrnE alone complements the DsbCD system when exported to the periplasm (19). The C-terminal residues of drFrnE monomer display high flexibility (RMSF ∼1 nm) in 2 ns MD simulations and are expected to be easily accessible to a primary electron donor.

E. coli thioredoxin reductase is strongly inhibited in the presence of Cd²⁺. Proteins involved in glutathione metabolism are not found in Deinococcus (or other frnE-positive species), and, therefore, GSH is an unlikely source for reductive equivalents to drFrnE. BSH has been detected in D. radiodurans (34). Also, we could confirm the presence of BSH reductase homolog in the Deinococcus genome. It is, thus, possible that bacillithiol might act as a reducing agent for FrnE in Deinococcus.

The presence of C-terminus tail and a second CxxxxC motif is unique to drFrnE and is not found in any of the structurally or functionally related proteins. The “true” drFrnE orthologs, detected from genome databases based on the conservation of C-terminal residues, revealed the presence of FrnE protein in at least 7 out of the 27 NCBI eubacterial phyla. Interestingly, the glutaredoxin system (GR, Grx1, Grx2, Grx3) could not be detected in most of the bacterial species carrying the “true” frnE gene (Supplementary Table S2). It may, thus, be conjectured that FrnE may be the only intracellular functional oxido-reductase system in the seven identified phyla in the presence of Cd²⁺. Recently, the thioredoxin system has been validated as a target for development of anti-microbial compounds, especially in bacteria lacking the Grx system (23, 26). We have observed the presence of FrnE in many pathogenic bacteria, including Mycobacterium tuberculosis. Thus, due considerations for the presence of FrnE protein in target pathogens may be needed in an effective drug discovery strategy.

In conclusion, FrnE is a cytoplasmic disulfide oxido-reductase found in eubacteria that lack proteins belonging to the glutaredoxin system. Unlike thioredoxin and thioredoxin reductase, drFrnE is functional in the presence of Cd²⁺. Different redox states of active sites obtained under different reducing environments provide structural snapshots for different stages of disulfide reductase activity of drFrnE. It has also been shown here that essential regeneration of the active site is aided by a flexible C-tail through its novel Cys-motif. Using a combination of diverse structural and molecular biology approaches, we demonstrate that FrnE represents a novel cytoplasmic thiol-disulfide oxido-reductase system that may protect eubacterial intracellular proteome under stress.

Materials and Methods

Bacterial strains, plasmids, and materials

D. radiodurans R1 (ATCC13939) was a gift from Professor J. Ortner, Germany (46). Cloning of wild-type drFrnE and construction of C22S/C25S mutants were described earlier (19). The C239S/C244S double mutant and truncation mutant ΔC222, lacking residues 223–252, were constructed by overlap extension PCR protocol. The E. coli strain NOVABLUE was used for cloning, and the E. coli strain BL21 (DE3) was used for the expression of recombinant proteins. E. coli was grown in Luria-Bertani broth and D. radiodurans was grown in TGY medium (1% trypton, 0.1% glucose, 0.5% yeast extract) with shaking at 180 rpm at 37°C and 32°C, respectively. Recombinant E. coli harboring expression vectors and their derivatives were grown in the presence of antibiotics as required. Shuttle expression vector pVHS559 (4) and its derivatives were maintained in E. coli strain NOVABLUE in the presence of spectinomycin (40 μg/ml) or spectinomycin (75 μg/ml) in the case of D. radiodurans. All recombinant techniques were performed as previously described (44). For microscopic examination, D. radiodurans cells harboring recombinant plasmid were grown at 32°C in TGY, and they were induced with 5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). GFP antibodies were from Roche Biochemical Germany. Molecular biology-grade chemicals, antibodies, and enzymes were procured from Sigma Chemicals Company, Roche Biochemicals (Mannheim, Germany), and New England Biolabs.

Construction of pFrnEGFP expression plasmids

Genomic DNA from D. radiodurans was prepared as described earlier (2). The 750 bp frnE gene (drfrnE) was PCR amplified by using forward primer (5′CAGGAATTCACAAACCTTGCACCCG3′) and reverse primer (5′CAGGATTCCAGTTGTTAGGGCGCTG3′). The PCR product was cloned at EcoRI and BamHI sites in pDSW208 (52) to yield pDSFrnE. The 1.5 kb drFrnE-GFP was further PCR amplified by using forward primer (5′GACACGCTTAAGATGACAAACCTT GCACCCG3′) and reverse primer (5′TAACTCGAGTCAGTTGTTAGGGCGCTG3′) and cloned at AflIII and XhoI sites in pVHS559 (4), and pFrnEGFP was obtained. The pFrnEGFP was transformed into both E. coli and D. radiodurans, and the expression of drFrnE-GFP fusion protein was confirmed by immunoblotting using antibodies against GFP.

