Iron Binding Properties of Recombinant Class A Protein Disulfide

Mar 30, 2017 - The protein disulfide isomerase (PDI) family comprises a wide set of enzymes mainly involved in thiol–disulfide exchange reactions in...
0 downloads 13 Views 2MB Size
Download PDF
Article pubs.acs.org/biochemistry

Iron Binding Properties of Recombinant Class A Protein Disulfide Isomerase from Arabidopsis thaliana William Remelli,†,‡,⊥ Stefano Santabarbara,‡ Donatella Carbonera,§ Francesco Bonomi,∥ Aldo Ceriotti,† and Anna Paola Casazza*,† †

Istituto di Biologia e Biotecnologia Agraria, Consiglio Nazionale delle Ricerche, Via Bassini 15a, 20133 Milano, Italy Istituto di Biofisica, Consiglio Nazionale delle Ricerche, Via Celoria 26, 20133 Milano, Italy § Dipartimento di Scienze Chimiche, Università di Padova, Via Marzolo 1, 35131 Padova, Italy ∥ Dipartimento di Scienze per gli Alimenti, la Nutrizione e l’Ambiente, DeFENS, Università di Milano, Via G. Celoria 2, 20133 Milano, Italy ‡

S Supporting Information *

ABSTRACT: The protein disulfide isomerase (PDI) family comprises a wide set of enzymes mainly involved in thiol− disulfide exchange reactions in the endoplasmic reticulum. Class A PDIs (PDI-A) constitute the smallest members of the family, consisting of a single thioredoxin (TRX) module without any additional domains. To date, their catalytic activity and cellular function are still poorly understood. To gain insight into the role of higher-plant class A PDIs, the biochemical properties of rAtPDI-A, the recombinant form of Arabidopsis thaliana PDI-A, have been investigated. As expressed, rAtPDI-A has only little oxidoreductase activity, but it appears to be capable of binding an iron−sulfur (Fe−S) cluster, most likely a [2Fe-2S] center, at the interface between two protein monomers. A mutational survey of all cysteine residues of rAtPDI-A indicates that only the second and third cysteines of the CXXXCKHC stretch, containing the putative catalytic site CKHC, are primarily involved in cluster coordination. A key role is also played by the lysine residue. Its substitution with glycine, which restores the canonical PDI active site CGHC, does not influence the oxidoreductase activity of the protein, which remains marginal, but strongly affects the binding of the cluster. It is therefore proposed that the unexpected ability of rAtPDI-A to accommodate an Fe−S cluster is due to its very unique CKHC motif, which is conserved in all higher-plant class A PDIs, differentiating them from all other members of the PDI family.

P

two or three a domains, whereas class 5 proteins possess only one TRX domain associated with additional domains or not. Photosynthetic organisms present a higher degree of variability in terms of domain composition. Therefore, starting from the phylogenetic analysis performed on rice, maize, and Arabidopsis,6,7 wheat,8 and Chlamydomonas reinhardtii,9 the original classification has been upgraded and extended to nine classes: PDI-A to -F, -L, -M, and -S.10 This nomenclature allows the classification of all known PDIs from photosynthetic organisms. Class A contains the smallest, ∼150 amino acids long, and simplest members of the PDI family, consisting in a single TRX module enclosing a CXXC motif, which is not associated with additional domains. The animal ortholog, ERp18, has been characterized as an ER-resident disulfide oxidase,11−13 but very little is known about members of the plant PDI-A class. They are expected to be ER-resident proteins because of the presence of an N-terminal signal peptide of 20−

rotein disulfide isomerases (PDIs) are the principal actors of oxidative protein folding in the endoplasmic reticulum (ER), catalyzing the formation, reduction, and isomerization of disulfide bonds in newly translated polypeptide chains.1 In general, PDIs contain at least one thioredoxin-like structural fold domain (β−α−β−α−β−α−β−β−α)2 and thus belong to the thioredoxin (TRX) protein superfamily (PF00085). Typically, they also contain at least one CXXC active site. In eukaryotes, PDIs are encoded by multigenic families, the main differences among the various members essentially lying in the number and position of TRX modules, in the active site sequence, and in the presence of additional protein domains. In mammals3 and yeast,4 PDIs are grouped in five classes on the basis of their architecture. Class 1 hosts the most extensively studied and founding member of the PDI family, i.e., the classical PDIs. These proteins possess two tandem repeats of TRX domains followed by a C-terminal acidic extension, c (a−b−b′−a′−c). The a and a′ domains are catalytically active, whereas the b′ domain provides the main substrate-binding site.5 PDIs from other classes contain either © XXXX American Chemical Society

Received: December 14, 2016 Revised: March 29, 2017 Published: March 30, 2017 A

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

disrupted by sonication (nine cycles of a 30 s burst at 10 μA in a MSE Soniprep 150 instrument, spaced by 1 min cooling on ice). After disruption, cellular debris was removed by centrifugation (2900g for 25 min at 4 °C), and the crude protein extract was filtered (0.25 μm) and immediately used for protein purification. For anaerobic protein extraction, all steps were performed in a N2 pre-equilibrated disposable anaerobic chamber (AtmosBag) using N2-saturated buffers. Protein Purification. Crude extracts were loaded on a 0.6 mL Ni-NTA resin (Qiagen) at a flow rate of 1 mL min−1. After being extensively washed with 50 mM Tris (pH 7.6), 300 mM NaCl, and 10 mM imidazole, proteins were eluted in the same buffer with a 10−50 mM step gradient of imidazole (from 10 to 300 mM). Purification was performed under either aerobic or anaerobic conditions. The latter took place in a N2 chamber using N2-saturated buffers. Purification yields were assessed by sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE),15 loading the fractions corresponding to the purified enzymes (300 mM imidazole for rAtPDI-A and its mutant derivatives and 100 mM imidazole for rAtPDIa). For long storage purposes (up to 6 months), the final elution from the Ni-NTA resin was performed in buffers supplemented with 40% (w/v) glycerol. The samples were either used immediately or stored in liquid nitrogen. Under these conditions, the sample formed a clear transparent matrix. The non-His-tagged version of rAtPDI-A (NrAtPDI-A) was partially purified by ionic exchange chromatography, using a Jasco high-performance liquid chromatography (HPLC) (PU1580/LG1580-02) system equipped with a diode array detector (MD1515). Crude cellular extracts were prepared as described above, if not for using 50 mM Tris (pH 7.6) and 50 mM NaCl. After the filtration step, cell extracts were loaded on a 1 mL DEAE column (Toyopearl 650S, Tosoh) at a flow rate of 0.25 mL min−1. Separation was performed by a stepwise NaCl gradient (from 10 to 500 mM). The richest NrAtPDI-A fraction eluted at 200 mM NaCl, as assessed by SDS−-PAGE. Western Blot Analysis. After SDS−PAGE, proteins were transferred on a 0.2 μm nitrocellulose membrane. The primary polyclonal antibody was produced in rabbits by Primm S.r.l. using rAtPDI-A as the antigen. The secondary antibody was a peroxidase-conjugated goat anti-rabbit immunoglobulin (Pierce). Signals were detected with the Supersignal West Pico Chemioluminescent Substrate (Pierce) using a ChemiDoc MP imaging system (Bio-Rad). Reductase Activity Assay. The ability of the proteins to catalyze the reduction of insulin was assayed by the turbidimetric method of Holmgren16 and expressed as reported by Martinez-Galisteo et al.17 The assay mixture contained 100 mM sodium phosphate (pH 7.0), 2 mM EDTA, 300 μM DTT, and 167 μM insulin. The protein concentration was 200 nM. Oxidase Activity Assay. The sulfhydryl oxidase activity was tested as described by Ruddock et al. 18 using a NRCSQGSCWN synthetic decapeptide (Primm S.r.l.) dissolved in 30% ACN and 0.1% TFA. The assay was performed in McIlvaine buffer (pH 6.5) supplemented with 0.5 mM GSSG, 2 mM GSH, and 3.2 μM decapeptide. The concentration of the recombinant enzyme was 200 nM. Quenching of the tryptophan fluorescence was followed for 900 s at 25 °C using a Peltier-equipped PerkinElmer LS-50B spectrofluorometer (λexc = 280 nm, and λem = 350 nm). The time-dependent fluorescence quenching is best described by a biexponential decay and a nondecaying component: F(t) = A1 exp(−t/τ1) + A2 exp(−t/τ2) + B. The activity was determined

