Effector role reversal during evolution: the case of frataxin in Fe-S cluster biosynthesis

Abstract

Human frataxin (FXN) has been intensively studied since the discovery that the FXN gene is associated with the neurodegenerative disease Friedreich’s ataxia. Human FXN is a component of the NFS1-ISD11-ISCU2-FXN (SDUF) core Fe-S assembly complex and activates the cysteine desulfurase and Fe-S cluster biosynthesis reactions. In contrast, the Escherichia coli FXN homolog CyaY inhibits Fe-S cluster biosynthesis. To resolve this discrepancy, enzyme kinetic experiments were performed for the human and E. coli systems in which analogous cysteine desulfurase, Fe-S assembly scaffold, and frataxin components were interchanged. Surprisingly, our results reveal that activation or inhibition by the frataxin homolog is determined by which cysteine desulfurase is present and not by the identity of the frataxin homolog. These data are consistent with a model in which the frataxin-less Fe-S assembly complex exists as a mixture of functional and nonfunctional states, which are stabilized by binding of frataxin homologs. Intriguingly, this appears to be an unusual example in which modifications to an enzyme during evolution inverts or reverses the mode of control imparted by a regulatory molecule.

Iron-sulfur (Fe-S) clusters are protein cofactors that are essential for most living organisms. Fe-S clusters primarily function in electron transfer, but also participate in substrate binding and activation, protein stabilization, regulation of gene expression or enzyme activity, radical generation, and as a sulfur donor (1–3). These biological functions require the Fe-S cluster to vary from solvent accessible to completely buried, and to be accommodated in a diverse set of protein scaffolds (4). Despite this diversity, conserved metallochaperone proteins recognize and insert Fe-S clusters into apo Fe-S proteins through mechanisms that remain poorly understood. This assembly system must also control the reactivity of the building blocks for Fe-S clusters, ferrous iron and sulfide, to reduce toxicity and cell death. In humans, ferrous iron is prone to undergo Fenton chemistry: reaction with hydrogen peroxide to generate highly reactive hydroxyl radicals that damage proteins, lipids, and DNA, whereas sulfide is more toxic than cyanide, is a potent inhibitor of cytochrome c oxidase, and targets olfactory nerves, the eyes, and the brain (5). Defects in these metallochaperone systems have profound effects on iron metabolism, and result in mitochondrial dysfunction and human disease (6–8).

In prokaryotes, three different Fe-S assembly systems have been characterized: the nitrogen fixation (NIF), iron-sulfur cluster assembly (ISC), and mobilization of sulfur (SUF) pathways (9–13). Escherichia coli contains both ISC and SUF systems and appears to use the ISC pathway under ‘normal’ growth conditions and the SUF system under conditions of iron limitation or oxidative stress (14). The ISC operon is controlled by IscR, a Fe-S cluster binding transcriptional repressor, and contains genes for a cysteine desulfurase (IscS), scaffold protein (IscU), ferredoxin (Fdx), IscA, molecular chaperones (hscA and hscB), and IscX (also known as yfhj and ORF3). The scaffold protein IscU (in this manuscript the E. coli homolog is referred to as IscU_ec, or U_ec) is the central protein in Fe-S cluster biosynthesis and is responsible for the assembly of [2Fe-2S] and [4Fe-4S] clusters (15). IscS (E. coli homolog referred to as IscS_ec or S_ec) catalyzes the PLP-dependent breakdown of cysteine to alanine and provides the inorganic sulfur for Fe-S clusters (16, 17). Fdx likely provides electrons for cluster assembly and IscA may be an iron donor or alternate scaffold (18–20). The function of IscX is unknown, but it appears to compete with CyaY for binding to IscS (21–23). Once the cluster is formed, the chaperones hscA and hscB facilitate the transfer of intact Fe-S clusters to apo target proteins (24–26). A homolog of human frataxin (E. coli homolog referred to as CyaY or as C_ec in protein complexes in this manuscript) is not part of the ISC operon, but has been shown to bind to the IscS_ec-IscU_ec (S_ecU_ec) complex (21, 27, 28) and inhibit Fe-S cluster assembly (29).