Expression and purification of drFrnE proteins

Recombinant drFrnE was purified from E. coli BL21 (DE3) to homogeneity as previously described (19). In brief, transgenic E. coli cells expressing drFrnE on recombinant plasmid were grown in Luria-Bertani medium supplemented with 100 μM ampicillin at 37°C. At an OD₆₀₀ of ∼0.6, protein expression was induced with 1 mM IPTG and the cell culture was further incubated for 4 h at 37°C. However, the protein yield was higher if incubation was maintained overnight at 20°C. The cells were then harvested by centrifugation, re-suspended in cold lysis buffer (50 mM Tris, 300 mM NaCl, pH 8.0), and lysed by sonication on ice. Cell debris was removed by centrifuging the cell lysate at 11,000 g for 10 min. Clear supernatant containing soluble proteins was mixed with Ni-NTA (nitrilotriacetic acid) agarose (Qiagen, Germany) and was kept at 4°C for 2 h with gentle shaking to allow binding of His-tagged drFrnE protein. The proteins bound to the Ni-NTA resin were washed extensively with lysis buffer containing an increasing concentration of imidazole and finally eluted with elution buffer (50 mM Tris, 300 mM NaCl, 250 mM imidazole, pH 8.0). The protein fractions thus eluted were analyzed on 12% SDS-PAGE. Pure drFrnE fractions were pooled and dialyzed against 50 mM HEPES, pH 7.0. Pure dialyzed protein was concentrated to 20 mg/ml by using 10 kDa cut-off centrifugal filters.

The drFrnE mutant proteins (C239S/C244S double mutant and ΔC222 truncation mutant) were purified to homogeneity by using protocol used for purification of the wild-type protein. The Western blot of drFrnE proteins was performed by using monoclonal Anti-polyHistidine antibody (produced in mouse; Sigma), alkaline phosphatase-conjugated anti-mouse IgG (Sigma), and 5-bromo-4-chloro-3-indolyl phosphate/nitro-blue tetrazolium (BCIP/NBT) (Fig. 7).

FIG. 7.

Immunolabeling of drFrnE proteins expressed with poly-His tag. The proteins resolved on SDS-PAGE (12%) were challenged with monoclonal anti-polyHistidine antibody (produced in mouse; Sigma) by using antibody concentrations recommended by the manufacturer. Alkaline phosphatase-conjugated anti-mouse IgG (Sigma) was employed as a secondary antibody, and color development was performed by using 5-bromo-4-chloro-3-indolyl phosphate/nitro-blue tetrazolium (BCIP/NBT). Lane 1: Marker protein, PolyHistidine tagged human translin protein (monomer, 27.5 kDa; dimer 55 kDa), Lane 3: purified drFrnE (28.5 kDa), and Lanes 5 and 7: purified C239S/C244S (28.5 kDa) and ΔC222 truncation mutants of drFrnE (25.5 kDa).

Biophysical studies

CD spectra of the protein samples in the absence or presence of DTT (upto 5 mM) were recorded on a Jasco J-815 CD spectrometer equipped with a peltier-type thermostatic cell holder. The CD spectra of drFrnE (9 μM) in buffer A (20 mM HEPES, 20 mM NaCl, pH 7.0) in the far-UV range (200–250 nm) were recorded by using a 1 mm path length cuvette at 20°C. The thermal denaturation curves were obtained by recording changes in the dichroic intensity at 220 nm in the temperature range of 20–95°C at a heating rate of 2°C/min. Fluorescence experiments in the wavelength range 250–400 nm were performed by using Jasco fluorescence spectrometer (model: FP-8500, JASCO Inc., Japan) at 20°C. The increase in fluorescence intensity in the presence of DTT (5 mM) at 335 nm was recorded (with excitation wavelength 280 nm) by using a 1 cm path length cuvette. The concentration and time required for the complete reduction of drFrnE by DTT was optimized by using this assay.

The particle size (i.e., hydrodynamic diameter, Dh) of the protein was measured by using the Zetasizer Nano ZS (Malvern Instruments Ltd, United Kingdom) DLS instrument at 20°C at pH values of 7.0 (in HEPES buffer), and in the presence of 10 mM DTT. The oligomeric state of the protein was estimated by SEC using Superdex 200 increase 10/300 GL column (GE Healthcare). The column was pre-equilibrated with buffer A, and elution profiles of molecular weight standards (Ferritin, 440 kDa; Conalbumin, 75 kDa and Carbonic anhydrase, 29 kDa) were used to estimate the oligomeric status of drFrnE under the conditions of experiments.

The order–disorder (native–unfolded state) transition in drFrnE was measured at a heating rate of 2°C/min, in the temperature range of 20–80°C by DSC (Model: DSC 822; Mettler Toledo, Switzerland) with or without DTT (2 mM). Temperature and enthalpy calibrations of the instrument were done by using cyclohexane and indium. All samples (25 μl of 300 μM drFrnE) were hermetically sealed in aluminum crucibles.