30 amino acids and a C-terminal KDEL-like ER retention signal. However, the only experimental evidence of subcellular localization in maize indicates that ZmPDI-A (also called ZmPDIL5-1), which is strongly induced in response to ER stress, is not associated with the endosperm endomembrane fraction.6 Published expression profiles, relative to different tissues and developmental stages, are available only for maize, where the highest transcription level of ZmPDI-A has been detected in kernels,6 and for wheat, where the level of expression of TaPDI-A (TaPDIL6-1) was higher in roots than in all other tissues.8 There is, at present, a lack of information concerning both the PDI-A catalytic activity and the physiological role. The most recent report on this class of PDI isoforms regards barley PDI-A (HvPDIL5-1), which has been suggested to be a susceptibility factor to plant viral infections,14 as its absence makes barley resistant to Bymovirus. To acquire more information about single-domain higher-plant PDI-A isoforms, which all present a modified CKHC putative active site motif, we performed a biochemical and spectroscopic characterization of the recombinant protein from Arabidopsis thaliana (rAtPDIA). Recombinant AtPDI-A possesses only a marginal in vitro PDI activity. In addition, when accumulating in a suitable cellular environment, such as the cytoplasm of Escherichia coli, it is able to bind an iron−sulfur (Fe−S) cluster at the interface between two monomers. Site-specific mutations at the level of the CKHC motif allowed the identification of the key residues involved in iron binding.

■

EXPERIMENTAL PROCEDURES Chemicals and Various Materials. Unless otherwise stated, all chemicals were from Sigma. Plasmid Construction. The At1g07960 coding sequence corresponding to the mature form of AtPDI-A was amplified via reverse transcription polymerase chain reaction (RT-PCR) from total RNA extracted from 4-week-old A. thaliana plants, using primers carrying restriction sites for NdeI and XhoI. For the untagged version of recombinant AtPDI-A (NrAtPDI-A), a stop codon was inserted before the XhoI restriction site. To obtain control recombinant protein AtPDIa, consisting of the isolated a domain of a classical PDI, the At1g21750 gene encoding AtPDI-L1 was chosen. Only the sequence comprising the first catalytic domain (a domain) and the short linker sequence connecting it to the b domain were amplified as described above. The digested PCR products were cloned into the pET32b expression vector (Novagen) after excision of the Trx coding sequence. Single and double mutants were generated by site-directed mutagenesis (Stratagene) in one and two steps, respectively (see Figures S1−S3 for details of the constructs and oligonucleotide sequences). All constructs were checked by DNA sequencing and expressed in BL21 E. coli (Invitrogen). Bacterial Growth. E. coli was grown in Luria-Bertani medium supplemented with 2 mM Fe(III) citrate, 2 mM cystine, and 50 μg mL−1 ampicillin at 37 °C in an orbital shaker at 180 rpm. Protein overexpression was induced with 0.1 mM isopropyl β-D-thiogalactoside (IPTG) at an OD600 of 0.6 ± 0.1. After 20 h at 20 °C (100 rpm), cells were collected (2000g for 40 min at 4 °C), washed once in 50 mM Tris (pH 7.6) and 300 mM NaCl, gently pelleted, and stored at −20 °C until protein extraction was performed. Protein Extraction. Cells were suspended in homogenization buffer [50 mM Tris (pH 7.6) and 300 mM NaCl] and B

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 1. Oxidoreductase activity of recombinant AtPDI-A. (A) Reductase activity assay: (---) uncatalyzed, () rAtPDIa (200 nM), and (···) rAtPDI-A (200 nM). (B) Assay of the oxidase activity: (○) uncatalyzed, (△) rAtPDIa (200 nM), and (□) rAtPDI-A (200 nM). Solid lines are fits of the kinetic traces.

°C on anaerobically purified rAtPDI-A. Scans (350−750 nm) were repeated every 10 min at a scan speed of 100 nm min−1 for 17 h after admission of air to the cuvette. Circular Dichroism. CD spectra were recorded at 20 °C on freshly purified samples (∼0.1 OD cm−1 at 420 nm) in a N2saturated 1 mL anaerobic cuvette, using a Jasco J600 spectropolarimeter. The scan rate was 100 nm min−1 at 2 nm optical resolution. Presented CD spectra are the means of three independent biological replicates, each averaged over 21 scans. Electron Paramagnetic Resonance (EPR). Continuous wave X-band (9.38 GHz) EPR spectra were recorded on wildtype rAtPDI-A (300 μM Fe) in a Bruker Elexsys E580 spectrometer equipped with an ER4102ST cavity and a helium flow cryostat (ESR900, Oxford Instruments). Acquisition parameters were as follows: temperature, 10−50 K; modulation amplitude, 1 mT; microwave power, 10.2 mW. g values were estimated by calibration with a strong-pitch sample. EPR samples were prepared anaerobically and frozen in liquid nitrogen. When indicated, the protein was reduced by anaerobic addition of 20 mM sodium dithionite (final concentration), added as a small volume from a concentrated stock solution in anaerobic buffer. Dithionite-reduced samples were incubated for 10 min before being frozen. All samples were stored in liquid nitrogen until spectra were recorded.