A similar ISC Fe-S cluster assembly pathway functions in the matrix space of eukaryotic mitochondria (8, 30). The Fe-S assembly proteins are encoded by nuclear DNA and often contain a targeting sequence that is cleaved upon import into mitochondria. For the human pathway, the cysteine desulfurase NFS1 (59% identical to IscS_ec), which forms a eukaryotic-specific functional complex with ISD11 (31–33), the scaffold protein ISCU2 (also known as Isu2; 70% identical to IscU_ec), and FXN (20% identical to CyaY) form a NFS1-ISD11-ISCU2-FXN (SDUF) complex (34, 35) that is analogous to the IscS_ec-IscU_ec-CyaY (S_ecU_ecC_ec) complex. Recent in vitro studies indicate that FXN is an allosteric activator of Fe-S cluster assembly that increases the catalytic efficiency (k_cat/K_M) for the cysteine desulfurase and the rate of Fe-S cluster biosynthesis (34). Intriguingly, in vitro data indicate E. coli CyaY and human FXN have opposing regulator functions in Fe-S cluster biosynthesis (29, 34). Different physiological functions may also be supported by different in vivo phenotypes for E. coli and mouse model systems lacking frataxin (36, 37). However, both CyaY and FXN can at least partially complement strains of yeast lacking the frataxin homolog Yfh1 (38, 39), which implies similar functions.

To resolve this discrepancy and better understand the role of frataxin in Fe-S cluster biosynthesis, enzyme kinetic experiments were performed for the human and E. coli systems in which analogous components were interchanged. Surprisingly, our results reveal that the cysteine desulfurase component, IscS_ec/NFS1-ISD11, rather than CyaY/FXN or IscU_ec/ISCU2 dictates the functional differences in how the frataxin homologs modulate Fe-S cluster assembly.

Experimental Procedures

Preparation of human recombinant proteins

Human NFS1 (Δ1–55) and ISD11 were coexpressed in Escherichia coli strain BL21(DE3) as previously described (34, 40). Cells were lysed by French press in 50 mM Tris pH 8.0, 500 mM NaCl, and 5 mM imidiazole. The NFS1-ISD11 (SD) complex was purified as previously described (34) except that a cation exchange column was added between the Ni-NTA and S300 gel filtration chromatography steps. The yellow SD fractions from the Ni-NTA column were combined with 100 μM pyridoxal 5′-phosphate (PLP), 5 mM dithiothreitol (DTT), and 2 mM ethylenediaminetetracetic acid (EDTA), diluted 3-fold with 50 mM Tris pH 8.0, loaded onto a cation exchange column (16/14, POROS 50HS), and eluted with a linear gradient from 0–1000 mM NaCl. Human ISCU2 (Δ1–35) was purified as previously described (34) except the fractions for the cation exchange column were eluted with a linear gradient from 0 – 1000 mM NaCl in 50 mM Tris, pH 7.5 buffer and a S100 rather than a S300 gel filtration column was used. Human FXN (Δ1–55) was purified as previously described (34) except a S100 rather than a S300 column was used. The Bradford method (41) or extinction coefficients of 42670, 8250, and 26030 M⁻¹cm⁻¹ at 280 nm (42) were used to estimate the protein concentration of SD, ISCU2, and FXN, respectively.

Preparation of E. coli recombinant proteins

A plasmid containing the IscS_ec gene (29) was transformed into BL21(DE3) E. coli cells and grown at 37 °C. Protein expression was induced with 0.5 mM IPTG when the OD₆₀₀ was ~0.6. Cells were harvested 5 hours later and lysed by sonication in 50 mM Tris pH 8.0, 500 mM NaCl, 5 mM imidiazole, and 100 μM PLP. After centrifugation, the supernatant was loaded on a Ni-NTA column (16/13, GBiosciences) and eluted with a linear gradient from 5 to 250 mM imidazole. Yellow fractions were concentrated and further purified on a Sephacryl S300 column (26/60, GE Healthcare) equilibrated in 50 mM HEPES pH 8.0, 250 mM NaCl. Plasmids that contained either IscU_ec or CyaY genes (29) were transformed into BL21(DE3) cells and grown at 37 °C. Protein expression was induced with 0.5 mM IPTG when the OD₆₀₀ was ~0.6. Cells were harvested 3.5 hours later and lysed by sonication in 50 mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA, and 1 mM DTT. The supernatant was loaded onto a GST column and eluted with a linear gradient from 0–7 mM glutathione. The fractions that bound the column were concentrated and dialyzed into 50 mM Tris pH 8.0, 0.5 mM EDTA, and 1 mM DTT and then incubated at room temperature overnight with TEV protease. The tag cleavage product was loaded onto the GST column and the flow through was further purified on a S100 gel filtration column equilibrated in 50 mM HEPES pH 7.5, 150 mM NaCl. The Bradford method (41) or extinction coefficients of 41370, 11460, and 28990 M⁻¹cm⁻¹ at 280 nm (42) were used to estimate the protein concentration of IscS_ec, IscU_ec, and CyaY, respectively.

