Introduction

Iron-sulfur clusters are essential protein cofactors that are evolutionarily conserved and usually coordinated by cysteines, although non-cysteinyl ligands, such as histidine, serine, and water, are occasionally used in place of a cysteine. They serve a vital role in diverse biological pathways, varying over electron transport, regulation of gene expression, substrate binding and activation, and radical generation [1]. Accordingly, iron-sulfur cluster biosynthetic pathways and the requisite protein machinery are generally conserved across all kingdoms of life [2]. The molecular and mechanistic basis of Fe–S cluster biosynthesis and cluster assembly in apo proteins has been the subject of intense investigation for more than two decades, stemming from the discovery that the reaction chemistry in cluster biogenesis is not spontaneous, but is, in fact, mediated by carrier proteins and chaperones [3–5]. The biogenesis of Fe–S clusters is complex, and, in general, can be accomplished on three main types of systems: namely, Isc (iron-sulfur cluster), Nif (nitrogen fixation), and Suf (sulfur utilization factor) [6]; with each system utilizing their own type of scaffold protein. In the case of eukaryotes, cellular cluster chemistry is also divided into mitochondrial iron-sulfur cluster (ISC) assembly and export machineries, and the cytosolic iron-sulfur protein (CIA) machinery. Details of how the mitochondrial and cytosolic Fe–S cluster assembly pathways is connected remain unclear, and have been the subject of extensive research with multiple proteins implicated, even if their roles are not entirely defined [7–10]. There is, however, no precedent for cellular chemistry involving a non-protein-bound Fe–S cluster, and so the potential involvement of an entirely glutathione-coordinated Fe–S cluster ([2Fe-2S](GS)4) in cluster assembly and delivery to target proteins, either directly or via pathways mediated by glutaredoxins and/or iron-sulfur cluster scaffold proteins, is of clear significance and interest. While small molecule iron-sulfur cluster compounds are typically stable only in non-nucleophilic solvents [11], we have demonstrated the [2Fe-2S] core to be stabilized in aqueous solution following coordination by four molecules of glutathione [12, 13]. A combination of optical, Mössbauer, mass spectrometry, and NMR characteristics is all consistent with a [2Fe-2S](GS)4 composition and strong antiferromagnetic coupling between the iron centers [13]. This novel cluster complex also appears to possess interesting and significant solution chemistry as a possible component of the labile iron pool, and as a substrate for the ABCB7 export protein that moves essential iron-sulfur clusters from the mitochondrion to the cytosol [14]. The glutathione-complexed cluster is capable of reconstituting the human iron-sulfur cluster scaffold protein IscU with a [2Fe-2S] cluster that can be further transferred to human ferredoxin [13]. Accordingly, such a cluster complex is readily exchangeable, could unify several sensory, regulatory, and transport pathways for cellular iron-sulfur clusters, and supports the presence of such clusters in the labile iron pool [15].

This report seeks to understand the functional and physiological roles of this novel [2Fe-2S](GS)4 complex in iron-sulfur cluster biosynthesis pathways, explores mechanisms of cluster exchange, and develops a possible role in the broader network of Fe–S cluster transfer reactions that underlie cluster trafficking.

Materials and methods

Materials

The gene for Saccharomyces cerevisiae S288c glutaredoxin 3 (Grx3) in pET28b(+) vector was ordered from GenScript. Primers (Forward: TAAGCACATATGCCGGTCATCGAA ATTAATGAT and Reverse: TGCTTAAAGCTTTTAGGATTGCAGTGCGTGTTG) were purchased from Integrated DNA Technologies, Inc. PD-10 desalting columns were purchased from GE Healthcare. Ferric chloride, sodium sulfide, and l-cysteine were obtained from Fisher.

Protein expression and purification

Purification of Homo sapiens IscU and Thermatoga maritima (Tm) NifS was performed as previously reported [16–18]. The expression vector for human ferredoxin-1 (Hs Fdx1) was kindly provided by Dr. J. Markley, and protein was expressed and purified according to the literature procedures [19]. Purification for human ferredoxin-2 (Hs Fdx2) was performed as previously reported [20]. The ferredoxins were isolated and purified as holo proteins and subsequently converted to apo forms by treatment with 100-mM EDTA (pH 7.5), 5-mM DTT, and 8-M urea in 50-mM Hepes, 100-mM NaCl, and pH 7.5. A construct of human Grx2 (comprising residues 56–161) with a tobacco etch virus cleavable N-terminal His6 tag in expression vector pNic-Bsa4 was kindly provided by Drs. Kavanagh, Muller-Knapp, and Oppermann, and protein was expressed and purified as previously reported [21]. Schizosaccharomyces pombe Isa1 protein was expressed and purified as previously reported [22]. The yeast Grx3 (Q03835, residues 36-285) open reading frame encoding both the Trx-like and Grx-like domains, but lacking the first 35 amino acids (Δ1–35), since those residues were not found to be utilized in vivo [23], was cloned into pET28b(+) between HindIII and NdeI restriction sites and expressed in Escherichia coli BL21 (DE3).

Transformed cells were grown in 10 mL of Luria–Bertani (LB) broth media-containing kanamycin (50 μM) at 37 °C overnight and then transferred to 3 L of fresh LB media-containing kanamycin and incubated at 37 °C until the OD600 reached 0.8. Protein expression was induced by addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) (300 μM) and the cell culture incubated at 30 °C for 10 h. Cell pellets were collected by centrifugation at 4330 g for 15 min, suspended in 30 mL of buffer-containing 50-mM Hepes, 100-mM NaCl, pH 7.5, and lysed by use of a dismembranator. The lysate was centrifuged at 28,928 g for 30 min, the supernatant subsequently applied to a Ni–NTA column, and the protein eluted with buffer-containing 50-mM Hepes, 100-mM NaCl, 0.3-M imidazole, and pH 7.5, prior to analysis on a 12 % SDS-PAGE gel that was then visualized by Coomassie Blue staining. Dialysis was carried out at 4 °C against 50-mM Hepes, 100-mM NaCl, and pH 7.5. Protein concentration was determined by use of the Bradford assay.