as the inverse of the average decay lifetime, defined as τav = ΣiAiτi/ΣiAi. To allow comparison of enzyme activities obtained in this study with those previously reported for the human counterpart,11 the values of τav describing the oxidation kinetics of the respective positive controls were determined. For the human PDI a domain,11 τav was estimated to be ∼140 s, and for our positive control (AtPDIa), it was 165 ± 15 s. These τav values are therefore quite comparable. Analytical Gel Filtration. Analytical gel filtration was performed on a Sephacryl S200HR column [50 cm (length), 2 cm (diameter)] (GE Healthcare) connected to the same HPLC system described above. The column was typically equilibrated in 50 mM Tris (pH 7.6) and 300 mM NaCl and run at a flow rate of 0.5 mL min−1. As stated in the text, runs were also performed at different ionic strength and pH values, in the presence or absence of the reducing agent DTT and the surfactant NDSB201. Anaerobic samples were taken from the anaerobic chamber with gastight syringes and immediately injected into the column. Calibration was performed using the following markers: Dextran Blue, glucose oxidase, bovine serum albumin, chicken egg albumin, carbonic anhydrase, cytochrome c, and vitamin B12 (each at a concentration of 1 mg mL−1). Iron Quantification. The amount of Fe bound to the purified protein was estimated either using an average molar extinction coefficient of 8000 M−1 cm−1 at 420 nm for [2Fe-2S] clusters19,20 or by the bathophenanthroline sulfonate (BPS) method as described by Bonomi and Pagani21 with little modification. The assay mixture contained 1 mM BPS and 20 mM sodium dithionite added to extensively degassed buffers. The amount of adventitious iron present in the reaction mixture was taken into account by following the absorbance changes at 535 nm until a stable reading was established. Aliquots of the purified protein were then added to the solution, and any further absorbance increase was monitored for 700 s. An extinction coefficient of 22100 M−1 cm−1 at 535 nm (on iron basis) was used for calculation. Absorption Spectroscopy. Freshly purified samples were transferred to a N2-saturated 1 mL anaerobic cuvette. ultraviolet−visible (UV−vis) absorption spectra (260−750 nm) were recorded at 25 °C using either a V360 or a V550 Uvidec spectrophotometer (Jasco) at a scan speed of 100 nm min−1. The presented spectra are the mean of at least eight measurements. Oxygen stability assays were performed at 10

■

RESULTS Recombinant AtPDI-A Is Not an Efficient Thiol/ Disulfide Oxidoreductase in Vitro. Considering their overall primary sequence, PDI-A isoforms resemble the isolated a (or a′) domain of classical PDI. This is usually considered the minimal unit required for the catalysis of thiol−disulfide exchange reactions,22 whereas the presence of an additional high-affinity substrate-binding domain is required for disulfide bonds isomerization.23 The main determinants of catalytic activity are the sequence of the active site and the presence of conserved residues that influences the pKa of active site cysteines. All eukaryotic PDIs from class A possess the minimal features required for acquiring a TRX-like fold and for performing thiol−disulfide exchange reactions.10,11 However, the most striking difference among them resides in the sequence of the putative CXXC active site (Figure S4). Class A PDIs from algae and mosses possess the classical PDI motif (CGHC) and are expected to be efficient oxidases, although no C

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry experimental evidence is available. All known animal PDI-A isoforms present a nonconservative substitution at position +3, where an alanine replaces the histidine (CGAC vs CGHC). These proteins can catalyze in vitro the formation of a disulfide bond in a peptide substrate, as reported for the human isoform ERp18.11 Higher-plant PDI-A isoforms are unique in having a lysine at position +2 instead of the glycine (CKHC), and no obvious prediction can be made about their catalytic activity. To investigate the biochemical properties of higher-plant single-domain PDI-A, we choose as a model protein that from A. thaliana (AtPDI-A), which is encoded by the single gene At1g07960. We overexpressed its mature form in E. coli (rAtPDI-A) and purified it to homogeneity by affinity chromatography, using a His tag at its C-terminus. For comparison, we also produced, employing the same experimental strategy, the recombinant form of a truncated a−b−b′− a′ PDI, consisting of only the isolated a domain (rAtPDIa). These overexpressed proteins were predominantly found in the soluble fraction of cell lysates. Their apparent molecular mass, as determined by SDS−PAGE under reducing conditions, agreed with that predicted from the engineered gene sequences. The reductase activity was tested by monitoring the ability to catalyze the cleavage of the disulfide bonds between the α and β chains of insulin.16 Figure 1A shows that the activity of rAtPDIA (4.25 ± 1.14 units/mg of protein) is 3-fold lower than that of rAtPDIa (13.48 ± 2.55 units/mg of protein), the isolated a domain of classical PDI. Although insulin reduction by rAtPDIA is significantly distinguishable from that caused by DTT alone in the uncatalyzed reaction, its extent may be too marginal to account for a significant biological function. The oxidase activity was assessed by following the formation of a disulfide bond in a synthetic decapeptide,18 as previously described for human ERp18.11 Data presented in Figure 1B show that the extent of substrate oxidation is not distinguishable from that of the uncatalyzed reaction. On the other hand, rAtPDIa shows significant dithiol oxidizing activity, which is comparable to that of the human PDI isolated a domain (ref 11 and Experimental Procedures). To verify if the lack of PDI-like activity observed for rAtPDIA may be due to the presence of lysine at position +2, we converted the CKHC site into the canonical CGHC site and tested the mutant rAtPDI-A K56G (K56G) for catalytic activity. Figure 2 clearly shows that this is not the case, because no significant differences could be detected with respect to the wild-type protein. Another peculiar feature differentiating animal and higherplant PDI-A is the presence of an additional cysteine (C51) four amino acids before the putative catalytic site in the plant proteins. To verify whether this residue can influence the activity of rAtPDI-A, we generated and analyzed single-point mutant rAtPDI-A C51S (C51S). As shown in Figure 2, the conservative substitution of C51 with a serine also has no effect on the in vitro thiol/disulfide oxidoreductase activity of rAtPDIA. Recombinant AtPDI-A Can Bind an Fe−S Cluster. The presence of C51 results in a peculiar CXXXCXΦC amino acid stretch (where Φ is an aromatic residue), which is consistent with the consensus sequence of a putative Fe−S cluster-binding motif (e.g., ref 24 and references therein). In particular, this binding motif resembles that found in most members of the radical S-adenosylmethionine (SAM) superfamily (PF04055).25 In radical SAM enzymes, a [4Fe-4S] cluster is coordinated by the three cysteine residues of this motif, which leaves one of the

Figure 2. Oxidoreductase activity of recombinant AtPDI-A and its sitespecific mutants. (A) Insulin reduction assay. Activity is expressed as reported by Martinez-Galisteo et al.17 The significance of the data was assessed by the Student’s t test against the value of the uncatalyzed reaction: *p < 0.05, and **p < 0.01. (B) Peptide oxidation assay. Activity was calculated as the inverse of the average quenching lifetimes as described in Experimental Procedures. Data are the means from at least three independent biological replicates. When present, the recombinant protein concentration was 200 nM in all assays.

four iron atoms free to coordinate SAM.26 It seemed therefore interesting to investigate whether rAtPDI-A could bind an Fe− S cluster. The recombinant protein was expressed under growth conditions favoring the assembly of a cluster and purified under anaerobiosis, because numerous Fe−S proteins are known to be sensitive to oxygen. The anaerobically purified rAtPDI-A has the typical brown coloration of a non-heme iron-binding protein, and its absorption spectrum (Figure 3A) is consistent with the binding of an Fe−S cluster. In particular, the spectrum shows a broad absorption across the full visible region, displaying defined peaks at 330 and 417 nm, together with shoulders around 465 and 550 nm, which are proposed to arise from dipolar allowed sulfur and iron charge transfer transitions.27 This spectral pattern is similar to that of several [2Fe-2S] proteins and different from that of proteins containing a [4Fe-4S] cluster.19 All these spectral features are essentially suppressed either when rAtPDI-A is purified under aerobic conditions or when the anaerobically purified protein is reduced with dithionite. Quantification of the iron bound to the protein, based on the extinction coefficients of (i) [2Fe-2S] clusters19,20 and (ii) aromatic residues of the protein, gave ∼0.4 metal atom per polypeptide, indicating that the cluster is bound substoichiometrically. Figure 3A also shows that the absorption spectrum of a recombinant version of AtPDI-A lacking the Cterminal His tag (NrAtPDI-A) is virtually indistinguishable D