For estimating guanidine hydrochloride (GdmCl)-induced unfolding equilibrium, drFrnE (100 μM) samples in buffer A containing increasing concentrations of GdmCl were incubated at room temperature (25°C) for 24 h. The unfolding experiments were also carried out with drFrnE in the presence of 5 mM DDT (reduced drFrnE). The GdmCl-induced unfolding curves of the native and reduced drFrnE were recorded by measuring the change in the far-UV CD signal from 200 to 250 nm at 20°C. The equilibrium unfolding curves as a function of GdmCl were obtained by plotting the CD signal at 220 nm. To generate unfolding profiles, the raw data were normalized to indicate the fraction unfolded, taking the assumption that the unfolding transition of drFrnE is complete at the highest denaturant concentration used in these experiments. Data were evaluated according to the two-state model of folding by using a six-parameter fit (29, 45).

Di-thiobis-(2-nitrobenzoic acid) (DTNB) assay

It was earlier shown that D. radiodurans Trx1 can cross-talk with E. coli thioredoxin reductase (ecTrxR) (35). DTNB assay was used to find interaction of drFrnE proteins (wild-type drFrnE, and C239S/C244S and ΔC222 truncation mutants) with ecTrxR. The assay mixture contained 50 mM phosphate buffer (pH 7.0), 200 μM NADPH, 0.1 μM ecTrxR, and 5 μM drFrnE. The reaction was started by adding DTNB (50 μM) and monitored by measuring the increase in absorbance at 412 nm by using the Shimadzu UV-1800 double-beam spectrophotometer. Fifty millimolar phosphate buffer (pH 7.0) containing 200 μM NADPH and 50 μM DTNB was used as reference in these assays. Reduction of DTNB in the presence of ecTrxR and NADPH, but without drFrnE, was used as a control. EcTrxR/NADPH by itself did not significantly reduce DTNB.

DTNB assay was also performed for quantitative estimation of the free cysteines in drFrnE and mutants. Fifty micromolar DTNB was added to the assay mixture containing 10 μM purified drFrnE proteins in 100 mM Tris (pH 8.0) and 6 M guanidine hydrochloride. Increase in the absorbance at 412 nm was monitored. An absorption coefficient of 13,700 M ⁻¹/cm was used for quantitative estimation of free cysteine groups in the protein.

Insulin reductase assay

Insulin reductase assays in the presence of DTT as electron donor were performed with wild-type and mutant drFrnE by using Synergy H1 multimode reader (BioTek). One hundred thirty micromolar insulin was mixed with 10 μM purified drFrnE in a reaction mixture containing 50 mM HEPES (pH 7.0). The reactions were initiated by adding 330 μM DTT. The rate of insulin precipitation was monitored as the increase in turbidity at 650 nm for 60 min. No protein was added to the control reaction. To understand the effect of Cd²⁺ on the activity of drFrnE, CdCl₂ (50 and 100 μM) was added to the reaction mixture containing drFrnE and DTT before the addition of insulin. The activity of drFrnE in the presence of Cd²⁺ was monitored at 650 nm and was compared with wild-type drFrnE. All the reactions were performed in triplicate.

Redox potential estimation

The drFrnE sequence carries four Trp residues. The intrinsic fluorescence of the Trp was observed to increase significantly in the presence of reducing agent (5 mM DTT). This change in fluorescence intensity was used to measure the redox potential of drFrnE by measuring the equilibrium constant of oxidation of reduced drFrnE by oxidized glutathione. Oxidized drFrnE (8 μM) was equilibrated in the presence of different ratios of GSH (0–3 mM) and GSSG (0.1 mM) in sodium phosphate buffer (pH 7.0) at 25°C. After 12 h, the change in fluorescence intensity of the sample was recorded at 337 nm (excitation wavelength 280 nm), where the difference between the emission of oxidized and reduced drFrnE was maximum. The equilibrium concentrations of GSH, GSSG, and oxidized and reduced drFrnE were calculated by using the Equations (1 )–(3), where [GSH_o] and [GSSH_o] are the initial concentrations of GSH and GSSH, R is the relative amount of reduced drFrnE at equilibrium, [FrnE_o] is the initial concentration of oxidized drFrnE, F is the measured fluorescence intensity, and F _ox and F _red are the fluorescence intensities of the completely oxidized and reduced drFrnE, respectively. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \left[ {{ \rm{GSH}}} \right] = \left[ {{ \rm{GS}}{{ \rm{H}}_{ \rm{o}}}} \right] \, - \, 2R \left[ {{ \rm{Frn}}{{ \rm{E}}_{ \rm{o}}}} \right] \tag{1} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \left[ {{ \rm{GSSH}}} \right] = \left[ {{ \rm{GSS}}{{ \rm{H}}_{ \rm{o}}}} \right] \, + \, R \left[ {{ \rm{Frn}}{{ \rm{E}}_{ \rm{o}}}} \right] \tag{2} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} R = ( F - {F_{{ \rm{ox}}}} ) / ( {F_{{ \rm{red}}}} - {F_{{ \rm{ox}}}} ) \tag{3} \end{align*} \end{document}