Titrations to favor protein complex formation

Cysteine desulfurase assays were used to determine reaction stoichiometries to favor protein complexes for different combinations of human and E. coli proteins. Cysteine desulfurase assays (see below for details) were initiated with 100 μM L-cysteine and included 0.5 μM of the cysteine desulfurase (SD or IscS_ec) and increasing amounts of the scaffold protein (ISCU2 or IscU_ec). Once the saturating number of equivalents of the scaffold protein were determined (Table 1), additional titrations were performed with increasing amounts of frataxin (FXN or CyaY).

Table 1.

Protein complexes and reaction stoichiometries.

Name	Components	Protein ratios
SDU	human NFS1/ISD11 and ISCU2	1:3
SDUec	human NFS1/ISD11; E. coli IscUec	1:3
SDUF	human NFS1/ISD11, ISCU2, and FXN	1:3:3
SDUecF	human NFS1/ISD11; E. coli IscUec; human FXN	1:3:200
SDUCec	human NFS1/ISD11 and ISCU2; E. coli CyaY	1:3:80
SDUecCec	human NFS1/ISD11; E. coli IscUec and CyaY	1:3:30
SecUec	E. coli IscS and IscUec	1:5
SecU	E. coli IscS; human ISCU2	1:5
SecUecCec	E. coli IscS, IscUec, and CyaY	1:5:10
SecUCec	E. coli IscS; human ISCU2; E. coli CyaY	1:5:10
SecUecF	E. coli IscS and IscUec; human FXN	1:5:15
SecUF	E. coli IscS; human ISCU2 and FXN	1:5:15

Cysteine desulfurase activity measurements

Assay mixtures (800 μL) contained 0.5 μM cysteine desulfurase (SD or IscS_ec), 10 μM PLP, 2 mM DTT, 5 μM Fe(NH₄)₂(SO₄)₂, and increasing or saturating amounts (Table 1) of the scaffold protein (ISCU2 or IscU_ec) and frataxin (FXN or CyaY) in 50 mM Tris pH 8.0, and 250 mM NaCl (34, 40, 43). Samples were incubated in an anaerobic glovebox (10~14 °C) for 30 minutes. The cysteine desulfurase reactions were initiated by addition of L-cysteine (0.01 to 1 mM) at 37 °C and quenched after 10 minutes by the addition of 100 μl of 20 mM N,N-dimethyl-p-phenylenediamine in 7.2 N HCl and 100 μl of 30 mM FeCl₃ in 1.2 N HCl, which also initiated the conversion of sulfide to methylene blue. After a twenty minute incubation at 37 °C, the absorption at 670 nm due to methylene blue formation was measured and compared with a Na₂S standard curve to quantitate sulfide production. Cysteine desulfurase reactions were also performed for mixed species Fe-S assembly complexes with 0 to 80 equivalents of Fe(NH₄)₂(SO₄)₂ and 100 μM L-cysteine. Units are defined as μmol sulfide/μmol cysteine desulfurase per minute at 37 °C. Rates were fit to the Michaelis-Menten equation using KaleidaGraph (Synergy Software).

Fe-S Cluster Formation

Assay mixtures contained 8 μM of the cysteine desulfurase (SD or IscS_ec), 5 mM DTT, 200 μM Fe(NH₄)₂(SO₄)₂, 100 μM L-cysteine, 50 mM Tris pH 8.0, and 250 mM NaCl in a total volume of 0.2 mL. In addition, the number of equivalents (relative to the cysteine desulfurase) for the scaffold protein (ISCU2 or IscU_ec) and frataxin (FXN or CyaY) required to saturate the activity (above; Table 1) were included in the assay mixture. The scaffold protein was incubated with 5 mM DTT in 50 mM Tris pH 8, 250 mM NaCl in an anaerobic glovebox for 30 minutes prior to mixing with the remaining assay components in an anaerobic cuvette and then incubated at 10 °C for 30 minutes in the cuvette holder. The reaction was initiated by injecting L-cysteine to a final concentration of 100 μM with a gas-tight syringe. Fe-S cluster formation was monitored at 456 nm at 10 °C and then the first 3000 seconds were fit as first-order kinetics using KaleidaGraph. The rate was converted to the activity using an extinction coefficient of 5.8 mM⁻¹cm⁻¹ at 456 nm for [2Fe-2S] cluster absorbance (15). Units are defined as the μmol of [2Fe-2S] cluster / μmol cysteine desulfurase per minute at 10 °C.