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 3. Spectroscopic characterization of recombinant AtPDI-A. (A) UV−vis spectra of rAtPDI-A purified under anaerobic () or aerobic (−·−) conditions. Also shown are the spectra of the anaerobically purified rAtPDI-A incubated with 10 mM sodium dithionite (---) and the non-His-tagged variant, NrAtPDI-A (−··−). The protein concentration was 77 μM for all samples, and the absorption was further normalized at 280 nm for ease of comparison. Note the break in the scale to highlight the spectral features in the visible region. (B) Visible region CD spectra of rAtPDI-A purified under anaerobic conditions before () and after the addition of 10 mM sodium dithionite (···). The spectrum of rAtPDI-A purified under aerobic conditions is also shown (−·−). The intensity of the CD signal was normalized to the absorption at 420 nm of each untreated sample.

from that of the His-tagged version. This indicates that the histidine tail is not involved in iron binding. To gain further insight into the nature of the cluster, the samples were also analyzed by visible circular dichroism (CD) spectroscopy. In comparison to the absorption spectrum, the CD spectrum of anaerobically purified rAtPDI-A exhibits a more defined structure, demonstrating that the broadness of the absorption spectrum is due to the overlap of several unresolved transitions. The CD spectra present relative maxima at 368, 478, and 552 nm and relative minima at 405 and 635 nm. In agreement with those observed in the absorption spectra, these features are strongly suppressed when the protein is purified under aerobic conditions or when it is reduced with dithionite (Figure 3B). Again, the general shape of the CD spectrum agrees with that reported for other biological [2Fe2S] clusters,28,29 although the positions of the peaks appear to be red-shifted. With the aim of investigating the type of Fe−S cluster bound to rAtPDI-A, electron paramagnetic resonance (EPR) spectra of the anaerobically purified protein were recorded in the presence or absence of dithionite. As isolated, rAtPDI-A is EPR silent, whereas upon reduction with dithionite, an EPR-active species is formed (Figure S5). The signal is, however, extremely weak, suggesting that the Fe−S cluster is highly unstable upon the addition of dithionite. This instability hampered further characterization. Many, but not all, Fe−S clusters are known to be sensitive to molecular oxygen. To investigate the stability of the cluster, anaerobically purified rAtPDI-A was exposed to air and the changes in absorbance were monitored over time (Figure 4). The kinetics of O2-induced degradation (inset of Figure 4), as monitored by the decrease in the absorption at 465 and 550 nm, indicate that the cluster is stable for approximately 1 h with a half-life of ∼4 h. It is worth noting that the absorption signal is not completely suppressed even after O2 exposure for 17 h, yet most of the spectral features centered at 465 and 550 nm cease to be resolved. Identification of Cluster Ligands. To identify the residues involved in the binding of the Fe−S cluster, a mutational survey of all cysteine residues in rAtPDI-A was

Figure 4. Effects of air exposure on the visible absorption spectra of rAtPDI-A purified under anaerobic conditions. Spectra corresponding to the initial and final points in the time course (air exposure for 17 h) are presented as solid lines. Intermediate points are presented as dashed lines. The inset shows the decrease in absorption monitored at 465 nm (□) and 550 nm (●) as a function of air exposure time.

performed. In addition to the three conserved cysteines of the putative catalytic site, all higher-plant PDI-A isoforms possess two additional cysteines (C85 and C92), separated by six nonconserved residues [C(X6)C motif]. Three single-site Cysto-Ser mutants (C51S, C55S, and C58S) and one double mutant (C85S/C92S) were generated and expressed under conditions leading to the incorporation of an Fe−S cluster in the wild-type protein. Even though all mutants could coordinate iron atoms, as inferred from the presence of absorption around 420 nm (Figure 5A), the extent of incorporation was lower than in the wild type (∼50%). Moreover, only the absorption spectra of C55S and C58S showed significant differences with respect to that of the wildtype protein, as the shoulders centered at 465 and 550 nm were E

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 5. Spectroscopic characterization of wild-type rAtPDI-A and its mutants. (A) Absorption and (B) CD spectra of rAtPDI-A (black solid lines) and its mutant derivatives, rAtPDI-A K56G (gold dashed−dotted lines), rAtPDI-A C51S (red dashed−dotted lines), rAtPDI-A C55S (green dashed−dotted lines), rAtPDI-A C58S (blue dotted lines), and rAtPDI-A C85S/C92S (magenta dotted lines). The absorption spectra are the means of at least three independent biological replicates and are normalized to 1 OD cm−1 at 280 nm. The inset of panel A shows a magnification of the visible absorption with spectra normalized at 420 nm to allow band shape comparison. The CD intensity is scaled to yield equal visible absorption of 0.1 OD cm−1 at 420 nm.

demonstrate that neither the residues directly involved in cluster binding nor those affecting the cluster geometry led to the restoration of any significant oxidoreductase activity. The rAtPDI-A Dimer Is the Minimal Oligomeric State Capable of Binding the Fe−S Cluster. As two cysteines are not sufficient to coordinate an Fe−S cluster, it could be envisaged that the cluster is bound at the interface of two rAtPDI-A monomers, i.e., by a dimer, as already described, for instance, for some plant and human glutaredoxins.30,31 The oligomeric state of the anaerobically purified rAtPDI-A was investigated by size exclusion chromatography, in the presence or absence of a reducing agent (Figure 6). Control experiments were also performed on soluble extracts of E. coli overexpressing rAtPDI-A, either His-tagged or not (Figure 7).

largely suppressed (inset of Figure 5A). C/S substitutions at residues 55 and 58 also led to considerable alterations in the CD spectra and to a strong decrease in the signal intensity (Figure 5B). Conversely, substitution of the cysteine at position 51 with serine caused only minor differences in both the CD and the absorption spectrum with respect to those of rAtPDI-A. However, the subtle changes in the C51S CD spectrum might suggest a slight modification in the geometry of the Fe−S cluster coordination. Similarly, substitution of both cysteines at positions 85 and 92 (C85S/C92S) did not significantly alter the position of the diagnostic absorption shoulders and of CD maxima, although some differences in intensity were observed. These data indicate that the two cysteines at positions 55 and 58 are primarily involved in cluster coordination. Residues C85 and C92 appear to play only a marginal role, and C51 is not involved altogether. Thus, rAtPDI-A does not belong to the radical SAM enzyme superfamily despite the presence of the CXXXCXΦC consensus motif. The analysis of site-specific mutants clearly showed that the first and second cysteine residues of the CKHC motif are involved in cluster coordination. However, the control protein rAtPDIa, which possesses a canonical CGHC motif, cannot bind a cluster upon overexpression in E. coli under permissive conditions (data not shown). We have therefore investigated the impact of the lysine at position +2 in rAtPDI-A, which corresponds to a glycine in rAtPDIa, on its ability to incorporate an Fe−S cluster. To address this question, we analyzed the anaerobically purified K56G mutant. As observed for the cysteine to serine substitutions, the K56G mutant also appears to be capable of coordinating iron atoms, as inferred from its absorption spectrum (Figure 5A). However, the CD signal (Figure 5B) is considerably altered with respect to that of the wild-type protein regarding both the intensity and the general band shape. Thus, the substitution of lysine strongly affects the iron binding properties of rAtPDI-A. Catalytic Activity of Site-Specific Mutants. To test the influence of the residues putatively involved in cluster coordination on the catalytic activity of rAtPDI-A, disulfide reduction and dithiol oxidation assays were performed for all site-directed mutants. The results shown in Figure 2