The data were fitted according to Equation (4) (12, 13), and the equilibrium constant for the drFrnE/glutathione system was found to be 2.78 × 10⁻⁴ M on nonlinear regression. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} R = \left( {{{ \left[ {{ \rm{GSH}}} \right] }^2} / \left[ {{ \rm{GSSH}}} \right] } \right) / \left( {{K_{{ \rm{eq}}}} + {{ \left[ {{ \rm{GSH}}} \right] }^2} / \left[ {{ \rm{GSSH}}} \right] } \right) \tag{4} \end{align*} \end{document}

The standard redox potential of drFrnE at 25°C and pH 7.0 (E _o ^FrnE) was calculated by using the Nernst Equation (5), where standard redox potential of glutathione redox couple at 25°C and pH 7 (E _o ^GSH/GSSG) is −240 mv, R = 8.315 J K⁻¹ mol⁻¹, F = 96,485 C mol⁻¹, T = 298 K, and n = 4. \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}{E_{ \rm{o}}}^{{ \rm{FrnE}}} = {E_{ \rm{o}}}^{{ \rm{GSH / GSSG}}} - \left( {RT / nF} \right) \times { \rm{ln}}{K_{{ \rm{eq}}}} \tag{5} \end{align*} \end{document}

Microscopic studies

Fluorescence microscopy studies were carried out essentially as described earlier (4). In brief, the mid-logarithmic-phase cells of D. radiodurans harboring plasmid pFrnEGFP were induced with 5 mM IPTG overnight. The expression of drFrnE-GFP was confirmed by using polyclonal antibodies against GFP. Cells expressing drFrnE-GFP were imaged by using a fluorescence microscope (Model IX83; Olympus, Japan). Image processing was done by using Image J software.

Glutaraldehyde cross-linking experiment

To determine the quaternary structure of drFrnE, glutaraldehyde cross-linking experiment was performed according to protocol described earlier (7). Briefly, 3 μl drops containing drFrnE protein (0.5 mg/ml) were placed on siliconized coverslips, inverted onto wells containing 50 μl of 25% (v/v) glutaraldehyde and the wells were sealed with vacuum grease. The plate was maintained at 30°C at Eppendorf thermomixer C. Protein drops were removed at different time intervals (10, 20, and 30 min) and were analyzed on 12% SDS-PAGE to monitor covalently linked oligomers.

Crystallographic analysis

The recombinant wild-type drFrnE was crystallized by the hanging-drop vapor-diffusion method by using 10% PEG-8000 as precipitant at pH 5.5 (36). The drFrnE was also crystallized in the presence of different reducing agents (10 mM DTT, 2 mM TCEP, or 10 mM βME) under similar conditions. The three-dimensional diffraction intensity data were acquired on the Agilent supernova X-ray diffractometer for drFrnE crystals grown in the presence of βME and TCEP, and using BL-21 beamline, INDUS-2 synchrotron (21), for all the other crystals. Crystallographic data collected on INDUS-2 synchrotron were processed by using XDS software suite (18). The data collected on the Agilent X-ray diffractometer were processed by using CrysAlisPro software suite (Agilent Technologies, Oxford, England). The highest resolution of 1.65Å was achieved for the native crystals and those grown in the presence of DTT (Table 1).

Table 1.

Data Collection and Refinement Statistics

	FrnE native (PDB id 5CNW)	FrnE-DTT (PDB id 5CO3)	FrnE-βME (PDB id 5COH)	FrnE-TCEP (PDB id 5E59)
Data collection
Space group	P2₁22₁	P2₁22₁	P2₁22₁	P2₁22₁
Cell dimensions
a, b, c (Å)	47.8, 63.0, 87.3	47.7, 63.5, 86.3	47.9, 62.4, 86.7	47.8, 62.9, 86.5
α, β, γ (°)	90, 90, 90	90, 90, 90	90, 90, 90	90, 90, 90
Resolution (Å)	32.23–1.65 (1.68–1.65)^a	47.73–1.65 (1.68–1.65)^a	15.77–1.8 (1.84–1.80)^a	15.76–2.0 (2.05–2.0)^a
R _merge	0.041 (0.52)	0.034 (0.57)	0.073 (0.551)	0.126 (.671)
I/σI	17.2 (1.9)	23.5 (2.2)	15.9 (1.9)	19.1 (6.1)
Completeness (%)	99.2 (93.6)	98.7 (88.7)	99.6 (98.2)	99.8 (99.9)
Redundancy	3.6 (2.7)	5.0 (3.9)	4.9 (2.7)	10.1 (9.0)
Refinement
Resolution (Å)	1.65	1.65	1.8	2.0
No. of reflections	32,094	31,447	24,554	18,200
R _work/R _free	0.170/0.198	0.186/0.215	0.174/0.218	0.157/0.206
No. of atoms
Protein	1756	1788	1771	1771
Ligand/ion	6	6	0	12
Water	197	146	228	252
B-factors (Å²)
Protein	30.3	39.3	27.1	28.5
Ligand/ion	32.0	43.3		64.3
Water	30.9	41.6	35.3	38.4
R.m.s. deviations
Bond lengths (Å)	0.006	0.006	0.007	0.007
Bond angles (°)	0.92	0.87	0.88	0.90