RESULTS

Ratio of protein complex components for activity measurements

The objectives of this study were to provide insight into functional differences in frataxin-mediated regulation of Fe-S cluster biosynthesis for prokaryotes and eukaryotes by measuring cysteine desulfurase and Fe-S formation activities of mixed species protein complexes. The sulfur generated by the cysteine desulfurase reaction can be directly measured or can be used as a component of the Fe-S formation reaction. Titrations were performed to determine the relative ratios of the different components for maximum activity. Previous experiments revealed that adding 3 equiv of ISCU2 minimized the cysteine desulfurase activity of the NFS1-ISD11 (SD) complex (Table 1), likely through formation of a NFS1-ISD11-ISCU2 (SDU) species (34). Likewise, adding 3 equiv of FXN to SDU maximized the cysteine desulfurase activity by forming a NFS1-ISD11-ISCU2-FXN (SDUF) complex (34). A similar strategy was used here to determine and then compensate for weaker binding of E. coli proteins to the human assembly complex and human proteins to the E. coli assembly complex.

First, protein reaction ratios were determined for the human SD complexes by measuring cysteine desulfurase activities with increasing amounts of the scaffold protein, either ISCU2 or IscU_ec, and frataxin, either FXN or CyaY. The cysteine desulfurase activity of SD decreased with addition of either ISCU2 or IscU_ec and minimized after the addition of ~2 equiv/SD (Figure 1A). Three equiv of ISCU2 or IscU_ec were therefore included with SD for kinetic experiments of the SDU or NFS1-ISD11-IscU_ec (SDU_ec) complexes (Table 1). The binding of frataxin homologs to the SDU complex increased the cysteine desulfurase activity that maximized after the addition of either 3 equiv of FXN or 80 equiv of CyaY (Figure 1B). Similarly, the binding of frataxin homologs to the SDU_ec complex increased the cysteine desulfurase activity that maximized after either 200 equiv of FXN or 30 equiv of CyaY (Figure 1C).

Figure 1.

Determination of protein ratios that favor complex formation. Cysteine desulfurase activities were measured for (A) the human SD complex with increasing amounts of ISCU2 or IscU_ec, and for (B) the SDU and (C) SDU_ec complexes with increasing amounts of FXN or CyaY. (D) The cysteine desulfurase activity was measured for E. coli IscS with increasing amounts of ISCU2 or IscU_ec.

Second, protein reaction ratios were determined for IscS_ec using similar titrations. The cysteine desulfurase activity of IscS_ec slightly decreased with addition of IscU_ec and slightly increased with addition of ISCU2 (Figure 1D). This perturbation in IscS_ec activity maximized after the addition of 3–5 equiv of IscU_ec or ISCU2. To favor complex formation, five equiv of IscU_ec or ISCU2 were therefore added to IscS_ec (Table 1) for all kinetic experiments of IscS_ec-IscU_ec (S_ecU_ec) and IscS_ec-ISCU2 (S_ecU) complexes, respectively. Titrations of up to 50 equiv of CyaY or FXN did not significantly change the cysteine desulfurase activity of either the S_ecU_ec or S_ecU complex (data not shown). Therefore, protein complex stoichiometries of 1:5:10 were used for the S_ecU_ecC_ec and S_ecUC_ec complexes based on the number of equivalents of CyaY previously determined to inhibit Fe-S cluster formation (29). Similarly, protein complex stoichiometries of 1:5:15 were arbitrarily chosen for the Michaelis-Menten kinetics (below) of the S_ecU_ecF and S_ecUF complexes (Table 1).

IscU_ec and CyaY stimulate the cysteine desulfurase activity of the human SD complex

Saturating amounts of ISCU2 or IscU_ec and FXN or CyaY determined above were added to the SD complex and the rates of the cysteine desulfurase reaction were measured as a function of the L-cysteine concentration. The human SD complex exhibited a k_cat/K_M of 660 M⁻¹s⁻¹ for the cysteine desulfurase reaction (Table 2), which is slightly higher (primarily due to a lower K_M for L-cysteine) than previously reported values of ~100 M⁻¹s⁻¹ (34, 40). The addition of ISCU2 or IscU_ec and formation of the SDU or SDU_ec complex slightly decreased the k_cat (Figure 2 and Table 2). The binding of FXN to form the SDUF and SDU_ecF complexes stimulated the k_cat for the cysteine desulfurase reaction by ~7-fold relative to the SDU and SDU_ec complexes (Figure 2). Similarly, CyaY binding and formation of the SDUC_ec and SDU_ecC_ec complexes resulted in a 4–6 fold increase in the k_cat relative to SDU and SDU_ec (Figure 2). These experiments show that IscU_ec can functionally substitute for ISCU2 and that CyaY can replace FXN as an activator of the cysteine desulfurase reaction for the human Fe-S assembly complex.