[2Fe-2S](GS)4 synthesis

The cluster was synthesized as previously reported [13]. Briefly, stock solutions of ferric chloride and sodium sulfide were added to 10-mL 40-mM glutathione (GSH) solution and pH 8.6 to final concentrations of 20 mM. A volume (40 mL) of ethanol was added to the mixture and mixed well by vortexing. The precipitate was collected by centrifugation at 13,000 rpm for 10 min, washed twice with ethanol, and dried under vacuum.

Iron quantitation [24, 25]

A solution of [2Fe-2S](GS)4 (0.05 mM, 200 μL) in H2O or holo protein was acidified by concentrated HCl (60 μL) and heated to 100 °C for 15 min. The resulting suspension was centrifuged at 14,000 rpm for 2 min, to remove precipitated protein, while the iron remained in solution. The supernatant (100 μL) was diluted with Tris–HCl (0.5 M, 1.3 mL, pH 8.5). Solutions of sodium ascorbate (0.1 mL, 5 %) and bathophenanthroline disulfonate (0.4 mL, 0.1 %) were sequentially added to the neutralized reaction solution with mixing between each addition. The solution was incubated at 25 °C for 1 h and iron was quantitated by measuring the absorbance at 535 nm on a UV–Vis spectrophotometer and calculated from a calibration curve made with 0.01–0.2-mM FeCl3 standard solutions. Subsequent analysis of the pellet revealed no bound iron.

In vitro reconstitution of apo proteins

In vitro reconstitution of Grx2 and Grx3 was performed as previously reported [21]. A solution of 200-μM apo protein was incubated with 2-μM Tm Nifs, 3-mM GSH, and 5-mM DTT, argon-purged for 30 min, and then made up to 0.6 mM in Fe2+ and 0.6 mM in l-cysteine. After incubation for 2 h, the reaction mixture was concentrated and passed through a PD10 desalting column to remove excess Fe2+, S2−, and l-cysteine, and the UV–Vis spectrum was obtained for both apo and holo Grx2 and Grx3.

For IscU and Isa1 [24], 200-μM apo protein was incubated with 50-mM DTT and 8-M urea, argon-purged for 30 min, and then made up to 1 mM in Fe2+ and 1 mM in S2−. After incubation for 1 h, the reaction mixture was concentrated and passed through a PD-10 desalting column which was equilibrated with 50 mM Hepes, 100 mM NaCl, and pH 7.5 to remove excess salts, urea, Fe2+, and S2−, and the UV–Vis spectrum was obtained for both apo and holo IscU and Isa1.

Cluster transfer experiment by electronic spectrophotometry

UV–Vis measurements were obtained on a Varian Cary 50 UV–Vis spectrophotometer. The absorbance at 330 nm and 415 nm corresponds to cluster charge transfer bands that can be used to monitor cluster formation, cluster transfer from the [2Fe-2S](GS)4 complex to apo proteins, and cluster extraction from holo proteins by glutathione to form [2Fe-2S](GS)4 complex. Glutathione has previously been shown to extract the iron-sulfur cluster core from holo protein to form the [2Fe-2S](GS)4 complex by monitoring the change in the charge transfer bands at 330 nm and 420 nm on UV–Vis [13]. The extinction coefficient for the [2Fe-2S](GS)4 complex is 7600 M−1 cm−1 at 330 nm [13]. At 420 nm, the extinction coefficient of the complex is 4900 M−1 cm−1. The extinction coefficient for the cluster of all the holo proteins in this report lies in the range of 8000 to 11,000 M−1 cm−1 based on published data [26, 27] and also matches the observed trend in absorbance. Hence, a decrease in the absorbance at those wavelengths can be attributed to cluster extraction by glutathione from holo proteins. Degassed and reconstituted holo protein in 50-mM Hepes, 100-mM NaCl, and pH 7.5 was incubated with GSH from 80 μM to 3 mM in an anaerobic cuvette and the absorbance measured from 700 to 200 nm on a Cary Win UV spectrophotometer every 2 min over the course of 1 h. The change in absorbance at 420 nm or 330 nm was plotted against time and fit to an exponential to obtain kobs. The second-order rate constants for GSH extraction were determined by use of observed rate constants (k obs) and the concentrations of glutathione and the [2Fe-2S] cluster. Cluster extraction by GSH to form the [2Fe-2S](GS)4 complex was confirmed by use of ESI mass spectrometry on a Bruker Micro-TOF (ESI) spectrometer and data were analyzed by use of the Data Analysis software (Bruker) [13, 15].

Cluster transfer experiment by circular dichroism

CD spectra were recorded by use of a Jasco J815 spectropolarimeter. The time course for cluster transfer from [2Fe-2S](GS)4 to apo proteins was monitored under anaerobic conditions at 25 °C by use of UV–visible CD spectroscopy in small volume 1-cm quartz cuvettes. Reactions were carried out in 50-mM Hepes, 100-mM NaCl buffer, and pH 7.5 with 200-μM DTT. Solid [2Fe-2S](GS)4 was resuspended in argon-purged 50-mM Hepes, 100-mM NaCl buffer, and pH 7.5, and added to the argon-purged anaerobic cuvette by use of a gas-tight syringe to a final concentration of 100 μM to initiate the reaction. CD scans from 300 to 600 nm were collected to analyze signature cluster-bound protein peaks at a scan rate of 200 nm/min. Data were processed by use of JASCO Spectramanager II Analysis software and analyzed using the chemical kinetics and equilibria program DynaFit by BioKin [28]. The deconvolution function from Spectramanager II analysis software was used for analysis of bands in the spectra that contained overlapping Lorentzian curves having the same full width at half maximum value that accurately distinguishes the peak positions for each band.