One crystal was used for each structure.

Values in parentheses are for highest-resolution shell.

βME, β-mercaptoethanol; DTT, 1,4-dithiothreitol; TCEP, tris(2-carboxyethyl)phosphine.

Initial phases for the native protein crystals were obtained by the molecular replacement method by using a homologous structure available in the PDB (PDB ID, 3GL5; Chang et al., 10.2210/pdb3gl5/pdb) as search model, which was identified based on sequence identity detected by BLAST. The truncated coordinates of the search model (coordinates of C-terminus residues removed from the list and using upto Cβ atoms for nonidentical residues) were used in the molecular replacement calculations using PHENIX (1) implementation of the PHASER (30). The molecular replacement search clearly identified unique solution with an LLG score of 143.4. Electron density maps computed with these phases and observed amplitudes were of good quality such that most of the truncated residues/atoms could be located in these. The initial model for drFrnE was subsequently refined by PHENIX using maximum-likelihood target and with intermittent model fitting using COOT (6). The refinement progressed well with the final model including all the residues, except for about 10 residues on the N-terminus and 13 residues on the C-terminus for which electron density could not be observed.

The structures of drFrnE in the presence of reducing agents were solved by using the truncated model of the refined native structure and PHASER. These structures were also refined by using PHENIX and COOT suites (Table 1). The solvents in all the structures were identified during PHENIX cycles of refinement and were validated in the electron density maps by using COOT. Also, during the final stages of refinement, the TLS model implemented in PHENIX was employed for each structure. The refined structures were validated by using PHENIX. Each structure has been refined to high precision by using all the observed data. Nearly 98% of the residues are observed in the most favored regions of the Ramachandran plot. Residues 9–239 for structures in the presence of βME and TCEP, and residues 10–247 for structure in the presence of DTT (without 232–240), were clearly visible in the electron density maps computed with 2mFo-DFc coefficients at 1.0 sigma contour levels.

MD simulations

The structural model of drFrnE revealed two independent domains and extended the conformation of C-terminals residues harboring unique 239-CXXXXC-244 motif. Two different binding modes of C-tail residues were also seen in the drFrnE crystal structures. To elucidate flexibility of these domains and the C-terminus region, MD simulations were performed by using OPLS-AA/L force field and GROMACS package (39). The monomer of drFrnE was subjected to initial energy minimization, 100 ps temperature equilibration at 298 K using Berendsen thermostat, and 100 ps pressure equilibrium at 1 atm using Berendsen barostat. The production run trajectory was simulated for 2 ns in 1,000,000 steps of 2 fs each, and structures were written every 10 ps. The MD simulation estimated very high flexibility of the C-terminus residues with an estimated RMSF of ∼1 nm.

Bioinformatics and phylogenetic analysis

A BLASTp search with drFrnE sequence was carried out against each of the 27 bacterial phylum individually to mine “true” FrnE proteins in the bacterial kingdom based on the E-value and presence of both the active site CXXC and C-tail CX₄C motifs, and to probe for the global phylum-level diversity of FrnE. A small “representative” database of FrnE sequences was created by pooling a maximum of two sequences from each phylum. The selected two sequences from each phylum shared the highest and the lowest identity with drFrnE among the listed top 500 sequences (maximum) detected by BLAST search, such as to construct a diverse pool of FrnE proteins available in the bacterial kingdom.

Subcellular localization of representative sequences of FrnE protein, from each phylum, was predicted by using two independent servers, CELLO2GO (53) (http://cello.life.nctu.edu.tw/cello2go) and SignalP4.1 (38) (www.cbs.dtu.dk/services/SignalP). Signal P is a neural network-based method that uses both position and composition information to reliably identify signal peptide in protein sequences. For validation of the prediction scores, E. coli DsbA (ecDsbA) sequence, which is well known to be localized in the periplasmic space, and an ecFrnE sequence, which was found only in a few of E. coli strains, were also used in the analysis.