Figure 6. Analysis of the oligomeric state of the anaerobically purified rAtPDI-A. Size exclusion chromatography profiles of rAtPDI-A in the presence (gray lines) or absence (crimson lines) of the reducing agent DTT, monitored at 280 nm (solid lines), to detect the overall protein content, and at 330 nm (dashed lines), to specifically detect bound iron. On the basis of column calibration, peaks and shoulders in the chromatograms can be identified as follows: 38 mL (>200 kDa oligomers), 48 mL (∼112 kDa octamers), 64 mL (28 kDa dimers), and 71 mL (14 kDa monomers). F

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

Figure 7. Analysis of the oligomeric state of rAtPDI-A in the anaerobic cellular extracts. (A) Size exclusion chromatography profiles of soluble protein extracts obtained from E. coli cultures in which overexpression of rAtPDI-A was either induced (crimson lines) or not induced (black lines), monitored at 280 nm (thin lines) and at 330 nm (thick lines). The inset shows the Western blot analysis of the fractions corresponding to the peaks attributed to the polymeric (elution volume of 38 mL), dimeric (elution volume of 64 mL), and monomeric (elution volume of 71 mL) forms of rAtPDI-A. Also shown in the blot are the total cellular extract before being loaded on the column [CE, crude cellular extract; (+), induced, positive control; (−), not induced, negative control]. A polyclonal anti-rAtPDI-A antibody was used for immunodetection. (B) Comparison of the size exclusion chromatograms obtained for soluble protein extracts overexpressing either the C-terminally His-tagged version of rAtPDI-A (+Tag; crimson lines) or its untagged version (NrAtPDI-A, −Tag; blue lines). Profiles were monitored at 280 nm (thin lines) and 330 nm (thick lines).

respectively, in general agreement with the observations discussed above for the bulk purified protein. Thus, iron appears to be substoichiometrically bound irrespective of the oligomeric state of rAtPDI-A. Even using the highly sensitive bathophenanthroline sulfonate method (see Experimental Procedures), Fe quantification fell behind the detection threshold for the monomeric and dimeric fractions, yet the chromatograms monitored at 330 nm (Figure 6) clearly showed Fe absorption at elution volumes corresponding to dimers, which decreased significantly for monomers. This indicates that dimers represent the smallest oligomeric state capable of coordinating the cluster. Similar conclusions were obtained for the analysis of crude extracts (Figure 7).

In all samples investigated, the chromatograms indicated the presence of a heterogeneous protein population. A reliable deconvolution of the anaerobically purified rAtPDI-A chromatograms, in terms of a linear combination of asymmetric Gaussian sub-bands, was obtained from the analysis of several independent replicates. This indicated that ∼75% of the protein eluted as soluble high-molecular weight oligomers, having a molecular mass of >200 kDa. The other bands, corresponding to molecular masses equivalent to those expected for an octamer (∼112 kDa), a dimer (28 kDa), and a monomer (14 kDa), accounted for 11−15, 8−10, and 1−2% of the total, respectively. Analogous results were obtained for crude extracts (Figure 7). Although the relative contribution of each fraction varied slightly according to the type of preparation, highmolecular weight oligomers were always predominant (>70%) and stable under a variety of conditions. In fact, the elution profile was not appreciably affected by (a) the presence of the reducing agent DTT (Figure 6), (b) a decrease in ionic strength (10−200 mM NaCl), (c) a change in the pH (6.0− 7.8) of the elution buffer, or (d) performing extraction and analysis in the presence of the surfactant NDSB201 (500 mM)32 in an attempt to preclude or disrupt the tight association of the protein into larger polymers and/or to promote their solubilization once they are formed. Noticeably, this was observed also for the nontagged version of the protein (NrAtPDI-A), demonstrating that the His tail acts neither as a nonspecific cluster ligand nor as an aggregation-promoting region (Figure 7B), contrary to what was reported, for example, for animal His-tagged ERp29.33 The same holds true for the C85S/C92S double mutant, indicating that the propensity of rAtPDI-A to form high-molecular weight oligomers is not due to the formation of intermolecular disulfides, at least not to those involving these two distal cysteines (Figure S6). Quantification of the iron content in the purified protein was possible for only the two highest oligomers. The Fe/protein ratios were ∼0.2 and ∼0.4 for the largest aggregation state (>200 kDa) and the putative octamer (∼112 kDa),

■

DISCUSSION In silico analysis of the product of the At1g07960 gene from A. thaliana (AtPDI-A) led to its classification as a PDI-like protein, which is predicted to be localized in the endomembrane system and to be involved in cell redox homeostasis. The results presented here strongly suggest that the main role of AtPDI-A may not be disulfide bond formation and/or reduction, even though, like all higher-plant PDI-A isoforms, it possesses several conserved features known to be important for oxidoreductase activity (see Figure S4 for sequence comparison). When its recombinant form (rAtPDI-A) was tested in canonical in vitro reductase/oxidase assays, it showed only marginal activity (Figures 1 and 2), which is different from what was observed for its animal counterpart, ERp18.11 This is, however, not so surprising, as besides sharing the same modular organization, AtPDI-A and ERp18 significantly diverge at the level of the active site (CKHC vs CGAC) as well as for the overall protein primary sequence, exhibiting only ∼15% identity. The lack of PDI-like activity observed for rAtPDI-A may be therefore linked to its very unique active site (CKHC), which differs from that encountered in classical PDIs (CGHC). In fact, the XX dipeptide of the CXXC motif is known to be the main factor G

