Defining the domains of Cia2 required for its essential function in vivo and in vitro

Amanda T. Vo , Nicholas M. Fleischman , Melissa D. Marquez , Eric J. Camire§ , Stephanie U. Esonwune , John D. Grossman , Kelly A. Gay , Jessica A. Cosman and Deborah L. Perlstein *
Department of Chemistry, Boston University, Boston, MA, USA. E-mail:

Received 12th June 2017 , Accepted 23rd August 2017

First published on 23rd August 2017

The cytosolic iron–sulfur cluster assembly (CIA) system biosynthesizes iron–sulfur (FeS) cluster cofactors for cytosolic and nuclear proteins. The yeast Cia2 protein is the central component of the targeting complex which identifies apo-protein targets in the final step of the pathway. Herein, we determine that Cia2 contains five conserved motifs distributed between an intrinsically disordered N-terminal domain and a C-terminal domain of unknown function 59 (DUF59). The disordered domain is dispensible for binding the other subunits of the targeting complex, Met18 and Cia1, and the apo-target Rad3 in vitro. While in vivo assays reveal that the C-terminal domain is sufficient to support viability, several phenotypic assays indicate that deletion of the N-terminal domain negatively impacts CIA function. We additionally establish that Glu208, located within a conserved motif found only in eukaryotic DUF59 proteins, is important for the Cia1–Cia2 interaction in vitro. In vivo, E208A–Cia2 results in a diminished activity of the cytosolic iron sulfur cluster protein, Leu1 but only modest effects on hydroxyurea or methylmethane sulfonate sensitivity. Finally, we demonstrate that neither of the two highly conserved motifs of the DUF59 domain are vital for any of Cia2's interactions in vitro yet mutation of the DPE motif in the DUF59 domain results in a nonfunctional allele in vivo. Our observation that four of the five highly conserved motifs of Cia2 are dispensable for targeting complex formation and apo-target binding suggests that Cia2 is not simply a protein–protein interaction mediator but it likely possesses an additional, currently cryptic, function during the final cluster insertion step of CIA.

Significance to metallomics

It is not well understood how the cytosolic iron sulfur cluster assembly (CIA) system or any FeS biogenesis system identifies its apo-protein substrates. Since Cia2 is the central scaffold of the targeting complex that executes apo-protein recognition, we systematically mutate each of its conserved motifs to identify those that mediate its protein–protein interactions. While this approach reveals a binding interface required to interact with the Cia1 subunit of the targeting complex, it also suggests that Cia2 has an additional function. Cia2's structural similarity to FeS cluster carriers indicates an active role in transferring the cluster during target maturation.


Iron–sulfur (FeS) clusters are essential and evolutionarily ancient cofactors.1 While it has been known for quite some time that FeS cofactors are required for primary metabolism and respiration within the mitochondria, it has been more recently recognized that FeS clusters are also required for DNA replication, transcription, and translation. Each of these fundamental processes require at least one FeS protein.2,3 The cytosolic iron sulfur cluster assembly (CIA) pathway assembles and inserts [Fe4S4] clusters into these nuclear and cytosolic metalloproteins.2

CIA culminates with insertion of the FeS cofactor into an apo-target, an FeS binding protein that is the substrate of the CIA system. Presently, it is not well understood how the CIA system recognizes the >20 different targets, each with a unique sequence, structure, and function. The CIA targeting complex, comprised of Cia1, Cia2, and Met18 in yeast, is proposed to mediate apo-target recognition.4–11 However, the function(s) of the individual subunits of this complex have not been clearly delineated. Do all of the subunits participate in target identification via direct protein–protein interactions or do they also play a more active role during the insertion of the FeS cluster into target? Additionally, the mechanism of target identification is not well defined.7–10 How does the same complex identify the nuclear Rad3 helicase and the cytosolic Leu1 isomerase when these proteins share not detectable sequence or structural homology?11 Insight into these important questions has been slow due in part the inability to reconstitute target identification in the test tube so that the roles of the individual subunits and their conserved residues can be investigated.

Herein we begin addressing these questions by investigating the function of the CIA targeting complex subunit Cia2. Like a majority of cluster biogenesis genes, Cia2 is essential.12 Recent work has demonstrated that Cia2 is the central component of the targeting complex, tethering Met18 to Cia1.7,8,11 The C-terminal half of Cia2 contains a domain of unknown function 59 (DUF59). Proteins harboring this domain not only play an essential role in CIA, but they are also frequently genetically and functionally associated with cluster biogenesis proteins in bacteria and archaea.13,14 For example, some bacteria encode a DUF59 protein called SufT in the SUF FeS cluster biogenesis operon.13S. aureus SufT was recently demonstrated to be an FeS biogenesis protein which shares functional overlap with the Nfu FeS cluster carrier.14–16 Thus, the developing picture is that DUF59 proteins play a role in cluster biogenesis across all domains of life.

To gain insight into the role of Cia2 and its DUF59 domain, we report the first biochemical and biophysical investigation of yeast Cia2. To begin defining the role(s) of its conserved residues, we identify three motifs found in eukaryotic Cia2 homologs, but not in bacterial or archaeal DUF59 proteins. Mutagenesis reveals that one of these motifs is vital for the Cia1–Cia2 interaction. While the DUF59 motifs are vital for Cia2 function in vivo, but they do not diminish any of Cia2's protein–protein interactions in vitro. These results suggest that Cia2 has an additional function beyond mediating protein–protein interactions required for formation of the targeting complex and/or identification of apo-targets.


Bioinformatic analysis

Sequences of DUF59 domain proteins from organisms indicated in Fig. S15 (ESI) of Tsaousis et al. were collected.13 After eliminating proteins that are DUF59-MRP fusions, fragments, and redundant sequences, the sequences were aligned using Clustal Omega. For structural alignment of human Cia2a (PDBID 2m5h) with Arabidopsis thaliana Nfu (PDBID 2z51), the proteins were aligned with DALI and visualized with PyMol. Predictions of disordered regions were completed with the MFDp2 server.17 The mean net charge and hydrophobicity were calculated via the PONDR server (

Plasmid construction for E. coli expression

PCR amplified inserts were ligated into restriction enzyme digested vectors by the method of Gibson.18 DNA sequencing was used to confirm successful construction of the plasmids. For expression of untagged Cia2 in E. coli, the gene was amplified from yeast genomic DNA and ligated between the NcoI and NdeI digested pET15b plasmid. For double-tagged Cia2 (dtCia2), Cia2 was amplified and ligated between the BamHI and SalI sites of pET52b to encode a protein with the following features in the following order: Strep-II tag, thrombin protease cleavage site, Cia2, HRV3C protease cleavage site, and a 6xHis-tag. For Δ102-Cia2, the C-terminal half of Cia2 was amplified and ligated between the NdeI and XhoI sites of pET24b. The resulting construct encodes residues 103–231 of Cia2 followed by a C-terminal 6xHis-tag. The E208A mutation was made by QuikChange mutagenesis (Agilent). All other mutations were constructed via Q5 mutagenesis (New England Biolabs) with primers designed by the NEBaseChanger webserver.