A second-order rate constant for cluster uptake was obtained via a method similar to the iron-sulfur cluster transfer method of Johnson and colleagues [27, 29]. Protein and the [2Fe-2S](GS)4 complex were mixed in a 1:1 ratio, 100 μM of each, and the peak at 330 nm for IscU, 370 nm for Isa1, 460 nm for Grx3, 365 nm for Grx2, 445 nm for Fdx1, and 440 nm for Fdx2 was monitored by CD spectroscopy. The data were converted to percent cluster transfer and analyzed using the DynaFit program [28] to determine the second-order rate constants for the various reactions by best-fit simulation to the second-order kinetics. The signal for the holo protein at the end of the reaction was compared with the reconstituted holo protein signal for the same concentration of [2Fe-2S] cluster concentration to determine the percentage of cluster transfer.

Results

Cluster transfer from [2Fe-2S](GS)4 complex to human IscU

Glutathione is an important cellular redox agent and is a cofactor for glutathione peroxidase [16, 30–32]. In addition, it has been shown that glutathione depletion impairs the maturation of cytosolic Fe–S cluster proteins, but has no effect on mitochondrial Fe–S cluster proteins, implying a significant role for glutathione in mediating mitochondrial cluster export in vivo [33]. Accordingly, glutathione is not only an important cellular redox agent or an exogenous ligand for glutaredoxins, but may also be intimately involved in Fe–S cluster transport by creating a labile pool of iron-sulfur clusters via the [2Fe-2S](GS)4 complex.

The sensitivity and intensity of the UV–Vis CD spectra of iron-sulfur centers make this technique the method of choice for monitoring the kinetics of cluster uptake or cluster transfer and reflect the chirality of the cluster environment in holo proteins. The [2Fe-2S](GS)4 complex is CD silent in the UV–Vis region, lacking any signature CD profile from 300 to 600 nm [13]. The CD profiles for the cluster center for each of the proteins in this work (Hs IscU, S. pombe Isa1, Hs Grx2, S. cerevisiae Grx3, Hs Fdx1 and Hs Fdx2) are distinct and can be distinguished from each other (Figure S1), thereby providing a convenient and quantitative measurement of cluster transfer between these proteins by measuring changes in ∆ε values at specific wavelengths.

A physiologically stable glutathione cluster complex [2Fe-2S](GS)4, has previously been shown to undergo reversible cluster exchange with the human iron-sulfur cluster scaffold protein IscU by use of UV–Vis spectroscopy [13]. Our previous studies with IscU suggest a role for the [2Fe-2S](GS)4 complex downstream of de novo iron-sulfur cluster biosynthesis on IscU, in which the complex can function to facilitate transfer to other protein targets or create a labile pool of [2Fe-2S] cluster [13]. Successful cluster transfer from [2Fe-2S](GS)4 complex to apo human IscU can be seen from the UV–Vis spectrum following completion of the cluster uptake reaction (Fig. 1). This spectrum matches that of the holo IscU spectrum obtained following the traditional chemical reconstitution by use of Fe3+ and S2−. UV–Vis CD spectroscopy was used to quantitatively investigate potential cluster transfer reactions and to evaluate the kinetics of cluster transfer from the [2Fe-2S](GS)4 complex to apo human IscU. The time course for cluster transfer from holo [2Fe-2S](GS)4 complex to apo IscU was monitored at room temperature by use of a 1:1 donor-to-acceptor cluster stoichiometry; and by continuously monitoring the time-dependent reaction at 330 nm, as shown in Fig. 1, with quantitative transfer occurring over the course of 30 min as determined by the change in CD ∆ε values for holo IscU. The rate constant was determined by use of chemical kinetics program DynaFit, and a lower limit for the apparent second-order rate constant for cluster uptake was determined to be 4100 ± 1500 M−1 min−1. While not in the preferred pseudo-first-order range, as a result of limitations from detection sensitivity and protein solubility, this is the method typically used to provide estimates for second-order rate constants in cluster transfer reaction [27, 29, 34], and as we show later in a test case is similar to the actual rate constant determined from more traditional concentration dependence measurements. This measured rate of cluster uptake is in the physiologically relevant range and is significantly faster than the previously reported A. vinelandii IscU reconstitution (9 M−1 min−1) [34], or even transfer via the HscA/HscB/ATP mediated pathway from [2Fe-2S] cluster-bound IscU to apo Isc Fdx (800 M−1 min−1) [34], both of which were determined by use of a similar CD method under conditions comparable with those described herein. Control experiments involving [2Fe-2S](GS)4 complex and no apo IscU did not show cluster breakdown over a period of an hour under anaerobic conditions as monitored by UV–Vis spectroscopy (Figure S2), supporting a model where cluster is transferred as an intact species from the [2Fe-2S](GS)4 complex to apo IscU and not via cluster degradation and reassembly on the target IscU protein. In fact, the use of solution inorganic iron and sulfide with glutathione, at the same total concentrations used for the cluster complex, did not yield any reconstitution of the apo proteins described here. This glutathione-complexed iron-sulfur cluster can, therefore, provide a mechanism for iron-sulfur cluster storage in vivo that is both stable and available at the cellular level for iron-sulfur cluster proteins [13, 15].

Fig. 1
figure 1

a UV spectra for human IscU after cluster uptake from [2Fe-2S](GS)4. b CD spectra for IscU after cluster uptake from [2Fe-2S](GS)4. c Time course measurement of cluster transfer from [2Fe-2S](GS)4 to apo human IscU in the absence of excess glutathione monitored by UV–visible CD spectroscopy under anaerobic conditions in semi-micro 1-cm cuvettes at room temperature. The change in extinction values was based on the initial [2Fe-2S]2+ cluster concentration. d The data were obtained by monitoring the change in CD intensity at 330 nm for a 1:1 reaction mixture and the solid-red line is the best-fit simulation by DynaFit to the second-order kinetics with a rate constant 4100 M−1min−1

Full size image

Cluster transfer from [2Fe-2S](GS)4 complex to S. pombe Isa1

In addition to IscU, two Isa homologues, Isa1 and Isa2, represent another set of scaffold proteins that have been identified in the yeast mitochondrial ISC assembly machinery; however, their molecular function is unknown. Isa has been implicated either as an alternative scaffold protein for Fe–S cluster assembly and transfer [35, 36], or as an iron-delivery protein [3, 37]. In this report, we have investigated if the [2Fe-2S](GS)4 complex was capable of successfully promoting cluster transfer to and reconstitution of S. pombe Isa1.