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

It is generally recognized that four ligands are mandatory for coordinating most types of Fe−S clusters, although on several occasions interactions among similar (or different) polypeptides may be necessary to fulfill this requirement. Monothiol glutaredoxins (Grx) and scaffold proteins involved in the assembly of Fe−S clusters are among the most studied example of this behavior.42,43 In the case of AtPDI-A, C55 and C58 alone should not be sufficient to bind the cluster. The analysis of size exclusion chromatograms (Figures 6 and 7) suggests that the cluster could be coordinated via the four cysteine thiolates of the CKHC motif, assuming they are located at the interface of two adjacent rAtPDI-A monomers. A similar behavior was observed for E. coli TrxA mutants, in which the canonical catalytic site CGPC was mutated into CACC or CACA. These mutations turned this protein from a monomeric disulfide reductase into a [2Fe-2S] bridged dimer capable of catalyzing O2-dependent sulfhydryl oxidation in vitro.44 The crystal structure of the CACA variant may explain how mutations within an existing scaffold can allow a protein to bind an Fe−S cluster and to acquire a markedly divergent function. As opposed to wild-type TrxA, also the second cysteine in the CACA mutant is surface-exposed, allowing the coordination of an Fe−S cluster.45 Therefore, it is reasonable to hypothesize that the KH dipeptide in AtPDI-A may exert a similar effect, rendering the C-terminal cysteine more accessible and capable of coordinating a [2Fe-2S] cluster at the dimer interface. The importance of the lysine residue in cluster binding was confirmed by substituting it with an inherently more flexible residue such as glycine (K56G mutant). This led not only to a further reduction in the efficiency of cluster coordination with respect to that of the wild type but also to a strong geometrical distortion of the Fe−S cluster (Figure 5). Similarities can also be found with glutaredoxins (Grx), another group of small oxidoreductases that belong to the TRX superfamily, mainly involved in the reduction of glutathionylated proteins. Noticeably, some class I Grxs (including poplar GrxC130 and human Grx231) and some class II Grxs (characterized by a conserved CGFS active site sequence42,46) form homodimers that are bridged together by a [2Fe-2S] cluster. As opposed to what is suggested here for rAtPDI-A, in Fe−S Grxs the cluster is ligated by the N-terminal active site cysteines of each monomer, whereas the other ligands come from two glutathione molecules. Very recently, a study by Bisio et al.47 has been published regarding a new class of TRX-related proteins from the human pathogen Echinococcus granulosus (IsTRP) that can coordinate an Fe−S cluster in a GSHindependent manner. As proposed here for rAtPDI-A, the binding of the cluster in IsTRP was shown to involve the two cysteine residues of the CXXC motif and to rely on protein dimerization. However, Fe−S Grxs, which have been proposed to be involved in iron homeostasis (ref 48 and references therein), as well as the newly discovered IsTRP, are located in the cytosol or in subcellular organelles like chloroplasts and mitochondria, where specific Fe−S cluster assembly machineries are present.43,49 AtPDI-A is assigned to the secretory pathway instead. This suggestion is corroborated by proteomic data50 that identify an N-terminal tryptic peptide (EVITLTPETFSDK) that lacks the tryptic start, indicating that the protein has been processed and the first 25 amino acids have been removed accordingly to presequence cleavage site predictions. In addition, transient expression of fluorescent fusions indicates that (i) AtPDI-A possesses a functional N-terminal signal peptide and (ii) the C-terminal DKEL sequence works

modulating the redox-active properties of thiol/disulfide oxidoreductases.34−36 Because of its amphipathic nature, the lysine at position +2 may alter the local environment of the active site cysteines. On one hand, the positive charge on the lysine side chain may influence the electrostatic interactions affecting the pKa of the N-terminal cysteine and consequently the efficiency of the protein in catalyzing thiol−disulfide exchange reactions. On the other hand, its long hydrophobic side chain may induce conformational changes modifying the accessibility of the C-terminal cysteine, which is normally buried and unreactive. The substitution of lysine with glycine, which reverts the CKHC motif into a canonical PDI active site, did however not restore any significant catalytic activity (Figure 2). A similar observation was reported previously for the recombinant poplar PDI-A isoform,37 suggesting that the sole presence of the CGHC consensus sequence is not sufficient to confer oxidoreductase activity to higher-plant class A PDIs. The very marginal enzymatic activity of these proteins appears therefore to be related to other, yet unidentified, structural factors. Other evidence that higher-plant PDI-A may differ significantly from other members of the PDI family is that, when expressed in a heterologous system under permissive conditions, rAtPDI-A can coordinate, albeit substoichiometrically, an oxygen-labile Fe−S cluster (Figures 3 and 4). The type of cluster could not be determined by EPR, probably because of its instability upon reduction with dithionite (which was necessary to produce an EPR detectable oxidation state of the cluster), even under minimal exposure to the reducing agent followed by rapid freezing. However, optical spectroscopy data on the oxidized form of the protein strongly hint at the presence of a [2Fe-2S] cluster rather than a [4Fe-4S] cluster. Remarkably, of the five cysteines of AtPDI-A, only the two composing the putative active site are directly involved in cluster coordination, because their substitution with serine resulted in substantial changes in the visible absorption and in almost complete suppression of the CD spectra (Figure 5). The very weak residual CD signal observed in these mutants might arise from the formation of a cluster with mixed (cysteinyl and non-cysteinyl) coordination, as observed, for instance, in cysteine-to-serine mutants of cyanobacterial Anabaena ferredoxin38,39 and human ferredoxin.40 The histidine located at position +3 in the putative active site might also participate in cluster coordination. However, CD signal suppression in the rAtPDI-A C58S mutant argues against its direct involvement as well as its ability to substitute for the adjacent cysteine (C58) as a ligand. Cysteine mutants at other positions (rAtPDI-A C51S and rAtPDI-A C85S/C92S) can still bind a cluster, but with a slightly distorted geometry and with a lower efficiency. C51 lies very close to the CKHC site, whereas C85 and C92 are relatively distant. No structural information is available for class A PDIs of higher plants, or for other PDIs from photosynthetic organisms containing the C(X6)C motif. However, the distribution of cysteines in their sequence is comparable to that of the third catalytic domain of animal ERp46, for which a crystal structure is available.41 In the third domain of ERp46, the two distal cysteines form a structural disulfide, which is in the proximity of the active site. Therefore, in analogy to ERp46, it can be hypothesized that amino acid substitutions at residues C85 and C92 can affect the protein conformation in the vicinity of the cluster, thereby explaining the different CD spectrum of the C85S/C92S mutant. H

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry

CD measurements, and Stefania Iametti (Università di Milano) for helpful discussion about Fe−S clusters.

efficiently in ER localization of soluble proteins (A. P. Casazza, personal observation). Because there is no evidence of the presence of a cluster assembly machinery in this compartment, the significance of the observed ability of rAtPDI-A to bind an Fe−S cluster remains unclear. However, at this stage, we cannot exclude the possibility that AtPDI-A, as shown for other members of the PDI family,51 may have dual/multiple localization (constitutively or in response to particular stress conditions) and may play different roles in different cellular compartments. When present in a suitable environment, AtPDI-A might coordinate an Fe−S cluster, exerting a function that remains a mystery. When located in the ER, AtPDI-A might also be involved in the binding of iron atoms. This ability might contribute, to some extent, to the scavenging of free iron, thus preventing, for instance, the generation of metal-catalyzed reactive oxygen species. However, the latter shall be considered only as a reasonable suggestion. The analysis of AtPDI-A subcellular localization as well as of its ability to bind iron/Fe− S clusters in plant cells will be required to further clarify these issues. However, at present, the very low level of AtPDI-A accumulation in plant tissues strongly limits the possibility of performing quantitative analysis in vivo.

■

■

ABBREVIATIONS ER, endoplasmic reticulum; Fe−S, iron−sulfur; Grx, glutaredoxins; PDI, protein disulfide isomerase; SAM, S-adenosylmethionine; TRX, thioredoxin.