For N-terminally His-tagged Cia1 (HisCia1), the gene was amplified from plasmid ScCD00012999 obtained from DNASU stock center and inserted between EcoRI and NotI sites of pETDuet-1. For double-tagged Cia1 (dtCia1), the amino acids for a StrepII tag and a TEV protease site were inserted into the His-Cia1 vector by an inverse PCR amplification followed by circularization of the PCR product. The resulting dtCia1 vector encodes the following features in the following order: His tag, TEV protease site, StrepII-tag, Cia1.

Cia2 expression and purification

E. coli BL21(DE3) transformed with the dtCia2 plasmid was grown at 37 °C, induced with IPTG (0.5 mM) at an OD600 = 0.7, and collected 4 h later. All protein purification steps were performed at 4 °C. Cell paste was resuspended in Buffer A (50 mM HEPES pH 8, 300 mM NaCl, 10% glycerol, 10 mM betamercaptoethanol (BME)) supplemented with 20 mM imidazole and 1% Triton X-100. DNaseI and protease inhibitors were added and the cells were disrupted by microfluidizer. The clarified lysate was mixed with His-Bind resin for 1 h, then the resin was collected, washed with 25 column volumes (CV) of Buffer A with 20 mM imidazole, 25 CV of Buffer A with 50 mM imidazole and eluted with Buffer A with 350 mM imidazole. Fractions containing Cia2 were pooled, 5 mM DTT was added, and the protein was buffer exchanged with a PD10 column into Buffer B (50 mM HEPES pH 8, 300 mM NaCl, 5 mM DTT, 10% glycerol). Cia2 was concentrated to ≤0.5 mg mL−1 and stored at −80 °C.

Refolding of Cia2

Cells were lysed as described for dtCia2. The inclusion bodies were washed with 50 mM Tris pH 8, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 0.5% Triton X-100, 5% glycerol and a second time in the same buffer without the Triton X-100. The pellet was resuspended in 50 mM Tris pH 8, 200 mM NaCl, 2 mM EDTA, 7 M GuHCl. Denatured Cia2 (1 mL) was added dropwise to 49 mL of 50 mM Tris pH 8, 25 mM NaCl, 5% glycerol, 1 mM EDTA, 1 mM DTT. Refolding of the Cia2 mutants was enhanced by the additional inclusion of 500 mM arginine in the refolding buffer. Two hours after the rapid dilution, the mixture was centrifuged, concentrated, buffer exchanged, and stored as described above for dtCia2.

Purification of Cia1

HisCia1 and dtCia1 were expressed as described for dtCia2. Cells were resuspended in 50 mM Tris pH 8.0, 100 mM NaCl, 10% glycerol, 5 mM betamercaptoethanol (BME) with protease inhibitor cocktail and DNase I. Following sonication and centrifugation, the clarified lysate was loaded on a HisBind column, washed with 50 CV resuspension buffer with 5 mM imidazole and eluted in the same buffer supplemented with 300 mM imidazole. Cia1 was dialyzed overnight against the resuspension buffer, concentrated, and stored at −80 °C.

To obtain Strep-Cia1, the His-tag was removed from dtCia1 by the addition of His-tagged TEV protease (1 mg mL−1). Following overnight incubation at room temperature, the mixture was passed over a HisBind column and Strep-Cia1 was recovered from the flow-through.

Affinity copurification experiments

Rad3 was purified as described.11 Met18 with an N-terminal His-SUMO tag (SUMOMet18) was purified as described except it was additionally chromatographed on a Superdex 200 Increase 10/300 GL (GE healthcare) column to remove low molecular weight contaminants.11 The SUMO-tag was removed by overnight incubation at 4 °C with SUMO protease. Untagged Met18 was recovered from the IMAC flow-through. Affinity purification experiments were performed as described.11 In all cases, a negative control omitting the bait was completed in parallel. In the cases where a mutation was found to be disruptive of an interaction, a positive control with wt-Cia2 was completed in parallel.

Circular dichroism

Circular dichroism (CD) spectra were acquired as previously described.19 Briefly, proteins (0.1–0.5 mg mL−1) were buffer exchanged into 10 mM KPO4, 100 mM KCl, 0.5 mM DTT, pH 8. Spectra were acquired with an Applied Photophysics CS/2 Chirascan CD spectrometer from 200–260 nm. Buffer interference below 200 nm prevented collection of data in this region. All spectra are an average of at least three scans and were baseline corrected to remove the buffer contribution. The manufacturer's software was used to smooth the data convert it to molar ellipticity or mean residue ellipticity as indicated. Analysis of the Θ222versus Θ200 was completed with CAPITO webserver.20


Untagged Cia2 was analyzed using a Bruker Autoflex Speed MALDI-TOF. The sample (∼0.1 mg mL−1 Cia2 in 10 mM Tris pH 8, 0.5 mM DTT) was mixed with an equal volume of α-cyano-4-hydroxycinnamic acid in 90% acetonitrile supplemented with 0.1% TFA. Samples (1.5 μL) were spotted on the MALDI target and allowed to air dry prior to acquisition of spectra.

Yeast complementation analysis

A yeast shuttle vector based on pRS313 (centromeric plasmid with a HIS3 marker) was constructed to express Cia2 with a C-terminal MYC-tag from an ADH (constitutive) promoter. Q5 mutagenesis (New England Biolabs) was performed to create the Δ102-Cia2 and NPQ-Cia2 mutants. E208A–Cia2 was created by the QuikChange mutagenesis (Agilent).

A Tet-regulated Cia2 strain (TH_3222; MATa URA3::CMV-tTa his3-1 leu2-0 met 15-0 KanR-TetO7-CIA2; GE-Dharmacon) was transformed with plasmids for expression of Cia2 alleles.21 Log phase cultures were serial diluted, applied to SD-Ura-His + G418 plates in the presence or absence of 50 μg mL−1 doxycycline (DOX) and in the presence of other additives as indicated, and incubated for 3 days at 30 °C before imaging. For methylmethane sulfonate (MMS) sensitivity, cells were applied to YPAD with 0.008% MMS the presence or absence of DOX. For sulfite reductase activity, cells were plated on YPAD supplemented with 0.001% (w/v) bismuth ammonium citrate, 0.003% (w/v) sodium sulfite, and 0.0003% (w/v) ferrous sulfate in the presence or absence of DOX. For hydroxyurea (HU) sensitivity, cells were plated on YPAD with 25 mM HU in the presence and absence of DOX. For NPQ-Cia2, a colony was suspended in sterile water, serial diluted and applied to plates since the strain was difficult to propagate in liquid culture.