Potential cluster transfer reactivity, and the kinetics of cluster transfer from the [2Fe-2S](GS)4 to apo Isa1, have been determined by use of UV–Vis CD spectroscopy. Successful cluster transfer from [2Fe-2S](GS)4 complex to apo Isa1 can be seen from the UV–Vis spectrum after completion of the cluster uptake reaction (Fig. 2). This spectrum matches that obtained for holo Isa1 following the traditional chemical reconstitution by use of Fe3+ and S2− ions. The time course of cluster transfer from holo [2Fe-2S](GS)4 complex to apo Isa1 was monitored at room temperature by use of a 1:1 donor:acceptor cluster stoichiometry and is shown in Fig. 2. Transfer occurs over a time course of 30 min, yielding a lower limit for the apparent second-order rate constant for cluster uptake of 7400 ± 1500 M−1 min−1. These kinetic data were obtained by continuously monitoring cluster transfer every 10 s for a total of 30 min at 370 nm.

Fig. 2
figure 2

a UV spectra for S. pombe Isa1 after cluster uptake from [2Fe-2S](GS)4. b CD spectra for Isa1 after cluster uptake from [2Fe-2S](GS)4. c Time course measurement of cluster transfer from [2Fe-2S](GS)4 to apo Isa1 in the absence of excess glutathione monitored by UV–visible CD spectroscopy under anaerobic conditions in semi-micro 1-cm cuvettes at room temperature. The change in extinction values was based on the initial [2Fe-2S]2+ cluster concentration. d The data were obtained by monitoring the change in CD intensity at 370 nm for a 1:1 reaction mixture and the solid-red line is the best-fit simulation by DynaFit to the second-order kinetics with a rate constant 7400 M−1min−1

Full size image

Cluster transfer from [2Fe-2S](GS)4 complex to monothiol (Grx3) and dithiol glutaredoxins (Grx2)

Glutaredoxins are widely distributed among all organisms, and play significant roles in Fe–S cluster trafficking within the cell, Fe–S protein biogenesis, and in regulatory and signaling mechanisms for Fe–S target proteins [38–40]. Accordingly, we have investigated cluster uptake by two types of glutaredoxins, namely, the monothiol S. cerevisiae Grx3 and the dithiol human Grx2. Human Grx2 (hGrx2) is a dithiol glutaredoxin primarily found in the mitochondria and has been shown to serve a central role in cellular apoptosis and mitochondrial redox regulation [41, 42]. By contrast, eukaryotic Grx3 (yeast, yGrx3) is a multidomain monothiol glutaredoxin that is localized in the nucleus and/or cytosol [43]. Both, Grx2 and Grx3 bind a [2Fe-2S] cluster, which is bridged between two monomers and involves two glutathione molecules as exogenous ligands [21, 44].

Interestingly, cluster uptake was observed only for Grx3 (Fig. 3) and not for Grx2 (data not shown). The same technique that was used to monitor the previously described transfer reactions for IscU and Isa1 was used to evaluate potential cluster transfer reactions and kinetics of cluster transfer, from [2Fe-2S](GS)4 to apo Grx3. Successful cluster transfer from the [2Fe-2S](GS)4 complex to apo Grx3 is seen from the UV–Vis spectrum after completion of cluster uptake (Fig. 3). This spectrum is similar to the spectrum obtained for holo Grx3 following the traditional chemical reconstitution by use of Tm NifS, Fe3+, and l-cysteine, with an absorbance band ~ 420 nm and a 330-nm band appearing as a shoulder (under the concentration conditions used). UV–Vis CD spectroscopy was used to both evaluate potential cluster transfer reactions and determine the kinetics of transfer from the [2Fe-2S](GS)4 complex to apo human Grx3. The time course for cluster transfer from holo [2Fe-2S](GS)4 complex to apo Grx3 was monitored at room temperature (Fig. 3) by use of a 1:1 donor:acceptor stoichiometry and following a previously reported method [27, 29, 34].

Fig. 3
figure 3

a UV spectra for S. cerevisiae Grx3 after cluster uptake from [2Fe-2S](GS)4 (20 µM). b CD spectra obtained for Grx3 after cluster uptake from [2Fe-2S](GS)4 are similar to reported data [29]. c Time course measurement of cluster transfer from [2Fe-2S](GS)4 to apo Grx3 in the absence of excess glutathione monitored by UV–visible CD spectroscopy under anaerobic conditions in semi-micro 1-cm cuvettes at room temperature. The change in extinction values was based on the initial [2Fe-2S]2+ cluster concentration. d The data were obtained by monitoring the change in CD intensity at 460 nm for a 1:1 reaction mixture and the solid-red line is the best-fit simulation by DynaFit to the second-order kinetics with a rate constant 1360 M−1min−1

Full size image

Quantitative transfer occurred over the course of 30 min (data fit to first 10 min), as reflected by the change in CD ∆ε values. The second-order rate constant for cluster uptake was determined to be 1360 ± 110 M−1 min−1, by continuously monitoring cluster transfer every 10 s for a total of 30 min over the course of the reaction at 460 nm.