Table 2.

Kinetic data for Fe-S assembly complexes.

Name	kcat (min−1)	KM (mM)	kcat/KM (M−1sec−1)	Fe-S cluster assembly activity (min−1)
SD	1.5 ± 0.1	0.038 ± 0.012	660 ± 200
SDU	1.0 ± 0.1	0.028 ± 0.008	600 ± 180	0.26 ± 0.22
SDUF	6.7 ± 0.4	0.011 ± 0.004	9700 ± 3900
SDUF + Fe	15.4 ± 1.1	0.015 ± 0.006	16800 ± 6300	30.39 ± 1.18
SDUCec	5.7 ± 0.4	0.027 ± 0.01	3500 ± 1300
SDUCec + Fe	6.7 ± 0.2	0.016 ± 0.003	7100 ± 1400	3.94 ± 0.48
SDUec	0.59 ± 0.02	0.018 ± 0.003	550 ± 120	1.30 ± 0.32
SDUecF	4.2 ± 0.1	0.009 ± 0.002	7900 ± 2000
SDUecF + Fe	6.1 ± 0.3	0.012 ± 0.003	8300 ± 2300	25.54 ± 1.17
SDUecCec	2.1 ± 0.1	0.012 ± 0.003	2900 ± 840
SDUecCec + Fe	1.9 ± 0.1	0.011 ± 0.002	2800 ± 480	1.21 ± 0.19
Sec	7.5 ± 0.1	0.017 ± 0.002	7300 ± 860
SecUec	4.7 ± 0.3	0.019 ± 0.006	4100 ± 1400	12.23 ± 0.59
SecUecCec	5.8 ± 0.2	0.025 ± 0.003	3800 ± 440
SecUecCec + Fe	4.2 ± 0.1	0.020 ± 0.004	3500 ± 680	0.79 ± 1.76
SecUecF	5.2 ± 0.1	0.017 ± 0.002	5100 ± 870
SecUecF + Fe	4.5 ± 0.2	0.024 ± 0.004	3200 ± 940	4.56 ± 0.40
SecU	6.5 ± 0.2	0.019 ± 0.003	5800 ± 1000	14.63 ± 0.95
SecUCec	7.4 ± 0.1	0.016 ± 0.002	7600 ± 900
SecUCec + Fe	7.3 ± 0.3	0.017 ± 0.004	7100 ± 1600	0.88 ± 0.64
SecUF	7.6 ± 0.3	0.020 ± 0.006	6300 ± 1600
SecUF + Fe	6.9 ± 0.1	0.016 ± 0.002	7200 ± 890	3.62 ± 0.24

Figure 2.

Stimulation of cysteine desulfurase reaction by addition of frataxin homologs. The k_cat for the cysteine desulfurase reaction was plotted for complexes of human SD with the scaffold protein (ISCU2 or IscU_ec) and frataxin (FXN or CyaY).

Iron-dependent stimulation of cysteine desulfurase reaction for SD complexes

Cysteine desulfurase activities were determined for SD complexes as a function of added iron. From 0 to 80 equiv of Fe(II) were added to the SDUF, SDUC_ec, SDU_ecF, and SDU_ecC_ec complexes and the cysteine desulfurase activity was measured (Figure 3A). The activity maximized after about 1, 1, and 5 equiv for the SDUF, SDUC_ec, and SDU_ecF complexes, respectively. In contrast, the SDU_ecC_ec complex did not exhibit a significant Fe-based activation. In addition, the SD complexes were incubated with 10 equiv of ferrous iron and the Michaelis-Menten parameters for the cysteine desulfurase activity were determined (Table 2). Ferrous iron had a large effect on the k_cat of the SDUF complex, as previously reported (34), a moderate effect on the SDUC_ec and SDU_ecF complexes, and no effect on the SDU_ecC_ec complex (Figure 3B).

Figure 3.