Leu1 activity assays

The TET-Cia2 strain (TH_3222) transformed with plasmids for wtCia2 or its mutant alleles were grown overnight in SD-Ura-His + G418 + DOX, diluted into YPAD supplemented with G418 and DOX (50 μg mL−1), and grown to mid-log phase at 30 °C. Crude extracts were generated in an anaerobic chamber and Leu1 assays carried out as described.22


Eukaryotic Cia2 homologs share five conserved motifs

We began by analyzing the sequences of eukaryotic DUF59 proteins to delineate their conserved residues, motifs and domains. Nearly all eukaryotes encode at least one Cia2 homolog and some organisms encode a paralogous pair, referred to as Cia2a (Fam96A) and Cia2b (Fam96B or Mip18).10,13 Cia2b is a general FeS assembly factor and is the ortholog of yeast Cia2. Cia2a is proposed to specifically direct maturation of iron responsive protein 1 (IRP1).10 Using a recent analysis of CIA evolution as our guide, we aligned a diverse set of 48 Cia2 homologs, including 17 Cia2a/Cia2b pairs.13 We found five conserved motifs distributed between an N-terminal acidic domain and a C-terminal domain containing the DUF59 (Fig. 1 and Fig. S1, ESI).
image file: c7mt00181a-f1.tif
Fig. 1 (A) The Cia2 sequence is annotated as follows: predicted regions of disorder, dark grey shading;17 DUF59 domain, light grey shading; Motif 1, red; Motif 2, orange; Motif 3, green; Motif 4, blue; Motif 5, purple. The positions of Leu103 (first residue of Δ102-Cia2), Cys161, and Glu208 are each marked with a black circle. (B) Surface representation of human Cia2a (PDB ID 2M5H)25 with Motifs 3, 4, and 5 colored as in Panel A. Cys161 and Glu208 are colored yellow and red, respectfully. (C) Comparison of the mean net charge versus mean hydrophobicity of full length Cia2 (black diamond) and the N-terminal domain (blue circle). The border (dashed line) between sets of natively folded (grey) and unfolded (red) protein standards is also shown.23

The N-terminal domain has two conserved regions, an NxNP motif and a patch of acidic residues (Motifs 1 and 2, Fig. 1). This N-terminal extension is missing in bacterial and archaeal DUF59 proteins and in the Cia2a paralogs (Fig. S2, ESI). There is little sequence conservation between these motifs except for an enrichment in polar and charged residues, suggesting this domain is intrinsically disordered (Fig. 1C and vide infra).23

The C-terminal region begins with the DUF59 domain and its characteristic motifs (Motifs 3 and 4, Fig. 1). These motifs are close to one another, forming a putative active site (Fig. 1B and Fig. S1, ESI).24–26 Motif 5 (purple), found C-terminal to the DUF59, is missing from archaeal and bacterial sequences (Fig. S2, ESI). Since Motifs 1, 2, and 5 are unique to eukaryotes, they likely play a CIA-specific function, such as interaction with Met18 or Cia1. The conservation of Motifs 3 and 4 in all DUF59 proteins suggests that they play a common function in cluster biogenesis across all domains of life.

Purification of Cia2

Since the proposed function of Cia2 is to bind Met18, Cia1 and apo-targets,7,8,10,11 we hypothesized that its conserved motifs could house its protein–protein interaction sites. To test this hypothesis, we developed methods to purify Cia2 and assess its interactions in vitro. When expressed in E. coli, most of the Cia2 was insoluble. However, some soluble double-tagged Cia2 (dtCia2) could be purified via IMAC (Fig. 2A). Since the yield was low (≤0.2 mg g−1 cell paste), we also developed a method to refold Cia2 via rapid dilution of solubilized inclusion bodies. Overall, refolding yielded protein of comparable purity as the IMAC purified protein with a 10-fold increase in the yield (Fig. 2A, lane 2).
image file: c7mt00181a-f2.tif
Fig. 2 (A) SDS-PAGE analysis of IMAC purified dtCia2 (lane 1, 31 kDa), refolded untagged Cia2 (lane 2, 25 kDa), and Δ102-Cia2 (lane 3, 15 kDa). Migration of molecular weight standards (M) are indicated in kDa. (B) CD spectra of refolded untagged Cia2, IMAC purified dtCia2 (orange), refolded Δ102-Cia2 (green), and a subtraction of Δ102-Cia2 from untagged Cia2 (blue). Data are plotted as mean residue ellipticity (MRE, 103Θ cm2 dmol−1 residue−1) versus wavelength. (C) The molar ellipticity ([Θ] in degrees cm2 dmol−1) at 222 and 200 nm of the CD spectra in Panel B (coloring as in Panel B).20 Also shown are the [Θ]222versus [Θ]200 values for a standard set of intrinsically disordered and folded proteins in light grey and dark grey, respectively. The regions containing unfolded coil and premolten globule standards are indicated.20,23

To assess refolding efficiency, the CD spectra of refolded and IMAC purified protein were compared (Fig. 2B). All preparations exhibit a minimum near 206 nm with a smaller shoulder at 222 nm. Although there were differences in the relative intensities of these negative peaks, the overall shape of the IMAC purified and refolded Cia2 spectra was similar. While the differences in the CD spectra indicate the two preparations could have minor structural differences, we have observed that both the refolded Cia2 and the IMAC purified Cia2 can form all of the protein–protein interactions required to form the targeting complex and tether the model apo-target Rad3 to that complex (Fig. 3A). We concluded that the refolding is an effective approach to access mg quantities of recombinant Cia2 for in vitro analysis.

image file: c7mt00181a-f3.tif
Fig. 3 SDS-PAGE analysis of affinity copurification assays. In all panels, a bait (*) is mixed with one or more prey proteins as indicated (input) and passed through an affinity resin specific to the tag appended to the bait, either IMAC (Panel B and C) or Streptactin (Panels A, D, and E). The column is washed then the bound proteins are eluted and analyzed by SDS-PAGE (Elution). A negative control omitting the bait is also included in each panel. Panel A compares the ability of wt-Cia2 (W) and Δ102-Cia2 (Δ) to tether Rad3 to the DTCia1 bait. Panel B compares the ability of wt-Cia2 (W) and E208A–Cia2 (M) to form the Met18–Cia2 or the Cia1–Cia2 binary complexes. Panel C compares the ability of wt-Cia2 (W) and E208A–Cia2 (M) to scaffold the targeting complex by simultaneously binding Met18 and Cia1. Panels D and E demonstrate C161A-Cia2 and NPQ-Cia2 can both form the targeting complex and bind Rad3. The relative migration of MW standards and positions of CIA targeting complex subunits and Rad3 are indicated to the left and to the right in each panel, respectively. The data shown is representative of at least three independent experiments.