The kinetics of cluster uptake by glutaredoxins from the glutathione-bound [2Fe-2S](GS)4 cluster is complex, because two of the ligands in holo glutaredoxins are glutathiones. To further understand the mechanism of cluster uptake by glutaredoxins, cluster uptake reactions for Grx3 were evaluated at different concentrations of cluster in the absence of excess glutathione. The second-order rate constant obtained from this experiment was 1600 ± 450 M−1 min−1 (Fig. 4a), which closely matches the second-order rate constant obtained from the mixed 1:1 donor:acceptor reaction. The effect of protein concentration dependence on the rate of cluster uptake by Grx3 in the absence of excess glutathione was also determined. The [2Fe-2S](GS)4 cluster was kept in excess and held constant to maintain pseudo-first-order conditions. The second-order rate constant determined from the protein dependence for cluster uptake was 2153 ± 281 M−1 min−1 (Fig. 4b). The plot of k obs for cluster uptake versus total apo protein concentration resulted in an inverse hyberbola (Fig. 4b). This can be fit to a substrate inhibition model, in which the addition of more apo Grx3 promotes the formation of Grx3 dimers, causing the monomer concentration to decrease. However, the monomeric form is the active species capable of taking up the cluster. This trend is reflected in the decay in the uptake rate with increasing protein concentration. As shown in Eq. (1), M corresponds to the monomeric protein and M 2 is the dimeric form, which exists in equilibrium. The reaction mechanism can be shown by Eq. (2), which then yields Eq. (3) [45, 46],

Fig. 4
figure 4

a Second-order rate constant determination for cluster transfer reaction from [2Fe-2S](GS)4 to apo Grx3 in absence of excess glutathione by UV–vis CD spectroscopy. The observed rate constant (kobs) of cluster uptake by Grx3 at different [2Fe-2S](GS)4 concentrations was plotted against cluster concentration to obtain the second-order rate constant, which was 1600 ± 450 M−1 min−1. b Protein concentration dependence on the rate of cluster uptake by Grx3 in the absence of excess glutathione. The [2Fe-2S](GS)4 cluster was kept in excess and held constant to maintain pseudo-first-order conditions. The concentration of Grx3 was varied and the data were fit to Eq. (3), reflecting inhibition by the dimeric form of the protein. The second-order rate constant determined from the protein dependence for cluster uptake was 2153 ± 281 M−1min−1

Full size image
$$ M_{2} \overset {K_{D} } \rightleftharpoons 2M $$
(1)
$$ {M} + [2{\text{Fe}}{\hbox{-}}2{\text{S}}]({\text{GS}})_{4} \mathop\rightarrow^{{k}_{1}}\limits_{-2 {\rm GSH} }{M}- [2{\text{Fe}}{\hbox{-}}2{\text{S}}]({\text{GS}})_{2} \mathop\rightarrow^{{M}}\limits_{-2 {\rm GSH} }{M}_{2} - [2{\text{Fe}}{\hbox{-}} 2{\text{S}}] $$
(2)
$$ k_{\text{obs}} = {\raise0.7ex\hbox{${k_{1} \left[ P \right]}$} \!\mathord{\left/ {\vphantom {{k_{1} \left[ P \right]} {K_{\text{D}} + \left[ P \right]\left( {1 + \frac{\left[ P \right]}{{K_{\text{D}} }}} \right)}}}\right.\kern-0pt} \!\lower0.7ex\hbox{${K_{\text{D}} + \left[ P \right]\left( {1 + \frac{\left[ P \right]}{{K_{\text{D}} }}} \right)}$}} $$
(3)

where k 1 is a first-order rate constant (min−1), P is the total protein concentration, and K D is an apparent dissociation constant for the monomer–dimer equilibrium given by the following equation:

$$ K_{\text{D}} = \frac{{\left[ M \right]^{2} }}{{\left[ {M_{2} } \right]}} $$
(4)

The constants k 1 and K D were calculated directly from Eq. (3) by use of a non-linear least-squares fit in the Origin software. At the lower protein concentrations in the plot of k obs versus protein concentration (Fig. 4b), the data fit well to the model with almost a linear trend, suggesting a greater dependence on protein concentration in this region. These data were used with Eq. 3 to obtain the second-order rate constant for the formation of the holo protein, ~2153 ± 281 M−1 min−1, which is comparable to the rate constant obtained by varying the cluster concentration. The first-order rate constant (k 1) was determined to be 1.69 min−1 and the apparent dissociation constant (K D) was 40 µM.

The influence of glutathione on cluster transfer kinetics was also investigated by performing the same cluster uptake reactions from the [2Fe-2S](GS)4 complex in the presence of increasing concentrations of glutathione. Previously, the K D for glutathione binding to glutaredoxins has been estimated by use of NMR titration experiments to lie in the mM range, and it has been shown that glutathione can bind to both, apo and holo glutaredoxin. Holo Grx3 has been shown to bind a [2Fe-2S] cluster in a dimeric form [44], whereas the apo protein can potentially exist in a monomer–dimer equilibrium, with or without excess glutathione. However, GSH binding to the apo form inhibits dimer formation when GSH is present above 20 mM [47, 48]. The binding of glutathione to apo Grx3 indicates that GSH is not only an exogenous ligand to the [2Fe-2S] cluster in glutaredoxins, but is also involved in making additional contacts with the protein even in the absence of the cluster.

Accordingly, in the absence of excess glutathione, the experimental data support a model, where the [2Fe-2S] cluster core from the [2Fe-2S](GS)4 complex is the subject of nucleophilic attack by the cysteine in the cluster-binding site of the apo Grx3 monomer, thereby releasing 1 GSH molecule from the complex and forming an intermediate monomeric holo species of Grx3. This is the rate-determining step followed by a very fast step involving binding of another Grx3 molecule to form the stable dimeric holo Grx3 (Scheme 1a). The intermediate monomeric holo Grx3 species is CD silent, and as it is followed by a very fast step, the reaction appears as a single-step process (Fig. 3d).