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.6b01257. Construct details, protein and oligonucleotide sequences, sequence alignment and comparison, EPR spectrum of purified protein, size exclusion chromatography of the rAtPDI-A C85S/C92S mutant, and analysis of the oligomeric states of all other site-directed mutants (PDF)

■

REFERENCES

(1) Hatahet, F., and Ruddock, L. W. (2009) Protein Disulfide Isomerase: A Critical Evaluation of Its Function in Disulfide Bond Formation. Antioxid. Redox Signaling 11, 2807−2850. (2) Ferrari, D. M., and Söling, H. D. (1999) The protein disulphideisomerase family: unravelling a string of folds. Biochem. J. 339, 1−10. (3) Appenzeller-Herzog, C., and Ellgaard, L. (2008) The human PDI family: versatility packed into a single fold. Biochim. Biophys. Acta, Mol. Cell Res. 1783, 535−548. (4) Norgaard, P., Westphal, V., Tachibana, C., Alsøe, L., Holst, B., and Winther, J. R. (2001) Functional differences in yeast protein disulfide isomerases. J. Cell Biol. 152, 553−562. (5) Pirneskoski, A., Klappa, P., Lobell, M., Williamson, R. A., Byrne, L., Alanen, H. I., Salo, K. E., Kivirikko, K. I., Freedman, R. B., and Ruddock, L. W. (2004) Molecular characterization of the principal substrate binding site of the ubiquitous folding catalyst protein disulfide isomerase. J. Biol. Chem. 279, 10374−10381. (6) Houston, N. L., Fan, C., Xiang, J. Q., Schulze, J. M., Jung, R., and Boston, R. S. (2005) Phylogenetic analyses identify 10 classes of the protein disulfide isomerase family in plants, including single-domain protein disulfide isomerase-related proteins. Plant Physiol. 137, 762− 778. (7) Lu, D. P., and Christopher, D. A. (2008) Endoplasmic reticulum stress activates the expression of a sub-group of protein disulfide isomerase genes and AtbZIP60 modulates the response in Arabidopsis thaliana. Mol. Genet. Genomics 280, 199−210. (8) d’Aloisio, E., Paolacci, A. R., Dhanapal, A. P., Tanzarella, O. A., Porceddu, E., and Ciaffi, M. (2010) The protein disulfide isomerase gene family in bread wheat (T. aestivum L.). BMC Plant Biol. 10, 101. (9) Lemaire, S. D., and Miginiac-Maslow, M. (2004) The thioredoxin superfamily in Chlamydomonas reinhardtii. Photosynth. Res. 82, 203− 220. (10) Selles, B., Jacquot, J. P., and Rouhier, N. (2011) Comparative genomic study of protein disulfide isomerases from photosynthetic organisms. Genomics 97, 37−50. (11) Alanen, H. I., Williamson, R. A., Howard, M. J., Lappi, A.-K., Jäntti, H. P., Rautio, S. M., Kellokumpu, S., and Ruddock, L. W. (2003) Functional Characterization of ERp18, a New Endoplasmic Reticulum-located Thioredoxin Superfamily Member. J. Biol. Chem. 278, 28912−28920. (12) Jeong, W., Lee, D.-Y., Park, S., and Rhee, S. G. (2008) ERp16, an Endoplasmic Reticulum-resident Thiol-Disulfide Oxidoreductase. J. Biol. Chem. 283, 25557−25566. (13) Rowe, M. L., Ruddock, L. W., Kelly, G., Schmidt, J. M., Williamson, R. A., and Howard, M. J. (2009) Solution Structure and Dynamics of ERp18, a Small Endoplasmic Reticulum Resident Oxidoreductase. Biochemistry 48, 4596−4606. (14) Yang, P., Lüpken, T., Habekuss, A., Hensel, G., Steuernagel, B., Kilian, B., Ariyadasa, R., Himmelbach, A., Kumlehn, J., Scholz, U., Ordon, F., and Stein, N. (2014) PROTEIN DISULFIDE ISOMERASE LIKE 5−1 is a susceptibility factor to plant viruses. Proc. Natl. Acad. Sci. U. S. A. 111, 2104−2109. (15) Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680−685. (16) Holmgren, A. (1979) Thioredoxinn catalyze the reduction of insulin disulfides by dithiothreitol and dihydrolipoamide. J. Biol. Chem. 254, 9627−9632. (17) Martinez-Galisteo, E., Padilla, C. A., Garcìa-Alfonso, C., LopezBarea, J., and Bàrcena, J. A. (1993) Purification and properties of bovine thioredoxin system. Biochimie 75, 803−809.

AUTHOR INFORMATION

Corresponding Author

*Istituto di Biologia e Biotecnologia Agraria, Consiglio Nazionale delle Ricerche, Via Bassini 15a, 20133 Milano, Italy. Telephone: +39 02 23699 406. Fax: +39 02 23699 411. Email: [email protected]. ORCID

Francesco Bonomi: 0000-0003-4556-2640 Anna Paola Casazza: 0000-0001-6129-3722 Present Address ⊥

W.R.: Sacco S.r.l., Via A. Manzoni 29/A, 22071 Cadorago (CO), Italy. Funding

This work was supported by the Program “Strategie innovative e sostenibili per la filiera agroalimentare-FilAgro”, as part of the activities defined within the Accordo Quadro Consiglio Nazionale delle Ricerche and Regione Lombardia. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Floriana Gavazzi and Aldo Grasso (IBBA-CNR Milano) for technical assistance, Marco Albertini and Marilena Di Valentin (Università di Padova) for helping in the acquisition and discussion of the EPR spectra, Giuseppe Zucchelli (IBF-CNR Milano) for useful suggestions concerning I

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX

Article

Biochemistry (18) Ruddock, L. W., Hirst, T. R., and Freedman, R. B. (1996) pHdependence of the dithiol-oxidizing activity of DsbA (a periplasmic protein thiol:disulphide oxidoreductase) and protein disulphideisomerase: studies with a novel simple peptide substrate. Biochem. J. 315, 1001−1005. (19) Tsibris, J. C. M., and Woody, R. W. (1970) Structural studies of iron-sulfur proteins. Coord. Chem. Rev. 5, 417−458. (20) Orme-Johnson, W. H. (1973) Iron-Sulphur proteins: structure and function. Annu. Rev. Biochem. 42, 159−204. (21) Bonomi, F., and Pagani, S. (1986) Removal of ferritin-bound iron by DL-dihydrolipoate and DL-dihydrolipoamide. Eur. J. Biochem. 155, 295−300. (22) Darby, N. J., and Creighton, T. E. (1995) Functional Properties of the Individual Thioredoxin-like Domains of Protein Disulfide Isomerase. Biochemistry 34, 11725−11735. (23) Darby, N. J., Penka, E., and Vincentelli, R. (1998) The multidomain structure of protein disulfide isomerase is essential for high catalytic efficiency. J. Mol. Biol. 276, 239−247. (24) Nicolet, Y., and Drennan, C. L. (2004) AdoMet radical proteinsfrom structure to evolution-alignment of divergent protein sequences reveals strong secondary structure element conservation. Nucleic Acids Res. 32, 4015−4025. (25) Broderick, J. B., Duffus, B. R., Duschene, K. S., and Shepard, E. M. (2014) Radical S-adenosylmethionine Enzymes. Chem. Rev. 114, 4229−4317. (26) Krebs, C., Broderick, W. E., Henshaw, T. F., Broderick, J. B., and Huynh, B. H. (2002) Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of pyruvate formate-lyase activating enzyme: A Mossbauer spectroscopic study. J. Am. Chem. Soc. 124, 912−913. (27) Noodleman, L., and Baerends, E. J. (1984) Electronic Structure, Magnetic Properties, ESR, and Optical Spectra for 2-Fe Ferredoxin Models by LCAO-Xa Valence Bond Theory. J. Am. Chem. Soc. 106, 2316−2327. (28) Stephens, P. J., Thomson, A. J., Dunn, J. B. R., Keiderling, T. A., Rawlings, J., Rao, K. K., and Hall, D. O. (1978) Circular Dichroism and Magnetic Circular Dichroism of Iron-Sulfur Proteins. Biochemistry 17, 4770−4778. (29) Jung, Y. S., Gao-Sheridan, H. S., Christiansen, J., Dean, D. R., and Burgess, B. K. (1999) Purification and Biophysical Characterization of a New [2Fe-2S] Ferredoxin from Azotobacter vinelandii, a Putative [Fe-S] Cluster Assembly/Repair Protein. J. Biol. Chem. 274, 32402−32410. (30) Rouhier, N., Unno, H., Bandyopadhyay, S., Masip, L., Kim, S. K., Hirasawa, M., Gualberto, J. M., Lattard, V., Kusunoki, M., Knaff, D. B., Georgiou, G., Hase, T., Johnson, M. K., and Jacquot, J. P. (2007) Functional, structural, and spectroscopic characterization of a glutathione-ligated [2Fe-2S] cluster in poplar glutaredoxin C1. Proc. Natl. Acad. Sci. U. S. A. 104, 7379−7384. (31) Johansson, C., Kavanagh, K. L., Gileadi, O., and Oppermann, U. (2007) Reversible sequestration of active site cysteines in a 2Fe−2Sbridged dimer provides a mechanism for glutaredoxin 2 regulation in human mitochondria. J. Biol. Chem. 282, 3077−3082. (32) Vuillard, L., Braun-Breton, C., and Rabilloud, T. (1995) Nondetergent sulphobetaines: a new class of mild solubilization agents for protein purification. Biochem. J. 305, 337−343. (33) Hermann, W. M., Cutfield, J. F., and Hubbard, M. J. (2005) Biophysical Characterization of ERp29, EVIDENCE FOR A KEY STRUCTURAL ROLE OF CYSTEINE 125. J. Biol. Chem. 280, 13529−13537. (34) Kortemme, T., Darby, N. J., and Creighton, T. E. (1996) Electrostatic Interactions in the Active Site of the N-Terminal Thioredoxin-like Domain of Protein Disulfide Isomerase. Biochemistry 35, 14503−14511. (35) Holst, B., Tachibana, C., and Winther, J. R. (1997) Active Site Mutations in Yeast Protein Disulfide Isomerase Cause Dithiothreitol Sensitivity and a Reduced Rate of Protein Folding in the Endoplasmic Reticulum. J. Cell Biol. 138, 1229−1238.