Cia2's intrinsically disordered domain is dispensable for its function in vitro and in vivo

We noticed that both the IMAC purified dtCia2 and refolded Cia2 exhibited aberrantly slow migration in SDS-PAGE. For example, dtCia2 (31 kDa) migrates with an apparent molecular weight of 44 kDa (Fig. 2A). MALDI-TOF analysis revealed this was not due to a post-translational modification or additional residues inadvertently introduced during cloning (Fig. S3, ESI). Since proteins with intrinsically disordered domains display this behavior,27 we examined Cia2's sequence with the Multilayered Fusion-based Disorder Predictor (MFDp2) which analyzes sequences with multiple different algorithms to return a disorder propensity score.17 This analysis revealed two regions in Cia2's N-terminal domain with a high probability of being disordered (dark grey, Fig. 1A).

To determine if the N-terminus is disordered, we removed the first 102 residues (Δ102-Cia2) and refolded the protein from insoluble inclusions. Δ102-Cia2's SDS-PAGE migration behavior normalized (Fig. 2A, lane 3). The shape of its CD spectrum also significantly changed (Fig. 2B, green). When we subtracted the CD spectrum of Δ102-Cia2 from that of untagged Cia2, we found that the molar ellipticity at 200 and 222 nm in the subtracted spectrum clusters with proteins forming pre-molten globules (blue, Fig. 2B and C).23 The data is consistent with the N-terminal half of Cia2 having a large amount of intrinsic disorder.

When we inspected the sequences of Cia2 homologs, we found all organisms examined have at least one Cia2 homolog with a disordered region. However, the intrinsically disordered domain (IDD) is missing in the Cia2a paralogs (Fig. S2, ESI). The recent report that Cia2a/b pairs direct maturation of different targets combined with our observation that the IDD is a distinguishing feature for this paralogous pair prompted us to investigate whether the N-terminal domain mediates interaction with targets.10 Using an affinity copurification assay, we compared ability of Cia2 and Δ102-Cia2 to bind to Met18, Cia1, and Rad3. When we mixed the double-tagged Cia1 (dtCia1, with both Strep- and His-tags) bait with SUMOMet18, Rad3, and Cia2, we observed similar amounts of each prey protein in the streptactin column elution fraction regardless of whether Cia2 or Δ102-Cia2 was used in the assay (Fig. 3A). Thus, the IDD is dispensable for Cia2's interactions in vitro.

We next wanted to know if the N-terminal domain is required for Cia2's essential function in vivo. For this experiment, we used a commercially available strain in which Cia2's promoter is replaced with a Tet-regulated promoter.21 In the absence of a plasmid-born Cia2, this TET-Cia2 strain cannot grow in the presence of doxycycline (DOX), which represses expression of genomic Cia2 which is essential.12 We introduced plasmids bearing Cia2 and Δ102-Cia2 into this strain and found that both constructs were able to complement the doxycycline-induced growth defect whereas the empty vector (EV) control could not (Fig. 4A). Thus, the C-terminal domain is sufficient to support CIA function in vivo.

image file: c7mt00181a-f4.tif
Fig. 4 Assays to assess functionality of Cia2 mutants in vivo. (A) For the complementation analysis, the Tet-Cia2 strain was transformed with plasmids for expression of Cia2, either wt or mutant alleles under control of an ADH promoter, or the empty vector (EV) control as indicated. Yeast were grown to mid-log phase in SD-Ura-His media and spotted on YPAD supplemented with additives. The inclusion of 50 μg mL−1 doxycycline (+Dox) represses expression of the genomic Cia2. (B) Leu1 activity in the soluble cell extract generated from Tet-Cia2 strain expressing the indicated Cia2 allele was compared. Yeast were grown to mid-log phase in the presence of DOX. The cells were collected, lysed and the soluble extract was assayed for Leu1 activity. The data shown represent the average ± the standard deviation of at least three independent determinations.

To probe the functionality of Δ102-Cia2 more deeply, we examined the activity of the FeS-dependent sulfite reductase by growing the complementation strains on media supplemented with bismuth sulfite. Active sulfite reductase results in formation of the brown Bi2S3 precipitate. Although the additives appeared to interfere with the DOX inhibition leading to slow growth of the empty vector control, we observed that the Δ102-Cia2 complemented strains were similar in color to the empty vector control on the plates containing doxycycline and lighter in color than the strain complemented with wt-Cia2 (Fig. 4A, second panel). This suggested Δ102-Cia2 might have compromised functionality. To quantitatively assess target maturation, we compared the activity of the cytosolic FeS protein Leu1. We found a significant decrease in the Leu1 activity in extracts derived from the Δ102-Cia2 complemented strain as compared to the wild-type control (Fig. 4B).

To also assess cluster targeting to nuclear FeS proteins, we examined the Δ102-Cia2 strain's sensitivity to the DNA damaging agent methylmethane sulfonate (MMS) and the DNA replication inhibitor hydroxyurea (HU).4,5,9,28,29 We observed diminished resistance to both reagents compared to the strain complemented with wt-Cia2 (Fig. 4A, bottom panels). Western blotting revealed that both the full length and the truncated Cia2 are expressed (Fig. S4, ESI). However, the expression level of Δ102-Cia2 appeared smaller than that of the full-length protein. Together, these results demonstrate that Δ102-Cia2 is sufficient to support cell viability but with a diminished ability to support CIA target maturation possibly due to lower stability of this construct in vivo.

Glu208 of Motif 5 is required for the Cia1–Cia2 interaction

The C-terminal half of Cia2 comprises the DUF59 domain and ∼40 additional amino acids which house Motif 5 (purple, Fig. 1). Motif 5 is unique in eukaryotic DUF59 proteins suggesting it could be important for formation of the targeting complex (Fig. S2, ESI). In fact, an in vivo protein–protein interaction study previously found that Glu208 found within Motif 5 (red, Fig. 1B) is important for the interaction between Met18 and Cia2.30

To pinpoint the function of Motif 5, we examined E208A–Cia2's interactions via affinity copurification. When wt-Cia2 or E208A–Cia2 was mixed with the SUMOMet18 bait, we were surprised to find similar amounts of the Cia2 prey in the elution fractions (Fig. 3B). This unexpected result prompted us to examine the mutant protein's ability to bind to Cia1 and found that the E208A mutation disrupts the Cia1–Cia2 complex (Fig. 3B). Consistent with Cia2's role as the bridge linking Met18 to Cia1 in the targeting complex,11 we also found that E208A–Cia2 was able to tether less SUMOMet18 to the HisCia1 bait as compared to the wt-Cia2 control (Fig. 3C). We concluded that E208A–Cia2 has a defect in its ability to bind Cia1, but has no observable defect in its ability to bind Met18.