The presence of well-known cytoplasmic thiol oxido-reductase systems (Trx and Grx) and a glutathione biosynthesis system (Gsh) was probed in each of the species listed in the representative sequence database. For this purpose, ecTrx1, ecTrx2, ecGrx1, ecGrx2, ecGR, ecGshA, ecGshB, and ecGshF protein sequences were used as queries in the BLASTp analysis. The mined sequences, which shared >30% sequence identity and functional description as to be components of Trx, Grx, or Gsh systems in these species, were used against the better annotated E. coli genome to confirm the identity. Only those sequences that showed the highest identity with E. coli Trx, Grx, and Gsh system proteins were taken to be components of Trx and Grx systems in the searched bacterial species.

The structural homologs of drFrnE were searched for by using the DALI server (15) (http://ekhidna.biocenter.helsinki.fi/dali_server) by using refined coordinates of the wild-type drFrnE protein and also for the domain (insertion domain) consisting of 53–178 residues. The best match (Z-scores 28.7 and 13.2 for full-length protein and for insertion domain, respectively) was observed with PDB entry (3GL5), which has been annotated as putative DsbA oxido-reductase. Among the top hits, several uncharacterized proteins were listed and functionally annotated proteins such as mitochondrial GST class-kappa (PDB ID, 1R4W), Pseudomonas HCCA isomerase (PDB ID, 2IMD), B. subtilis BdbD (PDB ID 3EU4), and Legionella pneumophila DsbA (PDB ID, 4JRR) were detected to share significant structural homology with drFrnE protein.

Phylogenetic and molecular evolutionary analyses of the representative FrnE sequences and structurally similar proteins were conducted by using MEGA version 7.0 (50). Only one FrnE sequence from each phylum in the representative database was used in the phylogenetic analysis. The tree was computed by using representative sequences from a multiple sequence alignment that was computed by using PROMALS3D web server (37) (http://prodata.swmed.edu/promals3d). The evolutionary history was inferred by using the maximum-likelihood (9) method. All positions with <95%, typically, site coverage were eliminated. The evolutionary distances were computed by using the Poisson correction method (54). An identical tree was also observed by the maximum parsimony (48) and UPGMA clustering (31) methods, as implemented in MEGA 7.

Accession Codes

Coordinates and structure factors for X-ray structures have been deposited in the Protein Data Bank under accession codes 5CNW, 5CO3, 5COH, and 5E59.

Footnotes

Acknowledgments

The authors are thankful to the staff of PX-BL21 beamline at Indus-synchrotron, RRCAT, for their help with diffraction data acquisition.

Authors' Contributions

The author(s) have made the following declarations about their contributions: Conceived and designed the experiments: S.C.B., L.P., Y.S.R., H.S.M., and V.K. Performed X-ray diffraction and biophysical experiments: S.C.B., L.P., and V.K. Performed biochemical experiments: S.C.B., L.P., and Y.S.R. Performed bioinformatics analysis: S.C.B. and V.K. Performed site-directed mutagenesis and fluorescence microscopy experiments: Y.S.R. Contributed to data analysis and discussion: H.S.M. and V.K. Wrote the article: S.C.B., L.P., and V.K.

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Adams

, Afonine

, Grosse-Kunstleve

, Read

, Richardson

, and Terwilliger

. Recent developments in phasing and structure refinement for macromolecular crystallography. Curr Opin Struct Bio, 19: 566–572, 2009.

Battista

, Park

, and McLemore

. Inactivation of two homologous proteins presumed to be involved in the desiccation tolerance of plants sensitizes Deinococcus radiodurans R1 to desiccation. Cryobiology, 43: 133–139, 2001.

Begg

, Eijkelkamp

, Luo

, Counago

, et al. Dysregulation of transition metal ion homeostasis is the molecular basis for cadmium toxicity in Streptococcus pneumonia . Nat Commun, 6: 1–11, 2015.

Charaka

and Misra

. Functional characterization of the role of the chromosome I partitioning system in genome segregation in Deinococcus radiodurans . J Bacteriol, 194: 5739–5748, 2012.

Collet

and Messens

. Structure, function, and mechanism of thioredoxin proteins. Antioxid Redox Signal, 13: 1205–1216, 2010.

Emsley

, Lohkamp

, Scott

, and Cowtan

. Features and development of Coot. Acta Crystallogr D Biol Crystallogr, 66: 486–501, 2010.

Fadouloglou

, Kokkinidis

and Glykos

. Determination of protein oligomerization state: two approaches based on glutaraldehyde crosslinking. Anal Biochem, 373: 404–406, 2008.

Fang

and Dos Santos

. Protective role of bacillithiol in superoxide stress and Fe-S metabolism in Bacillus subtilis . Microbiol Open, 4: 616–631, 2015.

Felsenstein

Evolutionary trees from DNA sequences: a maximum likelihood approach. J Mol Evol, 17: 368–376, 1981.