Scheme 1
scheme 1

Model for cluster uptake from [2Fe-2S](GS)4 complex by apo Grx3 in (A) the absence of additional glutathione, and in (B) the presence of excess glutathione. In (A), the apo protein exists in equilibrium between monomeric and dimeric states. As the total protein concentration increases, the concentration of dimer increases, which reflects the decrease in reactivity shown in Fig. 4B, where the dimeric form is unable to take up the [2Fe-2S](GS)4 complex to form the holo Grx3. Since there is no excess GSH, the GSH ligands required for the holo Grx3 form are provided by the GSH ligands on the complex. In (B), excess GSH is present in the solution. As such, the first step in Scheme 1B depends on GSH concentration, because it must bind to Grx3 prior to reaction. At higher concentrations (above the KD ~ 50 μM for GSH binding to Grx3), the GSH site is saturated and so the rate is maximized

Full size image

By contrast, in the presence of excess glutathione, a two-step pathway is followed (Scheme 1b). First, the apo monomeric Grx3 binds to one GSH per monomer. Similar to the reaction in the absence of additional glutathione (Scheme 1a), the [2Fe-2S](GS)4 complex is subject to nucleophilic attack by the Cys in the cluster-binding site of Grx3 monomer, followed by rapid binding of a second Grx3 to form an intermediate cluster-bridged dimer (Fig. 5a, step 1). This complex has chirality associated with it that is similar to, but exceeds that of native holo Grx3, and can be distinguished by CD spectroscopy (Fig. 5). Subsequently, in step 2, cluster-bound glutathiones are then released by intramolecular displacement by the Grx3-bound glutathiones with the formation of native holo Grx3.

Fig. 5
figure 5

a Time course measurement of cluster transfer from [Fe2S2](GS)4 complex to apo Grx3 monitored by UV–visible CD spectroscopy under anaerobic conditions in semi-micro 1-cm cuvettes at room temperature in the presence of 5-mM glutathione (GSH). Biphasic kinetics was observed which can be divided into two distinct steps. b Second-order rate constant determination for cluster transfer reaction from GSH cluster complex to apo Grx3 under varying GSH concentration from 0 mM to 10 mM for the two distinct steps was determined

Full size image

The [2Fe-2S] cluster transfer reactions from the [2Fe-2S](GS)4 complex to IscU and Isa1 were also performed in the presence of excess glutathione, however, as expected, neither of the cluster uptake reactions showed any glutathione dependence (data not shown). The [2Fe-2S] cluster transfer reactions from the [2Fe-2S](GS)4 complex to human Fdx’s were also performed, but neither Fdx1 nor Fdx2 accepted cluster from [2Fe-2S](GS)4 complex under our conditions.

Extraction of [2Fe-2S]2+ cluster from holo proteins by glutathione

Earlier work in our laboratory has demonstrated the glutathione-stabilized [2Fe-2S]2+ core to be exchangeable with iron-sulfur cluster scaffold protein IscU, in which the [2Fe-2S](GS)4 complex can be formed by cluster extraction from holo IscU by free glutathione [13]. Accordingly, under physiological concentrations of mitochondrial glutathione, this provides a possible pathway by which [2Fe-2S](GS)4 clusters are synthesized in vivo. We, therefore, investigated other proteins that could readily exchange cluster with free glutathione to form [2Fe-2S](GS)4 complex.

When holo human IscU was incubated with free glutathione, the absorbance at 330 nm and 420 nm was observed to decrease until it reached a plateau (Fig. 6). The change in absorbance was consistent with the difference in extinction coefficient for holo IscU and glutathione-coordinated cluster (as reported previously [13]). This was true for Isa1 and Grx3 as well (Fig. 6). For both, holo Isa1 and holo Grx3, the extinction coefficients for the cluster are higher compared with that of the [2Fe-2S](GS)4 complex (7600 M−1 cm−1 at 330 nm), which is reflected by the absorbance decrease in the region of the charge transfer bands (330 and 420 nm). A significantly smaller change in absorbance was noted in the absence of GSH, reflecting the hydrolytic instability of the protein-bound cluster. Cluster extraction is supported by the time-dependent change in the UV–Vis spectrum, where the concentration of cluster transferred can be calculated by use of the Beer–Lambert relationship, because the extinction coefficients for the [2Fe-2S](GS)4 complex [13] and holo proteins are known at 300 nm and 420 nm. From the absorbance change, the estimated [2Fe-2S](GS)4 complex formed at the end of 1 h is 5.1 µM for IscU, 15.7 µM for Grx3, and 7.6 µM for Isa1 from the initial concentration of [2Fe-2S] cluster of 20 µM of holo protein. This is consistent with the half-time determined for the reaction. Following the glutathione extraction reaction by the UV–Vis analysis, the presence of the [2Fe-2S](GS)4 complex was confirmed by use of ESI mass spectrometry. Major peaks were observed at 1413.3 and 1435.3 (Figures S6, S7) for Isa and Grx3, respectively, corresponding to intermediate species previously identified in mass spectra of [2Fe-2S](GS)4 complex [15], and also similar to reaction intermediates observed in the cluster assembly product for the scaffold proteins (IscS–IscU) [49]. For IscU, glutathione extraction of cluster results in the formation of the [2Fe-2S](GS)4 complex with a bound sodium atom [13]. This MS peak was also observed for Grx3 and Isa1 at 1427.3 (Figures S6, S7), although at a lower intensity.

Fig. 6
figure 6

For Grx3, Isa1, and IscU, the absorbance at 330 and 420 nm is observed to decrease, as shown as an example with Grx3 in (a), when 20-µM holo Grx3 (b), 20-µM holo Isa1 (c), and 20-µM holo IscU (d) are incubated with 1-mM GSH in 50 mM Hepes, 100-mM NaCl, pH 7.5. Red, holo protein in the absence of GSH; black, holo protein with 1-mM GSH added

Full size image

For glutathione extraction, different concentrations of glutathione were used (0–3 mM) for 10 µM holo protein. Figure 6 shows typical traces for cluster extraction by glutathione from holo [2Fe-2S] proteins in the presence of excess glutathione followed either at 330 nm or 420 nm. The data were fit to an exponential curve-fit and pseudo-first-order rate constants, k obs, for [2Fe-2S](GS)4 complex formation were obtained. The plot of k obs versus glutathione concentration is a straight line for [2Fe-2S](GS)4 complex formation from holo IscU (Fig. 7c). The slope from this fit resulted in a second-order rate constant of 130 ± 27 M−1 min−1. However, the plot of kobs versus glutathione concentration for holo Grx3 and holo Isa1 is a rectangular hyberbola (Figs. 7a, b, respectively). The plot levels off at high glutathione concentrations for these two reactions. This behavior can be explained by a simple pre-reaction-binding interaction between reactants according to the following equation:

Fig. 7
figure 7

Pseudo-first-order rate constants (k obs) were determined for [2Fe-2S] cluster extraction by glutathione from holo Grx3, Isa1, and IscU (Fig. 6.). The k obs determined from the exponential decrease in absorbance was then fit to Eq. 6 to obtain the second-order rate constants for [2Fe-2S](GS)4 formation from (a) holo Grx3, (b) holo Isa1, and from (c) holo IscU

Full size image
(5)

which yields the following equation:

$$ k_{\text{obs}} = \frac{{k_{1} }}{{1 + \frac{K}{{\left[ {\text{GSH}} \right]}}}} $$
(6)

where k 1 is a first-order rate constant (min−1) for [2Fe-2S](GS)4 complex formation and K (M) is an apparent dissociation constant for the pre-reaction complex of holo protein and GSH given by the following equation:

$$ K = \frac{{\left[ {\text{holo-protein}} \right]\left[ {\text{GSH}} \right]}}{{\left[ {\text{holo-protein-GSH}} \right]}}. $$
(7)

The constants k 1 and K were calculated directly from Eq. (6) by use of a non-linear least-squares fit by use of the Origin software. At lower glutathione concentrations, the plot of k obs versus glutathione concentration is linear and was used to obtain the second-order rate constant for the formation of [2Fe-2S](GS)4 complex from Grx3 and Isa1. Table 1 summarizes the second-order rate constants for [2Fe-2S] cluster uptake by apo proteins from the [2Fe-2S](GS)4 complex and formation of [2Fe-2S](GS)4 complex from holo proteins by free glutathione.

Table 1 Second-order rate constants for cluster uptake from [2Fe-2S](GS) 2−4 complex and cluster extraction by free glutathione from holo proteins
Full size table

Discussion

In this report, we have further established the [2Fe-2S](GS)4 complex to be stable under physiological conditions and demonstrated reversible cluster exchange with a range of Fe–S cluster proteins. Comparison of the second-order rate constants for cluster uptake by IscU, Isa1, and Grx3, reveals Isa1 to be most likely a preferred substrate for the [2Fe-2S](GS)4 complex, with subsequent transfer reactions to other chaperones or target proteins. Although the proteins examined in this work were from a variety of organisms, we can still conclude that the complex is capable of reacting with a variety of iron-sulfur cluster scaffold and delivery proteins. Investigations of other iron-sulfur cluster proteins that can serve as potential substrates for the [2Fe-2S](GS)4 complex are ongoing. There is an apparent preference for proteins with some element of tertiary structure. Unstructured proteins, such as the apo ferredoxins, that lack structure and are highly flexible in the apo state, were unable to interact with the cluster complex [13]. An additional measure of selectivity must be at play, given the difference in the determined rate constants and the inability of Grx2 to interact with the complex (Scheme 2). The exchange of cluster between any of Isa1, IscU, Grx3, and free glutathione further suggests a cellular buffering role for exchangeable [2Fe-2S]2+ cluster, with glutathione extraction preventing hydrolytic degradation, but allowing return of the cluster to apo proteins, such as IscU, Grx3, and Nfu, when protein-promoted delivery of cluster to target proteins is required. If a target protein is not readily available, then extrusion as a stable glutathione adduct preserves the cluster following its formation [16, 24], and we have successfully demonstrated the feasibility of direct reconstitution of several important apo target proteins by the [2Fe-2S](GS)4 complex.

Scheme 2
scheme 2

Schematic proposal representing cluster uptake and transfer pathways in the cell. Kinetic rate constants for iron-sulfur cluster uptake from the [2Fe-2S](GS)4 complex and for transfer from holo proteins to free glutathione. The [2Fe-2S](GS)4 complex is believed to be exported from the mitochondria via the ABCB7 transporter (51). Following export to the cytosol, the cluster can then be taken directly up by the cytosolic proteins

Full size image

Cluster uptake by IscU and Isa1 appears to follow a relatively simple pathway, with direct docking to the cluster-binding pocket, presumably mediated by Cys exchange of a cluster-bound glutathione, and subsequent rapid intramolecular ligand exchange to yield the final holo product. In the case of Grx3, in the absence of exogenous glutathione (Scheme 1a), cluster binding also involves direct substitution by Cys from Grx3 to form a monomeric intermediate, where a glutathione can also dock to the surface glutathione-binding site on the Grx3 protein, with subsequent recruitment of the second Grx3. The formation of a monomeric intermediate is consistent with the kinetic rate studies, and by the dependence on protein concentration, while the rapid uptake of the second Grx3 is most likely favored by additional bonding contacts to the protein that provides a template effect.

In the presence of exogenous glutathione, the kinetic data for formation of holo dimeric Grx3 are consistent with a model, where glutathione binds to apo monomeric Grx3 to form a glutathione-bound protein complex that can then recruit cluster by displacement of two cluster-bound glutathiones by active site cysteine, and cysteine from bound glutathione, to form a monomeric intermediate that rapidly recruits a second Grx3 to form an intermediate [2Fe-2S]-bridged species (Scheme 1b). Binding of the second Grx3 is most likely again favored by additional bonding contacts to the protein (similar to the proposed mechanism in the absence of exogenous glutathione, Scheme 1a). Under conditions of excess GSH, the rds is dependent upon GSH being bound to apo Grx3, in contrast to the pathway shown in Scheme 1a. Therefore, GSH binds to a Grx3 monomer before interacting with the cluster. This binding event depends on the concentration of GSH. With high GSH concentration, the binding sites on Grx3 will be occupied and greatly accelerate the rate of cluster uptake, in comparison to the scenario with low GSH and a slower rate of uptake, due to the lack of additional contacts for the cluster.