Since this result is at variance with the conclusions of Lev et al.,30 we additionally examined whether E208A–Cia2 could complement depletion of wt-Cia2 in the TET-Cia2 strain. In agreement with the previous study, E208A–Cia2 can support viability (Fig. 4A, top panel). We additionally examined the sulfite reductase activity and the HU and MMS sensitivity of the E208A–Cia2 complemented strain. We observed little, if any effect, of the E208A mutation on the color of the colonies grown on bismuth sulfite media and a slightly increased sensitivity to HU and MMS compared to wt-Cia2 control (Fig. 4A). The modest effects observed in the qualitative assays prompted us to quantitatively assess maturation of the Leu1. We observed the E208A mutation results in a 6-fold reduction in Leu1 activity (Fig. 4B) while Western blotting revealed an expression level comparable to that of wt-Cia2 (Fig. S4, ESI). All together our results demonstrate that E208 is important for association with Cia1 and that destabilization of the Cia1–Cia2 complex can negatively impact CIA function in vivo.

DUF59 motifs are essential for Cia2's function in vivo but are dispensable for its protein–protein interactions in vitro

The two remaining motifs, Motif 3 and 4, are found within Cia2's DUF59 domain. It was previously reported that mutation of C161 within Motif 4 results in a dominant negative phenotype, but no mutations of Cia2's Motif 3 have been reported.12 We mutated “DPE” sequence of Motif 3 to “NPQ” and examined how this mutation affects CIA function via the complementation assay. Although we could obtain transformants, the NPQ-Cia2 strain grew slowly in the absence of DOX and it failed to grow on plates supplemented with DOX (Fig. 5). Since it was difficult to propagate the NPQ-Cia2 complemented strain due to the apparent growth inhibitory effect of this allele, we could not assess its sensitivity to HU or MMS or its effect on Leu1 activity. However, on plates supplemented with bismuth sulfite, we observed that the NPQ-Cia2 strain was significantly lighter in color as compared to the strain expressing wt-Cia2. We concluded that NPQ-Cia2 is a nonfunctional allele and has a similar phenotype to that reported for mutation of the other conserved region of the DUF59 domain.12
image file: c7mt00181a-f5.tif
Fig. 5 The Tet-Cia2 strain was transformed with plasmids for the indicated Cia2 allele grown in the presence (right) or absence (left) of doxycycline, or bismuth sulfite (bottom).

Next, we tested whether the residues of Motifs 3 or 4 were required for any of Cia2's protein–protein interactions in vitro. We reasoned that if the DUF59 binds to apo-FeS proteins, this could explain the common function of this domain in cluster biogenesis. When we mixed the C161A-dtCia2 bait with SUMOMet18, HisCia1, and Rad3 prey proteins, we found all three prey proteins in the elution fraction (Fig. 3D). We observed a similar result when the NPQ variant was used in the copurification assay (Fig. 3E) or if Leu1 was used as the model apo-target in place of Rad3 (data not shown). We concluded that neither DUF59 motif is vital for any of Cia2's interactions in vitro.

The only validated biochemical function of Cia2 is mediation of interactions essential for formation of the targeting complex and binding apo-targets. Therefore, it was surprising that mutation of the invariant residues of the DUF59 domain did not affect Cia2's interactions, especially given that mutation of the invariant residues within the DUF59 domain are nonfunctional in vivo (Fig. 5).12 This observation suggested to us that Cia2 might have an additional function in target maturation. In fact, we noticed during our bioinformatics analysis that the DUF59 domain (PFAM family PF01883) belongs to the same PFAM clan as the “NifU domain” (PFAM family PF01106).31 NifU is a three-domain protein that serves as the FeS cluster scaffold for nitrogenase metallocofactor maturation.32,33 PF01106 corresponds to NifU's C-terminal domain and it is also found in Nfu FeS cluster carriers. These Nfu carriers bind a cluster at their homodimeric interface via a conserved CxxC motif (Fig. S5A, ESI).34–36 A structural alignment of Arabidopsis Nfu (CnfU) and the human Cia2a paralog reveals these two domains share the same fold (Fig. S5, ESI).31 Moreover, the cysteine of the DUF59 domain (C161 of Cia2) aligns in three dimensional space with the second cysteine in Nfu's cluster-ligating CxxC motif (Fig. S5, ESI).

The structural similarity between Nfu carriers and DUF59 proteins and the conservation of one of the two cluster ligating ligands prompted us to examine whether Cia2 binds an FeS cluster. The UV-Vis spectra of some Cia2 preparations purified by IMAC had absorption features in the low 400 nm region which were suggestive of [Fe2S2] or [Fe4S4] binding (Fig. S6, ESI). However, the intensity of this feature was always significantly lower than one would expect for stoichiometric FeS cluster binding. Furthermore, the intensity of these features in the as isolated protein varied between different preparations of the same construct. We tried unsuccessfully to increases cluster loading in the as-isolated protein via several approaches including anaerobic expression and purification, expression in iron supplemented media, and by coexpression with ISC operon or with the other targeting complex subunits Met18 and Cia1. We also attempted to chemically reconstitute a cluster on the refolded Cia2 or the Cia1–Cia2 complex. However, Cia2's instability and propensity to precipitate during or immediately following chemical reconstitution prevented us from isolating and characterizing an FeS-bound form of the protein.


Currently we have little information about the function of Cia2 during the last step of cytosolic iron sulfur cluster assembly. Herein, we began addressing these gaps in our understanding by first designating the conserved regions of Cia2 and subsequently probing the roles of each motif with in vivo and in vitro assays. We analyzed Cia2's eukaryotic homologs and compared these to DUF59 proteins derived from bacteria and archaea. This analysis combined with our CD study revealed that all eukaryotes encode at least one Cia2 homolog with a relatively long IDD at its N-terminus (Fig. 1 and Fig. S2, ESI). Since IDDs play well-recognized roles in scaffolding complexes and mediating promiscuous or plastic protein–protein interactions,37 we examined whether the N-terminal domain was required for Cia2's protein–protein interactions in vitro. We found that the first 100 residues of Cia2 are dispensable for its interaction with Cia1, Met18 and Rad3 (Fig. 3A). We additionally found that this domain is dispensable for Cia2's essential function in vivo, albeit with diminished capacity to support CIA function (Fig. 4).