10.

Fernandes

and Holmgren

. Glutaredoxins: glutathione-dependent redox enzymes with functions far beyond a simple thioredoxin backup system. Antioxid Redox Signal, 6: 63–74, 2004.

11.

Gon

, Faulkner

, and Beckwith

. In vivo requirement for glutaredoxins and thioredoxins in the reduction of the ribonucleotide reductases of Escherichia coli . Antioxid Redox Signal, 8: 735–742, 2006.

12.

Hawkins

and Freedman

. The reactivities and ionization properties of the active-site dithiol groups of mammalian protein disulphide-isomerase. Biochem J, 275: 335–339, 1991.

13.

Hawkins

, de Nardi

, and Freedman

. Redox properties and cross-linking of the dithiol/disulphide active sites of mammalian protein disulphide-isomerase. Biochem J, 275: 341–348, 1991.

14.

Helmann

JD.

Bacillithiol, a new player in bacterial redox homeostasis. Antioxid Redox Signal, 15: 123–133, 2011.

15.

Holm

and Rosenström

. Dali server: conservation mapping in 3D. Nucleic Acids Res, 38: W545–W549, 2010.

16.

Holmgren

Hydrogen donor system for Escherischia coli ribonucleoside diphosphate reductase dependent upon glutathione. Proc Natl Acad Soc U S A, 73: 2275–2279, 1976.

17.

Ito

and Inaba

. The disulfide bond formation (Dsb) system. Curr Opin Struct Biol, 18: 450–458, 2008.

18.

Kabsch

XDS. Acta Crystallogr D Biol Crystallogr, 66: 125–132, 2010.

19.

Khairnar

, Joe

, Misra

, Lim

, and Kim

. FrnE, a cadmium-inducible protein in Deinococcus radiodurans, is characterized as a disulphide isomerase chaperone in vitro and for its role in oxidative stress tolerance in vivo. J Bacteriol, 195: 2880–2886, 2013.

20.

Krisko

and Radman

. Biology of extreme radiation resistance: the way of Deinococcus radiodurans . Cold Spring Harb Perspect Biol, 5: a012765, 2013.

21.

Kumar

, Ghosh

, Poswal

, Pandey

, Jagannath , Hosur

, Dwivedi

, Makde

, and Sharma

. Protein crystallography beamline (PX-BL21) at Indus-2 synchrotron. J Synchrotron Radiat, 23: 629–634, 2016.

22.

Laurent

, Moore

, and Reichard

. Enzymatic synthesis of deocyribonucleotides. 4. Isolation and characterization of thioredoxin hydrogen donor from Escherischia coli B. J Biol Chem, 239: 3436–3444, 1964.

23.

Liang

, Fernandes

, Holmgren

, Li

, and Zhong

. Bacterial thioredoxin and thioredoxin reductase as mediators for epigallocatechin 3-gallate-induced antimicrobial action. FEBS J, 283: 446–458, 2016.

24.

López-Mirabal

and Winther

JR.

Redox characteristics of the eukaryotic cytosol, Biochim et Biophys Acta, 1783: 629–640, 2008.

25.

and Holmgren

. The thioredoxin antioxidant system. Free Radic Biol Med, 66: 75–87, 2014.

26.

, Vlamis-Gardikas

, Kandasamy

, Zhao

, Gustafsson

, Engstrand

, Hoffner

, Engman

, and Holmgren

. Inhibition of bacterial thioredoxin reductase: an antibiotic mechanism targeting bacteria lacking glutathione. FASEB J, 27: 1394–1403, 2013.

27.

, Chandrangsu

, Helmann

, Romsang

, Gaballa

, and Helmann

. Bacillithiol is a major buffer of the labile zinc pool in Bacillus subtilis . Mol Microbiol, 94: 756–770, 2014.

28.

Markillie

, Varnum

, Hradecky

, and Wong

. Targeted mutagenesis by duplication insertion in the radioresistant bacterium Deinococcus radiodurans: radiation sensitivities of catalase (katA) and superoxide dismutase (sodA) mutants. J Bacteriol, 181: 666–669, 1999.

29.

Maxwell

, Wildes

, Zarrine-Afsar

, De Los Rios

, Brown

, et al. Protein folding: defining a “Standard” set of experimental conditions and a preliminary kinetic data set of two-state proteins. Protein Sci, 14: 602–616, 2005.

30.

McCoy

, Grosse-Kunstleve

, Adams

, Winn

, Storoni

, and Read

. Phaser crystallographic software. J Appl Crystallogr, 40: 658–674, 2007.

31.

Michener

and Sokal

. A quantitative approach to a problem of classification. Evolution, 11: 490–499, 1957.

32.