The subsequent exchange of the cluster-bound glutathiones from the [2Fe-2S](GS)4 complex with the bound glutathiones on the monomeric Grx3 protein is an intramolecular event that is independent of the protein concentration. This is supported by our experimental data, in which there is no protein dependence on cluster uptake in the presence of excess glutathione (7 mM), indicating that the monomeric Grx3 is the reactive species for cluster uptake. Interestingly, the first phase (Fig. 5a, Step 1) is relatively independent of external glutathione, and is slightly inhibitory up to ~2 mM, most likely reflecting a preference for the absence of Grx3-bound glutathione prior to the initial uptake of cluster, where the cluster complex can provide the Grx3 additional glutathione. Prior binding of glutathione then becomes inhibitory. At higher concentrations of glutathione, both steps 1 and 2 (Fig. 5a) show an enhancement of activity with increasing glutathione concentration, consistent with pre-population of the glutathione-binding sites on Grx3-promoting formation of holo dimer. In addition, the intermediate monomeric holo Grx3 species observed in cluster uptake reaction Scheme 1(both a and b) can then serve as a potential substrate for formation of heterodimeric holo complexes, such as Grx3-Grx4, Grx3-Aft1, and Grx3-Fra2, all of which have been shown to be more stable compared with homodimeric holo Grx3 [43, 50].

Formation of [2Fe-2S](GS)4 complex by extraction of cluster from holo proteins by free glutathione is mechanistically different for the various holo proteins investigated. In the case of human IscU, there is no evidence of a pre-reaction complex [Eq. (5)], whereas, formation of a glutathione-bound holo protein complex was observed for both yeast Grx3 and Isa1, as seen by the saturation kinetics of this process (Fig. 7). Grx3 makes contacts with glutathione in both apo and holo forms. In the holo form, glutathione serves as an exogenous ligand for the cluster; however, additional GSH molecules could bind to the cluster prior to extraction by GSH to form the [2Fe-2S](GS)4 complex. In case of Isa1, the holo protein has been shown to be both monomeric as well as dimeric [22, 36, 51]. Free glutathione can make contacts with the dimeric protein forming an intermediate glutathione-bound holo protein complex. In the case of monomeric holo S. pombe Isa1, free glutathione most likely coordinates to the cluster directly at the non-cysteinyl ligand site, promoting the formation of pre-reaction complex as observed by the saturation kinetics (Fig. 7B) for facile cluster extraction. Structural studies on IscU have shown that the IscU protein exists in two distinct states: structured (S) and disordered (D) [52], where each structural state exhibits different behavior toward construction of an iron-sulfur cluster on IscU [53]. Each state has a different affinity for the cluster and this could potentially affect the extraction of the cluster by free glutathione. Holo IscU is primarily monomeric [54, 55] with the [2Fe-2S] cluster bound on the surface, which may prevent formation of stronger interactions with free glutathione and thereby prevent the formation of a pre-reaction complex.

Given the reversibility of transfer from the [2Fe-2S](GS)4 complex to apo proteins and then formation of the [2Fe-2S](GS)4 complex via GSH extraction from holo proteins, the question of the directionality of these reactions in vivo naturally arises. Rate constants for cluster extraction by GSH are significantly smaller than rate constants for cluster delivery, and so there is certainly a driving force for delivery, notwithstanding the high cellular concentrations of glutathione. Furthermore, a large excess of GSH has been utilized in the previous work examining cluster transfer reactions, and a direct protein–protein interaction is the apparent preferred route, since a clear one-step reaction is observed [29, 56]. Ultimately, the actual observed rates and the position of equilibria will depend on relative concentrations of species, which are not known; however, we propose that the glutathione-stabilized cluster complex can still serve as a vital storage component for cellular iron-sulfur clusters, especially in light of the potential for fluctuating cellular conditions (oxidative stress, iron levels, cellular compartments, etc.).

The role of DTT was also investigated, since recent studies have suggested formation of DTT-bound higher molecular weight protein species that could be involved in cluster transfer or change in nuclearity when excess of DTT (20 mM) is added to reaction mixtures. It is unclear if the effects of high concentrations of DTT reflect facilitated cluster transfer between putative protein partners [57, 58], or through structural destabilization induced by high DTT concentrations. In our own studies, we incubated holo proteins (holo IscU, holo Isa1, and holo Grx3) with excess DTT (Figures S3 to S5). No significant change in CD signal was observed for the holo proteins, clearly indicating that under our conditions, DTT is not forming any intermediate Fe–S species (supplementary information). In the case of holo Grx3, there was minimal change in the CD signal (<10 %) when incubated with excess DTT (Figure S5). In addition, the change is inconsistent over time indicating non-specific interaction of DTT with the holo protein. To further demonstrate that uptake of the [2Fe-2S](GS)4 complex was not dependent on the presence of DTT, we were able to utilize a non-thiol reducing agent, TCEP, as a replacement. These data are consistent with the view that both DTT and TCEP simply serve to reduce any disulfide bonds formed in the cysteine rich domains of the iron-sulfur cluster-binding proteins rather than acting as a vehicle to mediate cluster transfer.

Glutathione-complexed [2Fe-2S] cluster is a plausible component of the labile iron pool and is able to reconstitute apo protein to provide an alternative mechanism for reconstitution of holo derivatives (Scheme. 2). We have demonstrated interactions with both mitochondrial and cytosolic proteins, suggesting a role in both cellular compartments. Following de novo cluster biogenesis on IscU in the mitochondria, free GSH can abstract the cluster to form the [2Fe-2S](GS)4 complex and act as either a storage of iron-sulfur clusters or an alternative vehicle for cluster delivery to downstream proteins, such as Isa1, to promote cluster trafficking. This complex can then be a potential substrate for mitochondrial exporter protein ABCB7 and can serve similar roles in the cytosol for iron-sulfur storage and transfer. With these implications for the cellular role of [2Fe-2S](GS)4 complex, this complex is now amenable to a molecular understanding to fully probe its functional role as a part of the labile iron pool.

Ongoing research in our laboratory seeks to understand potential functional and physiological roles for such [2Fe-2S](GS)4 complexes in both iron-sulfur cluster biosynthesis and cellular trafficking of clusters in molecular and mechanistic detail.