Although we cannot rule out the possibility that the N-terminal domain plays a role in stabilizing one or more of Cia2's interactions, we think it is more likely that the IDD serves a regulatory function, such as providing sites for post-translational modifications.37 In support of this idea, the presence of this domain appears to be a distinguishing feature between the Cia2a/b paralogs (Fig. S2, ESI) and a recent study demonstrated that these paralogs differ not only in their target specificities but also in how their depletion affects cellular iron homeostasis.10 Moreover, it was recently reported that Cia2b appears to be the key factor for regulation of CIA by its degradation in the absence of Met18.38 Our observation that the amount of Δ102-Cia2 in extracts appears to be significantly lower than that of the full-length protein is also consistent with the notion the N-terminal domain plays a role in regulating Cia2's in vivo concentration (Fig. S4, ESI). Although further studies will be required to define the role of this domain and its NxNP motif (Motif 1), our study definitively establishes that the N-terminal domain is not vital for any of Cia2's protein–protein interactions in vitro (Fig. 3A).

One area of great interest is delineating the interaction interfaces important for formation of the targeting complex and the residues required for target recognition. Toward that end, we show that mutation of Glu208 within Motif 5 disrupts the Cia1–Cia2 complex in vitro (Fig. 3B). This observation was surprising since it was previously reported that E208G–Cia2 is defective in association with Met18 in vivo.30 However, it can be challenging to distinguish direct effects from indirect effects with in vivo protein interaction assays like the protein complementation approach employed by Lev et al. Indeed, these authors also reported that depletion of Cia1 diminished the Met18–Cia2 interaction.30 Several studies have also noted that depletion or overexpression of one targeting complex subunit can affect the concentrations of the other subunits, which could indirectly affect the amount of complex observed in co-IPs from extracts.6,7,38 We think it is likely the E208G mutation destabilizes the Cia1–Cia2 complex, which in turn affected the amount of the Met18–Cia2 complex detected via protein complementation. Our findings highlight how difficult it can be to dissect the molecular details underlying formation of multiprotein complexes such as the CIA targeting complex using in vivo approaches. With access to the fully reconstituted recombinant system in vitro analysis, hypotheses generated from cell-based approaches can now be directly assessed with the assays developed herein. Ultimately, we expect that a model of the Met18–Cia1–Cia2 complex could be constructed once key interface residues have been identified.

Finally, it was surprising how few of the mutations tested resulted in disruption of Cia2's interactions in vitro in light of the fact that mediation of protein–protein interactions is only validated activity for Cia2. In particular, it was unexpected that mutation of Motifs 3 and 4 produced no observable perturbation in Cia2's interaction with Cia1, Met18, or Rad3 (Fig. 3D and E). The recently reported structures of the human Cia2a paralog together with the structures of DUF59 proteins from bacteria and archaea have demonstrated that Motif 3 and 4 are in close proximity to one another (Fig. S1, ESI).24–26 Our finding that mutation of Motif 3 appears to inhibit cell growth even in the presence of the wt-allele is reminiscent of the report that the C161A variant of Cia2 is dominant negative. The similar phenotypes exhibited by mutation of Motifs 3 and 4 supports the notion that these two highly conserved regions cooperate to execute the same biochemical function.12 All together our in vitro and in vivo data suggest that these mutants compete with wt-Cia2 for binding to Met18 and Cia1, ultimately tying up these subunits to a nonfunctional CIA targeting complex.

All together, we conclude from our in vitro and in vivo studies that Cia2 must possess an additional biochemical activity on top of its established ability to bind to Met18, Cia1 and apo-targets. One possibility is that Cia2 has yet another protein–protein interaction partner and mutation of Motifs 3 and 4 disrupts this interaction. We do not favor this model based on the relatively modest phenotype observed for the E208A mutation which disrupts the Cia1–Cia2 complex as compared to growth inhibitory phenotypes observed for the Motif 3 and 4 mutations (Fig. 5).12

Instead, the functional overlap between DUF59 and Nfu in bacteria14–16 combined with the structural similarity between these two domains (Fig. S5, ESI) strongly suggests that Cia2's cryptic function could be related to interacting with the FeS cluster as it is inserted into apo-targets. Although Cia2 has just one highly conserved cysteine, this residue aligns with the cluster binding residue of Nfu carriers (Fig. S5, ESI). The human Cia2a structure also has revealed this cysteine is at the homodimeric interface, just like the cluster binding CxxC motif of Nfu (Fig. S1, ESI).26,39 Our recent isolation of Cia1·Cia22, Met18·Cia24, and Met18·Cia12·Cia24 complexes demonstrates that formation of Cia2 dimers or higher order complexes could be a conserved feature.11 Additionally, the demonstration that the dimeric monothiol glutaredoxins can bind an FeS cluster ligated via a single conserved cysteine and glutathione ligands provides evidence that a protein with one conserved cysteine could bind an FeS cluster.40

While the bioinformatic and genetic evidence strongly supports the cluster carrier hypothesis, we and others have been unsuccessful at isolating Cia2 or other DUF59 proteins bound to an FeS cluster.14 Since the structure of Cia2a shows that the putative FeS binding site is likely to be surface exposed, this ligation environment would likely make the cluster labile and prone to oxidative degradation. This could explain our inability to isolate an FeS-bound Cia2. However, our results could also indicate that Cia2 is not the direct FeS donor during target maturation and it is not a cluster carrier. If this is the case, we favor a model where Cia2 plays a role interacting with the nascent FeS cluster as it is being transferred from the immediate donor and inserted into an apo-target bound to the targeting complex. Undoubtedly, the ability to assemble the targeting complex with Cia2 mutants for in vitro analysis is a vital first step toward defining the function of Cia2 and the role of the DUF59 domain in FeS biogenesis.


This work is aimed at defining the function of Cia2 in CIA. By systematically mutating all of its conserved motifs, we identify surprisingly few protein–protein interaction sites given that the established function of this protein is to bind to Met18, Cia1 and apo-targets. Instead, we conclude that Cia2 exploits its DUF59 domain for an additional, currently cryptic function.

Conflicts of interest

There are no conflicts to declare.


We thank the NSF for support for ATV (DGE-0947950 and CHE-1555295), MDM (DGE-0947950 and DGE-1247312) and JDG (CHE-1555295) and for access to the CD spectrometer (MRI, CHE-1126545) and MALDI-TOF (MRI, CHE-1337811). We also thank Boston University for funding this work in the form of a startup grant. SUE and JAC also thank the Undergraduate Research Opportunities Program for support. We thank the Chemical Instrumentation Center staff for advice and training, Brahm Gardner for completion of the Cia2 refolding screen and Jin Gao for creation of the Cia2 and dtCia2 vectors.