Iron-dependent perturbation of cysteine desulfurase activity for SD complexes. (A) Cysteine desulfurase activity plotted as a function of added ferrous iron. (B) Plot of k_cat values for the cysteine desulfurase reaction of SD complexes with (black) and without (gray) 10 equiv of ferrous iron.

CyaY can substitute for FXN in Fe-S cluster synthesis by human SD complexes

The rate of Fe-S cluster assembly for the SDU, SDU_ec, SDUF, SDUC_ec, SDU_ecF, and SDU_ecC_ec complexes were determined by monitoring the increase in absorbance at 456 nm (15) at 10 °C and converting this change in absorbance to an activity measurement (μmol [2Fe-2S] / μmol of complex per minute). In the absence of FXN or CyaY, the SDU and SDU_ec complexes exhibited very low or no activity and were similar to a control reaction in which sulfide was substituted for cysteine (Table 2 and Figure 4). The addition of FXN greatly stimulated the Fe-S assembly activity of SDU and SDU_ec. In contrast, addition of CyaY resulted in a very modest or no increase in Fe-S assembly activity of the SDU and SDU_ec complexes, respectively. A CyaY-based stimulation was more readily observed when the Fe-S cluster assembly assays were repeated at room temperature (Figure 4A inset). Thus, E. coli CyaY can substitute for human FXN in the human Fe-S assembly complex and strongly stimulate the cysteine desulfurase activity. CyaY also weakly stimulates the Fe-S cluster formation reaction.

Figure 4.

Stimulation of Fe-S cluster assembly activity by frataxin homologs. (A) The Fe-S assembly activities were measured at 10 °C for the Na₂S control and SDU, SDUC_ec, and SDUF complexes. The inset shows the activities of the SDU, SDUC_ec, and Na₂S control at 25 °C. (B) The Fe-S assembly activities were measured at 10 °C for the Na₂S control and SDU_ec, SDU_ecC_ec, and SDU_ecF complexes.

CyaY and FXN have minor effects on the cysteine desulfurase activity of IscS_ec complexes

Purified recombinant IscS_ec exhibited a k_cat of 7.5 min⁻¹ (Figure 5A and Table 2), which is similar to the previously reported value of 8.5 min⁻¹ (44). Addition of IscU_ec diminished the k_cat for the cysteine desulfurase reaction by a factor of 1.6, whereas a more moderate decrease in k_cat was observed upon ISCU2 addition (Figure 5A). Previously, IscU_ec was shown to either have no effect on the activity of IscS_ec (45) or stimulate the activity 6-fold (46) using different experimental conditions and assays. The k_cat for S_ecU_ec and S_ecU were increased slightly upon addition of either CyaY or FXN (Figure 5A and Table 2). The IscS_ec cysteine desulfurase activity was slightly decreased by added ferrous iron, consistent with a previous report (47), that appeared to occur within the first equiv of added iron for the S_ecU_ecF and S_ecU_ecC_ec complexes and 10 equiv of iron for the S_ecUF complex (Figure 5B). Steady state kinetics also revealed the k_cat slightly decreased upon addition of 10 equiv of ferrous iron (Figure 5A). The slight decrease in cysteine desulfurase activity upon addition of iron (Figure 5) is opposite of the general Fe-based activation observed for SD complexes (Figure 3). Overall, these results indicate that binding of frataxin homologs and ferrous iron have minor effects on the cysteine desulfurase activity of IscS complexes.

Figure 5.

Cysteine desulfurase activities for IscS complexes with and without iron. (A) The k_cat for the cysteine desulfurase was plotted for complexes of E. coli IscS (S_ec) without (gray or colored) and with (black) 10 equiv of ferrous iron. (B) The cysteine desulfurase activities for IscS complexes were measured with 100 μM L-cysteine and increasing amounts of ferrous iron.

CyaY and FXN inhibit Fe-S assembly by IscS_ec complexes

The Fe-S cluster assembly rate for the S_ecU_ec and S_ecU complexes were similar to each other (Figure 6) and significantly higher than analogous SDU_ec and SDU complexes with NFS1-ISD11 (Figure 4). The addition of 10 equiv of CyaY to the S_ecU_ec and S_ecU complexes completely inhibited Fe-S cluster formation (Figure 6A and 6B), consistent with previous results (29). At 2 equiv, CyaY is a slightly more effective inhibitor of the native S_ecU_ec complex than the mixed species S_ecU complex. The addition of FXN to the S_ecU_ec and S_ecU complexes (Figure 6C and 6D) also lowered the rate and possibly the amount of cluster bound; however, FXN did not completely eliminate Fe-S cluster formation. Thus, bacterial IscS assembly complexes are inhibited in Fe-S cluster formation by the inclusion of either frataxin ortholog.