Min-Ho

, Jung

S-W

, Im

S-H

, Lim

S-Y

, Song

H-P

, Kwon

, and Kim

D-H

. Genome-wide response of Deinococcus radiodurans on cadmium toxicity. J Microbiol Biotechnol, 21: 438–447, 2011.

33.

Nebert

and Vasiliou

. Analysis of the glutathione S-transferase (GST) gene family. Hum Genomics, 1: 460–464, 2004.

34.

Newton

, Rawat

, La Clair

, Jothivasan

, Budiarto

, Hamilton

, Claiborne

, Helmann

, and Fahey

. Bacillithiol is an antioxidant thiol produced in Bacilli. Nat Chem Biol, 5: 625–627, 2009.

35.

Obiero

, Pittet

, Bonderoff

and Sanders

. Thioredoxin system from Deinococcus radiodurans . J Bacteriol, 192: 494–501, 2010.

36.

Panicker

, Misra

, and Bihani

. Purification, crystallization and preliminary crystallographic investigation of FrnE, a disulfide oxidoreductase from Deinococcus radiodurans . Acta Crystallogr F, 70: 1540–1542, 2014.

37.

Pei

, Kim

B-H

, and Grishin

. PROMALS3D: a tool for multiple sequence and structure alignment. Nucleic Acids Res, 36: 2295–2300, 2008.

38.

Petersen

, Brunak

, von Heijne

, and Nielsen

. SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods, 8: 785–786, 2011.

39.

Pronk

, Páll

, Schulz

, Larsson

, Bjelkmar

, Apostolov

, Shirts

, Smith

, Kasson

, van der Spoel

, Hess

, and Lindahl

. GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics, 29: 845–854, 2013.

40.

Rajkarnikar

, Strankman

, Duran

, Vargas

, Roberts

, Barretto

, Upton

, Hamilton

, and Rawat

. Analysis of mutants disrupted in bacillithiol metabolism in Staphylococcus aureus . Biochem Biophys Res Comm, 436: 128–133, 2013.

41.

Rajpurohit

, Gopalakrishnan

, and Misra

. Involvement of protein kinase activity inducer in DNA double strand break repair and radioresistance of Deinococcus radiodurans . J Bact, 190: 3948–3954, 2008.

42.

Rajpurohit

and Misra

. Characterization of a DNA damage-inducible membrane protein kinase from Deinococcus radiodurans and its role in bacterial radioresistance and DNA strand break repair. Mol Microbiol, 77: 1470–1482, 2010.

43.

Rohr

, Hammerstad

, and Anderson

. Tuning of thioredoxin redox properties by intramolecular hydrogen bonds. PLoS One, 8: e69411, 2013.

44.

Sambrook

and Russell

. Molecular Cloning: a Laboratory Manual, 3rd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 2001.

45.

Santoro

and Bolen

. Unfolding free energy changes determined by the linear extrapolation method. 1. Unfolding of phenylmethanesulfonyl α-chymotrypsin using different denaturants. Biochemistry, 27: 8063–8068, 1988.

46.

Schaefer

, Schmitz

, Facius

, Horneck

, Milow

and Ortner

. Systematic study of parameters influencing the action of Rose Bengal with visible light on bacterial cells: comparison between biological effect and singlet-oxygen production. Photochem Photobiol, 71: 514–523, 2000.

47.

Sengupta

and Holmgren

. Thioredoxin and glutaredoxin-mediated redox regulation of ribonucleotide reductase. World J Biol Chem, 5: 68–74, 2014.

48.

Sneath

PHA

and Sokal

. Numerical Taxonomy. San Francisco: Freeman, 1973.

49.

Stroher

and Millar

. The biological roles of glutaredoxins. Biochem J, 446: 333–348, 2012.

50.

Tamura

, Peterson

, Stecher

, Nei

, and Kumar

. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol, 28: 2731–2739, 2011.

51.

Tian

, Xu

, Sun

, Lin

, and Hua

. Evaluation of the antioxidant effects of carotenoids from Deinococcus radiodurans through targeted mutagenesis, chemiluminescence, and DNA damage analyses. Biochim Biophys Acta, 1770: 902–911, 2007.

52.

Weiss

, Chen

, Ghigo

, Boyd

, and Beckwith

. Localization of FtsI (PBP3) to the septal ring requires its membrane anchor, the Z ring, FtsA, FtsQ, and FtsL. J Bacteriol, 181: 508–520, 1999.

53.

, Cheng

, Su

, Chang

, Huang

, Hwang

, and Lu

. CELLO2GO: a web server for protein subCELlularLOcalization prediction with functional gene ontology annotation. PLoS One, 9: e99368, 2014.

54.

Zuckerkandl

and Pauling

. Evolutionary divergence and convergence in proteins. In: Evolving Genes and Proteins, edited by Bryson

, and Vogel

. New York: Academic Press, 1965, pp. 97–166.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.22 MB