  1. D. C. Johnson, D. R. Dean, A. D. Smith and M. K. Johnson, Structure, function, and formation of biological iron-sulfur clusters, Annu. Rev. Biochem., 2005, 74, 247–281 CrossRef CAS PubMed.
  2. V. D. Paul and R. Lill, Biogenesis of cytosolic and nuclear iron-sulfur proteins and their role in genome stability, Biochim. Biophys. Acta, 2015, 1853, 1528–1539 CrossRef CAS PubMed.
  3. J. O. Fuss, C. L. Tsai, J. P. Ishida and J. A. Tainer, Emerging critical roles of Fe-S clusters in DNA replication and repair, Biochim. Biophys. Acta, 2015, 1853, 1253–1271 CrossRef CAS PubMed.
  4. O. Stehling, A. A. Vashisht, J. Mascarenhas, Z. O. Jonsson, T. Sharma, D. J. A. Netz, A. J. Pierik, J. A. Wohlschlegel and R. Lill, MMS19 Assembles Iron-Sulfur Proteins Required for DNA Metabolism and Genomic Integrity, Science, 2012, 337, 195–199 CrossRef CAS PubMed.
  5. K. Gari, A. M. León Ortiz, V. Borel, H. Flynn, J. M. Skehel and S. J. Boulton, MMS19 links cytoplasmic iron-sulfur cluster assembly to DNA metabolism, Science, 2012, 337, 243–245 CrossRef CAS PubMed.
  6. S. Ito, L. J. Tan, D. Andoh, T. Narita, M. Seki, Y. Hirano, K. Narita, I. Kuraoka, Y. Hiraoka and K. Tanaka, MMXD, a TFIIH-independent XPD-MMS19 protein complex involved in chromosome segregation, Mol. Cell, 2010, 39, 632–640 CrossRef CAS PubMed.
  7. M. Seki, Y. Takeda, K. Iwai and K. Tanaka, IOP1 Protein Is an External Component of the Human Cytosolic Iron-Sulfur Cluster Assembly (CIA) Machinery and Functions in the MMS19 Protein-dependent CIA Pathway, J. Biol. Chem., 2013, 288, 16680–16689 CrossRef CAS PubMed.
  8. N. van Wietmarschen, A. Moradian, G. B. Morin, P. M. Lansdorp and E.-J. Uringa, The Mammalian Proteins MMS19, MIP18, and ANT2 Are Involved in Cytoplasmic Iron-Sulfur Cluster Protein Assembly, J. Biol. Chem., 2012, 287, 43351–43358 CrossRef CAS PubMed.
  9. A. A. Vashisht, C. C. Yu, T. Sharma, K. Ro and J. A. Wohlschlegel, The Association of the Xeroderma Pigmentosum Group D DNA Helicase (XPD) with Transcription Factor IIH Is Regulated by the Cytosolic Iron-Sulfur Cluster Assembly Pathway, J. Biol. Chem., 2015, 290, 14218–14225 CrossRef CAS PubMed.
  10. O. Stehling, J. Mascarenhas, A. A. Vashisht, A. D. Sheftel, B. Niggemeyer, R. Rösser, A. J. Pierik, J. A. Wohlschlegel and R. Lill, Human CIA2A-FAM96A and CIA2B-FAM96B integrate iron homeostasis and maturation of different subsets of cytosolic-nuclear iron-sulfur proteins, Cell Metab., 2013, 18, 187–198 CrossRef CAS PubMed.
  11. A. T. V. Vo, N. M. Fleischman, M. J. Froehlich, C. Y. Lee, J. A. Cosman, C. A. Glynn, Z. O. Hassan and D. L. Perlstein, Identifying the protein interactions of the cytosolic iron sulfur cluster targeting complex essential for its assembly and recognition of apo-targets, Biochemistry, 2017 DOI:10.1021/acs.biochem.7b00072.
  12. E. Weerapana, C. Wang, G. M. Simon, F. Richter, S. Khare, M. B. Dillon, D. A. Bachovchin, K. Mowen, D. Baker and B. F. Cravatt, Quantitative reactivity profiling predicts functional cysteines in proteomes, Nature, 2010, 468, 790–795 CrossRef CAS PubMed.
  13. A. D. Tsaousis, E. Gentekaki, L. Eme, D. Gaston and A. J. Roger, Evolution of the cytosolic iron-sulfur cluster assembly machinery in Blastocystis species and other microbial eukaryotes, Eukaryotic Cell, 2014, 13, 143–153 CrossRef CAS PubMed.
  14. A. A. Mashruwala and J. M. Boyd, Investigating the role(s) of SufT and the domain of unknown function 59 (DUF59) in the maturation of iron-sulfur proteins, Curr. Genet., 2017 DOI:10.1007/s00294-017-0716-5.
  15. A. A. Mashruwala, C. A. Roberts, S. Bhatt, K. L. May, R. K. Carroll, L. N. Shaw and J. M. Boyd, Staphylococcus aureus SufT: an essential iron-sulphur cluster assembly factor in cells experiencing a high-demand for lipoic acid, Mol. Microbiol., 2016, 102, 1099–1119 CrossRef CAS PubMed.
  16. A. A. Mashruwala, S. Bhatt, S. Poudel, E. S. Boyd and J. M. Boyd, The DUF59 Containing Protein SufT Is Involved in the Maturation of Iron-Sulfur (FeS) Proteins during Conditions of High FeS Cofactor Demand in Staphylococcus aureus, PLoS Genet., 2016, 12, e1006233 Search PubMed.
  17. M. J. Mizianty, W. Stach, K. Chen, K. D. Kedarisetti, F. M. Disfani and L. Kurgan, Improved sequence-based prediction of disordered regions with multilayer fusion of multiple information sources, Bioinformatics, 2010, 26, i489–496 CrossRef CAS PubMed.
  18. D. G. Gibson, L. Young, R. Y. Chuang, J. C. Venter, C. A. Hutchison, 3rd and H. O. Smith, Enzymatic assembly of DNA molecules up to several hundred kilobases, Nat. Methods, 2009, 6, 343–345 CrossRef CAS PubMed.
  19. E. J. Camire, J. D. Grossman, G. J. Thole, N. M. Fleischman and D. L. Perlstein, The Yeast Nbp35-Cfd1 Cytosolic Iron-Sulfur Cluster Scaffold Is an ATPase, J. Biol. Chem., 2015, 290, 23793–23802 CrossRef CAS PubMed.
  20. C. Wiedemann, P. Bellstedt and M. Gorlach, CAPITO–a web server-based analysis and plotting tool for circular dichroism data, Bioinformatics, 2013, 29, 1750–1757 CrossRef CAS PubMed.