Figure 6.

Frataxin-based inhibition of Fe-S cluster formation for IscS complexes. The rate of Fe-S cluster formation were measured for the S_ecU_ec complex with increasing amounts of (A) CyaY and (C) FXN, and for the S_ecU complex with increasing amounts of (B) CyaY and (D) FXN.

DISCUSSION

The physiological function of frataxin has been intensively investigated since it was associated with the neurodegenerative disease Friedreich’s ataxia (FRDA; reviewed in (48, 49)). A decrease in FXN levels in yeast and mouse model systems results in a phenotype that includes loss of Fe-S cluster enzyme activity, accumulation of iron in mitochondria, and susceptibility to oxidative stress (50–52). In addition, FXN levels are correlated with the age of disease onset for FRDA patients (53, 54), and FXN missense mutations were identified that show FRDA-like phenotypes in mouse (52) and yeast (55) model systems. A role for frataxin as an Fe chaperone and/or donor in Fe-S cluster biosynthesis is often cited due to the frataxin depletion phenotype (50–52), the ability of frataxin to bind Fe (56–63), and the acceleration of Fe-S cluster formation by Fe-bound FXN in vitro (58). Recent in vitro data as well as data presented here reveals that human FXN binds to the SDU complex and activates both the cysteine desulfurase and Fe-S formation reactions (34). Moreover, assays using FXN missense mutations exhibit defects in binding and activation that appear to correlate with the severity of clinical progression for FRDA patients (64, 65). The physiological function of CyaY is less clear. In contrast to FXN in eukaryotes, deletion or overexpression of CyaY does not affect cellular growth, iron content, and survival after exposure to H₂O₂ (36, 66). Part of the complication for these types of studies is the presence of both the ISC and SUF Fe-S assembly systems in E. coli. If the physiological role of CyaY were to inhibit Fe-S cluster formation by the ISC system, then one would expect overexpression of CyaY in the absence of the SUF system to give a loss of Fe-S cluster phenotype. To the best of our knowledge this experiment has not been reported. Unfortunately, the complicated phenotype for human FXN depletion, the lack of a phenotype for CyaY depletion, and the opposing in vitro functions in Fe-S cluster assembly have contributed to the general confusion over the physiological role of this protein.

Here we used a function replacement approach to understand the factors that mediate the dramatically different in vitro responses to frataxin homolog binding. More specifically, in vitro cysteine desulfurase and Fe-S cluster assembly activities were evaluated for protein complexes in which components from the recombinant human Fe-S assembly system: the NFS1-ISD11 cysteine desulfurase complex, the scaffold protein ISCU2, and FXN, and the recombinant E. coli system composed of the IscS cysteine desulfurase, the scaffold protein IscU_ec, and CyaY, were interchanged. Surprisingly, our results reveal that activation or inhibition by the frataxin homolog is not controlled by the frataxin homolog but rather dictated by the cysteine desulfurase.

Interestingly, even though IscS_ec and NFS1 are almost 60% identical, IscS_ec exhibited a k_cat/K_M for the cysteine desulfurase activity that is 10–70 fold higher than that of the SD complex and was largely unchanged by the addition of IscU_ec and CyaY (34, 40). In contrast, the low activity of the SD complex could be rescued by the addition of ISCU2 and FXN; the SDUF complex had a k_cat/K_M that is about double that of IscS_ec alone or as a S_ecU_ec or S_ecU_ecC_ec complex. The S_ecU_ec and SDU complexes also exhibit striking differences in their ability to facilitate Fe-S cluster formation. The S_ecU_ec complex is highly active in the Fe-S cluster assembly assay, whereas the SDU complex exhibits little ability to mediate this reaction in the absence of frataxin. These differences were not due to the identity of the scaffold protein: interchanging the IscU_ec and ISCU2 proteins did not substantially alter the higher rate of the complexes containing S_ec or the substantially inhibited rates of complexes containing SD (Figures 4, 6). Moreover, addition of CyaY or FXN to these reactions inhibited Fe-S assembly by S_ec-containing complexes and activated Fe-S assembly by SD-containing complexes (Figures 4, 6), albeit inhibition of the E. coli complex by CyaY was more efficient than by FXN and activation of the human complex by FXN was more efficient than by CyaY.