(36) Huber-Wunderlich, M., and Glockshuber, R. (1998) A single dipeptide sequence modulates the redox properties of a whole enzyme family. Folding Des. 3, 161−171. (37) Selles, B. (2011) Les glutathion peroxydases et protéine disulfure isomérases de peuplier: potentialités du repliement thiorédoxine pour la catalyse des réactions redox. Ph.D. Thesis, Université Henri Poincaré (Nancy 1), Nancy, France. (38) Cheng, H., Xia, B., Reed, G. H., and Markley, J. L. (1994) Optical, EPR, and 1H NMR Spectroscopy of Serine-Ligated [2Fe-2S] Ferredoxins Produced by Site-Directed Mutagenesis of Cysteine Residues in Recombinant Anabaena 7120 Vegetative Ferredoxin. Biochemistry 33, 3155−3164. (39) Hurley, J. K., Weber-Main, A. M., Hodges, A. E., Stankovich, M. T., Benning, M. M., Holden, H. M., Cheng, H., Xia, B., Markley, J. L., Genzor, C., Gomez-Moreno, C., Hafezi, R., and Tollin, G. (1997) Iron-sulfur cluster cysteine-to-serine mutants of Anabaena [2Fe-2S] Ferredoxin exhibit unexpected redox properties and are competent in electron transfer to ferredoxin: NADP+ reductase. Biochemistry 36, 15109−15117. (40) Xia, B., Cheng, H., Bandarian, V., Reed, G. H., and Markley, J. L. (1996) Human ferredoxin: Overproduction in Escherichia coli, reconstitution in vitro, and spectroscopic studies of iron-sulfur cluster ligand cysteine-to- serine mutants. Biochemistry 35, 9488−9495. (41) Gulerez, I. E., Kozlov, G., Rosenauer, A., and Gehring, K. (2012) Structure of the third catalytic domain of the protein disulfide isomerase ERp46. Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Commun. 68, 378−381. (42) Banci, L., Brancaccio, D., Ciofi-Baffoni, S., Del Conte, R., Gadepalli, R., Mikolajczyk, M., Neri, S., Piccioli, M., and Winkelmann, J. (2014) [2Fe-2S] cluster transfer in iron-sulfur protein biogenesis. Proc. Natl. Acad. Sci. U. S. A. 111, 6203−6208. (43) Lill, R., and Mühlenhoff, U. (2008) Maturation of iron-sulfur proteins in eukaryotes: mechanisms, connected processes and diseases. Annu. Rev. Biochem. 77, 669−700. (44) Masip, L., Pan, J. L., Haldar, S., Penner-Hahn, J. E., DeLisa, M. P., Georgiou, G., Bardwell, J. C., and Collet, J. F. (2004) An Engineered Pathway for the Formation of Protein Disulfide Bonds. Science 303, 1185−1189. (45) Collet, J. F., Peisach, D., Bardwell, J. C., and Xu, Z. (2005) The crystal structure of TrxA(CACA): Insights into the formation of a [2Fe-2S] iron−sulfur cluster in an Escherichia coli thioredoxin mutant. Protein Sci. 14, 1863−1869. (46) Picciocchi, A., Saguez, C., Boussac, A., Cassier-Chauvat, C., and Chauvat, F. (2007) CGFS-Type Monothiol Glutaredoxins from the Cyanobacterium Synechocystis PCC6803 and Other Evolutionary Distant Model Organisms Possess a Glutathione-Ligated [2Fe-2S] Cluster. Biochemistry 46, 15018−15026. (47) Bisio, H., Bonilla, M., Manta, B., Graña, M., Salzman, V., Aguilar, P. S., Gladyshev, V. N., Comini, M. A., and Salinas, G. (2016) A New Class of Thioredoxin-Related Protein Able to Bind Iron-Sulfur Clusters. Antioxid. Redox Signaling 24, 205−216. (48) Couturier, J., Przybyla-Toscano, J., Roret, T., Didierjean, C., and Rouhier, N. (2015) The roles of glutaredoxins ligating Fe−S clusters: Sensing, transfer or repair functions? Biochim. Biophys. Acta, Mol. Cell Res. 1853, 1513−1527. (49) Couturier, J., Touraine, B., Briat, J. F., Gaymard, F., and Rouhier, N. (2013) The iron-sulfur cluster assembly machineries in plants: current knowledge and open questions. Front. Plant Sci. 4, 259. (50) Baerenfaller, K., Hirsch-Hoffmann, M., Svozil, J., Hull, R., Russenberger, D., Bischof, S., Lu, Q., Gruissem, W., and Baginsky, S. (2011) pep2pro: a new tool for comprehensive proteome data analysis to reveal information about organ-specific proteomes in Arabidopsis thaliana. Integr Biol. (Camb) 3, 225−237. (51) Porter, B. W., Yuen, C. Y. L., and Christopher, D. A. (2015) Dual protein trafficking to secretory and non-secretory cell compartments: Clear or double vision? Plant Sci. 234, 174−179.

J

DOI: 10.1021/acs.biochem.6b01257 Biochemistry XXXX, XXX, XXX−XXX