  21. S. Mnaimneh, A. P. Davierwala, J. Haynes, J. Moffat, W.-T. Peng, W. Zhang, X. Yang, J. Pootoolal, G. Chua, A. Lopez, M. Trochesset, D. Morse, N. J. Krogan, S. L. Hiley, Z. Li, Q. Morris, J. Grigull, N. Mitsakakis, C. J. Roberts, J. F. Greenblatt, C. Boone, C. A. Kaiser, B. J. Andrews and T. R. Hughes, Exploration of Essential Gene Functions via Titratable Promoter Alleles, Cell, 2004, 118, 31–44 CrossRef CAS PubMed.
  22. A. J. Pierik, D. J. A. Netz and R. Lill, Analysis of iron-sulfur protein maturation in eukaryotes, Nat. Protoc., 2009, 4, 753–766 CrossRef CAS PubMed.
  23. V. N. Uversky, J. R. Gillespie and A. L. Fink, Why are “natively unfolded” proteins unstructured under physiologic conditions?, Proteins, 2000, 41, 415–427 CrossRef CAS.
  24. M. S. Almeida, T. Herrmann, W. Peti, I. A. Wilson and K. Wüthrich, NMR structure of the conserved hypothetical protein TM0487 from Thermotoga maritima: implications for 216 homologous DUF59 proteins, Protein Sci., 2005, 14, 2880–2886 CrossRef CAS PubMed.
  25. B. Ouyang, L. Wang, S. Wan, Y. Luo, L. Wang, J. Lin and B. Xia, Solution structure of monomeric human FAM96A, J. Biomol. NMR, 2013, 56, 387–392 CrossRef CAS PubMed.
  26. K. E. Chen, A. A. Richards, J. K. Ariffin, I. L. Ross, M. J. Sweet, S. Kellie, B. Kobe and J. L. Martin, The mammalian DUF59 protein Fam96a forms two distinct types of domain-swapped dimer, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2012, 68, 637–648 CAS.
  27. P. Tompa, Intrinsically unstructured proteins, Trends Biochem. Sci., 2002, 27, 527–533 CrossRef CAS PubMed.
  28. M. Chang, M. Bellaoui, C. Boone and G. W. Brown, A genome-wide screen for methyl methanesulfonate-sensitive mutants reveals genes required for S phase progression in the presence of DNA damage, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 16934–16939 CrossRef CAS PubMed.
  29. J. L. t. Hartman and N. P. Tippery, Systematic quantification of gene interactions by phenotypic array analysis, Genome Biol., 2004, 5, R49 CrossRef PubMed.
  30. I. Lev, M. Volpe, L. Goor, N. Levinton, L. Emuna and S. Ben-Aroya, Reverse PCA, a systematic approach for identifying genes important for the physical interaction between protein pairs, PLoS Genet., 2013, 9, e1003838 Search PubMed.
  31. R. D. Finn, A. Bateman, J. Clements, P. Coggill, R. Y. Eberhardt, S. R. Eddy, A. Heger, K. Hetherington, L. Holm, J. Mistry, E. L. Sonnhammer, J. Tate and M. Punta, Pfam: the protein families database, Nucleic Acids Res., 2014, 42, D222–230 CrossRef CAS PubMed.
  32. J. N. Agar, P. Yuvaniyama, R. F. Jack, V. L. Cash, A. D. Smith, D. R. Dean and M. K. Johnson, Modular organization and identification of a mononuclear iron-binding site within the NifU protein, J. Biol. Inorg. Chem., 2000, 5, 167–177 CrossRef CAS PubMed.
  33. P. C. Dos Santos, A. D. Smith, J. Frazzon, V. L. Cash, M. K. Johnson and D. R. Dean, Iron-sulfur cluster assembly: NifU-directed activation of the nitrogenase Fe protein, J. Biol. Chem., 2004, 279, 19705–19711 CrossRef CAS PubMed.
  34. S. Angelini, C. Gerez, S. Ollagnier-deChoudens, Y. Sanakis, M. Fontecave, F. Barras and B. Py, NfuA, a new factor required for maturing Fe/S proteins in Escherichia coli under oxidative stress and iron starvation conditions, J. Biol. Chem., 2008, 283, 14084–14091 CrossRef CAS PubMed.
  35. S. Bandyopadhyay, S. G. Naik, I. P. O’Carroll, B.-H. Huynh, D. R. Dean, M. K. Johnson and P. C. Dos Santos, A proposed role for the Azotobacter vinelandii NfuA protein as an intermediate iron-sulfur cluster carrier, J. Biol. Chem., 2008, 283, 14092–14099 CrossRef CAS PubMed.
  36. H. Gao, S. Subramanian, J. Couturier, S. G. Naik, S. K. Kim, T. Leustek, D. B. Knaff, H. C. Wu, F. Vignols, B. H. Huynh, N. Rouhier and M. K. Johnson, Arabidopsis thaliana Nfu2 accommodates [2Fe-2S] or [4Fe-4S] clusters and is competent for in vitro maturation of chloroplast [2Fe-2S] and [4Fe-4S] cluster-containing proteins, Biochemistry, 2013, 52, 6633–6645 CrossRef CAS PubMed.
  37. R. van der Lee, M. Buljan, B. Lang, R. J. Weatheritt, G. W. Daughdrill, A. K. Dunker, M. Fuxreiter, J. Gough, J. Gsponer, D. T. Jones, P. M. Kim, R. W. Kriwacki, C. J. Oldfield, R. V. Pappu, P. Tompa, V. N. Uversky, P. E. Wright and M. M. Babu, Classification of intrinsically disordered regions and proteins, Chem. Rev., 2014, 114, 6589–6631 CrossRef CAS PubMed.
  38. D. C. Odermatt and K. Gari, The CIA Targeting Complex Is Highly Regulated and Provides Two Distinct Binding Sites for Client Iron-Sulfur Proteins, Cell Rep., 2017, 18, 1434–1443 CrossRef CAS PubMed.
  39. T. Yabe, E. Yamashita, A. Kikuchi, K. Morimoto, A. Nakagawa, T. Tsukihara and M. Nakai, Structural analysis of Arabidopsis CnfU protein: an iron-sulfur cluster biosynthetic scaffold in chloroplasts, J. Mol. Biol., 2008, 381, 160–173 CrossRef CAS PubMed.
  40. H. Li and C. E. Outten, Monothiol CGFS glutaredoxins and BolA-like proteins: [2Fe-2S] binding partners in iron homeostasis, Biochemistry, 2012, 51, 4377–4389 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c7mt00181a
Current address: Zoom Rx, Inc. Cambridge, MA, USA.
§ Current address: Department of Chemistry, Emmanuel College, Boston, MA, USA.
This paper is dedicated to the memory of Jessica Cosman. Her passion for science and remarkable talent for research continue to inspire us all.

This journal is © The Royal Society of Chemistry 2017