Our data therefore suggest that intrinsic features of IscS_ec and human SD cysteine desulfurase components control both the basal activity as well as the effect of binding the frataxin homologs CyaY and FXN. In both cases, binding of the effector molecule switches the activity, either from “off” to “on” or from “on” to “off”. In this sense, NFS1 behaves less like its ortholog IscS and more like its paralogs, such as SufS, which are characterized by modest activity that can be increased or activated by accessory proteins (67–69). Evolutionary innovations in cysteine desulfurases from eukaryotes that control the basal activity and intrinsic features of frataxin homolog regulation are not known; binding of ISD11 and conserved amino acid changes in eukaryotes near the active site could play important roles in this process. Precise molecular details that regulate these features are currently under investigation.

Our data are consistent with FXN and CyaY binding playing a role in modulating Fe-S cluster formation by trapping an “on” conformation of SDU or inducing an “off” conformation of S_ecU_ec (Figure 7). This model is consistent with the fact that SDU forms a stable complex with FXN (34, 35) and S_ecU_ec forms a stable complex with CyaY (21, 27, 28). The fact that FXN and CyaY can functionally replace each other in vitro, albeit at lowered efficiencies, is consistent with the formation of a complex with a common structure and with the ability of both of these proteins to suppress defects caused by deletion of YFH1, which encodes the S. cerevisiae frataxin homolog (38, 39). These modifications to the cysteine desulfurase may function as part of a protein-level regulatory mechanism. Future experiments will be aimed at providing additional mechanistic insight into the differences between the human SD complex and IscS_ec and how these differences control Fe-S cluster assembly.

Figure 7.

Working model for frataxin regulation of Fe-S cluster biosynthesis. In eukaryotes, a pre-equilibrium model is proposed in the absence of FXN in which a cysteine desulfurase and Fe-S cluster assembly deficient form is favored over a functional form of the SDU complex. FXN binding stabilizes the functional form and promotes sulfur transfer from NFS1 to ISCU2 and Fe-S cluster synthesis activities. In contrast, the prokaryotic S_ecU_ec complex that lacks CyaY exhibits cysteine desulfurase and Fe-S assembly activities. CyaY binding may induce a conformational change in the S_ecU_ec complex that does not significantly affect the cysteine desulfurase activity, but abolishes Fe-S cluster synthesis. The “ON” and “OFF” labels indicate the Fe-S assembly activity of the respective complexes, the blue arrows from the PLP to the mobile Cys loop indicate the presence of cysteine desulfurase activity, and the blue arrows from the Cys loop to the scaffold protein (shown in green) indicate interprotein sulfur transfer. The displayed model represents half of the expected dimeric Fe-S assembly complexes.

Abbreviations

C_ec

E. coli frataxin homolog CyaY

DTT

dithiothreitol

EDTA

ethylenediaminetetracetic acid

FRDA

Friedreich’s ataxia

FXN

human frataxin

HEPES

N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid

IPTG

isopropyl β-D-1-thiogalactopyranoside

ISC

iron-sulfur cluster assembly pathway

IscS_ec

the E. coli cysteine desulfurase IscS

IscU_ec

the E. coli scaffold protein IscU

NIF

nitrogen fixation Fe-S cluster assembly pathway

PLP

pyridoxal-5′-phosphate

SD

protein complex composed of human NFS1 and ISD11

SDU

protein complex composed of human NFS1, ISD11, and ISCU2

SDU_ec

protein complex composed of human NFS1 and ISD11 with E. coli IscU

SDUF

protein complex composed of human NFS1, ISD11, ISCU2, and FXN

SDU_ecF

protein complex composed of human NFS1, ISD11, and FXN with E. coli Iscu2

SDUC_ec

protein complex composed of human NFS1, ISD11, and ISCU2 with E. coli CyaY

SDU_ecC_ec

protein complex composed of human NFS1 and ISD11 with E. coli IscU and CyaY

S_ec

E. coli IscS

S_ecU_ec

protein complex composed of E. coli IscS and IscU

S_ecU

protein complex composed of E. coli IscS with human ISCU2

S_ecU_ecC_ec

protein complex composed of E. coli IscS, IscU, and CyaY

S_ecUC_ec

protein complex composed of E. coli IscS and CyaY with human ISCU2

S_ecU_ecF

protein complex composed of E. coli IscS and IscU and human frataxin

S_ecUF

protein complex composed of E. coli IscS with human ISCU2 and frataxin

SUF

sulfur mobilization Fe-S cluster assembly pathway

Tris

Tris(hydroxymethyl)aminomethane

U_ec

E. coli IscU